Enjoy our introductory Primer to set you on the road to the trend following experience. The content outlined in this Primer focusses on the principles and practices we adopt here at ATS and provides a necessary prelude to prepare your mind for the trend following way. 

Introduction

Welcome to the shadowy world of Diversified Systematic Trend Following. This primer is an account of a curious special breed of trader who beats to the tune of their own drum and disregards the news, hot tips, and economic commentary in favour of applying a simple trading technique. Namely to simply follow price either long or short with the faith that liquid financial markets can at times exhibit trending behaviour.

For any financial market observer, they will sometimes experience moments in time when markets exhibit directional price moves that extend well beyond what can be produced by a random distribution of price action. There is some causal impetus that drives this trending price behaviour. These are the price moves that trend followers want to capitalise on.

While the focus of economic literature over the decades has been towards the notion that markets are normally distributed (or random) and are efficient in nature (e.g. The Efficient Market Hypothesis), while being applicable to the markets most of the time, such notions assigned to how markets behave are more academic in nature or gross simplified statements  as opposed to being true descriptors of how the actual market behaves. We find when plotting the market distribution of daily returns for any liquid market, opportunities do exist outside the Normal distribution curve where arbitrage exists for traders seeking a slight edge. However we need to take care as by far the majority of price action defined by the distribution of market returns are indeed random in nature.

Fortunately, we only need a slight edge in the market to then capitalise on this small bias in market data over a large trade data sample where that small edge compounds in a non-linear manner and can lead to exquisite long term net wealth over time.  Granted the market spends most of its time fooling the speculator with ‘false promises’ which makes us particularly cautious in the way we manage risk, however provided we leave ourselves open for the exotic ‘non-random’ market anomaly, we can reach for the stars in our long-term wealth ambitions.

Trend followers are regarded as ‘price followers’ as opposed to ‘price predictors’. We like to think that we react to current price as opposed to attempting to predict its next move. Our spoils are generated when trends extend into significant directional moves. Like a surfer we wait until a swell is observed and then we focus on simply entering an existing trending price series, either long or short, and simply follow the trending behaviour till its conclusion.  There is no prediction involved. We simply hitch a ride on any existing directional trending price series to see where it takes us.

Given our preference for only participating in a trade event when price is exhibiting trending behaviour, we thereby avoid everyday market machinations and focus our attention on the more exotic market conditions where markets only exhibit this trending condition. It is in these more unpredictable market regimes where predictive models tuned to exploit repetitive market behaviour can ‘break-down’.

Trend followers unlike other forms of trader, are more like spiders who simply place their trading traps in likely areas where predictive models break down. They capitalise on any stray victim who simply got their prediction wrong and whose failed trades fall into our webs.

As a trend follower, we are less concerned with ‘why’ markets can at times exhibit trending behaviour, and more concerned with the simple fact that markets ‘do’ occasionally exhibit this trending behaviour. The world in which we live constantly changes and that is the source of all trends. That is as specific as we need to be in our conclusion. In a world of change, then trends will be ubiquitous.

So by restricting our trading opportunities to those more exotic times, when markets trend, a trend follower hopes that once on board for the ride, the trending move continues. We simply wait for a directional price move of significance and hitch a ride on board the possible gravy train while always managing risk.

While trending price series can be constructed from random price data, a trending price series can at times possess directional momentum. In quantitative circles, we refer to the presence of a directional bias in a price series as serial correlation. Serial correlation means that price at say time t=1 may influence price at time t=5. Price at one point in time may be correlated with price at another moment in time. Under a serially correlated relationship, one moment in time may be dependent on another.

Now the presence of serial correlation in a price series is what makes that price series non-random and path dependent in nature. We will be regularly visiting this notion of path dependence in this primer series. Path dependence is a very important concept to understand due to its influence over long term wealth building. So, any form of quantitative trading method with a definitive edge is seeking to harvest this principle of serial correlation. In other words, looking for serial correlation is not solely a feature restricted to trend following.

We could imagine this effect as that created by the impact of repeated favorable news events over time. Each news event progressively confirms that conditions are improving leading to a building consensus amongst participants of this improving state. This leads to buying pressure which dominates selling pressure over the news cycle leading to a bullish trend.

The presence of serial correlation (also called auto-correlation) leads to a bias in the price series over time. Much like the physics principle of a moving objective with mass, once price is seen to trend either long or short, then we infer that there is overall buying or selling pressure imparted in the price (referred to as momentum) creating inertia in the time series which hopefully continues for an extended period.

Serial correlation is rarely omnipresent in a price series and typically clusters. It waxes and wanes over time and is a feature of complex adaptive systems that are never stationery. Most of the time price is independent to its time series and its behaviour approaches a random walk. At other times a trending series with serial correlation may be just part of a larger mean reversion cycle as opposed to an enduring directional event under momentum.

As a result, most of the time when we enter a trend thinking that momentum will persist, we are wrong in our assumption, and shortly after entry, price either retraces in the other direction to our trend following direction, or simply stalls in its momentum and stays where it is. So, to manage the downside risk associated with unfavourable price movement, trend followers are known to apply a simple rule of ‘always cutting losses short’. We do this by the rigorous application of stops and trailing stops for each market in our portfolio. This way, we ensure that while we incur many small losses along the way, no loss is material in nature to our overall level of trading capital.

However, for those less frequent times that we may be right in our trade entry, a trend follower simply ‘lets profits run’ until the market finally decides that the trend ends and the serial correlation in the price series, that leads to this trending bias in the price series, fades. Hopefully by the time the momentum dissipates, a trend follower has been able to exploit this directional bias with a significant gain. Trailing stops are a prerequisite for the principle of ‘letting profits run’. Having profit targets may offer psychological benefits, but their application compromises our ambition of possibly riding a trend of ‘infinite potential’. A trend follower places their faith in the notion that the less frequent outsized rewards over time will exceed all the small losses that are incurred with this technique along the way.

To capture the infrequent ‘outsized’ gains, trend followers need to be extensively diversified in geography, timeframe, system, and market type across liquid markets. Such diversification allows trend followers to only participate in the strongest trends where serial correlation is more likely to reside which have the potential to possibly turn into outlier moves. The typical portfolio of a diversified trend follower extends across different asset classes such as foreign exchange, soft commodities, metals, treasury instruments, equity indexes and even cryptocurrencies. The intent of such wide diversification across asset classes, that frequently display different market behaviour, is that somewhere, someplace, a market instrument is displaying trending behaviour.

Under a diversified portfolio where our philosophy is non-predictive in nature we adopt a perspective of making many small equally sized bets for all the composite return streams. Such a philosophy ensures that no single return stream dominates the portfolio and risk dollars are equally spread throughout. Given the need to allocate risk evenly across markets with differing volatility we typically use the Average True Range or other normalization methods to define our stops levels and trailing stop levels. While many believe that the use of recent volatility is a sign that trend followers actually do use predictive measures, we prefer to view this normalization as non-predictive in nature where we give an equal weight to any liquid market to trend within our diversified portfolios.

While most markets at any moment in time in a diversified portfolio offer no trending opportunities, the intent of diversification across markets, timeframes and systems is to improve your chances of participating in the occasional significant trending opportunity that pays for all the small losses incurred in this waiting game.

An ideal trend following portfolio equity curve is one which possesses a continuously rising equity curve across a time series. This continuous ascent means that at all points across the time series, trending markets were present and sufficient to pay for all the small losses along the way.

It is however unlikely for a diversified trend follower to achieve this steadily rising equity curve. The infrequent nature of trending environments makes our equity curves more volatile than many alternative trading techniques. The reality of adaptive markets is that they can enter protracted periods of noisy or mean reverting condition where enduring trends are simply absent. This means that our trend following portfolios enter periods of stagnation of building drawdowns while these conditions persist.

This does not mean that the portfolio systems are broken. This is a very natural symptom of trend following where our success is contingent on trending environments.

Any trends that may arise within these noisy or mean reverting regimes typically have no substantive momentum embedded within them resulting in a trend follower being victim to many ‘false trends’. This can lead to slow building drawdowns that can last for considerable stretches of time.

Such enduring periods of drawdown are a feature of price following techniques that patiently wait for trending conditions. While there is little skill in simply following price, your skill as a trend follower lies in how you mitigate the impact of these drawdowns through risk management approaches.

This skill relates to how a trend follower can survive over extended periods of unfavourable regime and preserve their capital by mitigating the impacts of drawdown so that when trending conditions resume, a trend follower can fully participate in these more favorable regimes.

Frequently we find that following extended periods of noisy or mean reverting market environment, markets can, on the turn of a pin, suddenly start to exhibit trending behaviour. We also find that, due to the interconnected nature of financial markets, a transition in one market can, like a set of dominoes falling, lead to these market transitions spreading across markets and asset classes.

Trend followers with their diversified portfolios can then capitalise on these events and find that in a very short period, previous drawdowns quickly disappear leading to far higher equity levels and new high-water marks being achieved.

Seasoned trend followers who manage risk well typically find that drawdowns take time to build yet can quickly evaporate as soon as market conditions become favorable.

What gives us faith in the enduring nature of trend following as a robust trading method is the decades of academic research that supports our cause combined with the very long-term track record of professional fund managers that apply this technique.

You see, for those that practice this trading method, we acknowledge that for a financial market to act as an efficient means of facilitating trade through perceptions of value, which is the primary role of a financial market, then a market must at times exhibit trending behaviour. If it did not, then there would be no reason for transacting price in the form of investing,  hedging or speculating in a financial market.

This transactional trade event which is facilitated by the market  leads to an emergent outcome over time where net wealth is transferred, from the many to the few. The winners and the losers change over time and this internal transference leads to trends (or flows) inside and across financial markets.

While speculation in a market is a zero sum game (a negative sum game when including the costs of trading), it is tempting to conclude that the market is random. However the flows that arise inside the market between participants is real and leads to non-random trending behaviour. You won’t necessarily see it on the outside as a gross statistical statement when looking from outside-in, but you will feel it as a participant privy to only a part of this gigantic game of transference from inside-out.

Given that trends are a natural consequence of a changing society and planet, for a liquid market to not display trending behaviour over a long-term data set, then something very strange is going on. It is being deliberately suppressed. Post 2010 we have seen coordinated action taken by the Central banks to suppress the very natural market tendency to want to re-equilibrate and offer trending environments via transition. Like any complex system, the changing nature of participation over time of its constituent parts makes trends a necessity to allow for a complex system to adapt over time and re-equilibrate in accordance with changing participant behaviour.

Our job as trend followers is a deceptively simple one. To simply participate in those trending conditions where and when they occur and at all other times to always manage risk and preserve our finite trading capital.

Now as simple as this statement sounds, to be a participant for trending opportunities requires enduring patience.

We can never be sure of ‘the where’ and ‘the when’ that a financial market will display trending behaviour and sometimes there is a very ‘long wait between drinks’. We are at the mercy of the market in this regard.

So as opposed to other trading techniques that are predictive in nature and exploit an opportunity based on a repeatable pattern of behaviour, trend following is regarded as a non-predictive method that simply jumps on board existing trending price behaviour and hitches a ride while the trending condition lasts.

Over-trading arising from the frustration of patiently waiting for trending conditions is the enemy of a trend follower. We must wait for the market decide to trend and not try to force the situation by attempting to predict the top or bottom of trends or trading any price series that inconsequentially moves. Over-trading inevitably leads to many further small losses that can compromise our ability to ‘pay for them all and then some’ when we enter favorable trending market regimes.

While this technique is deceptively simple, it flies in the face of human intuition as our brains have been sculpted to predict, so any non-predictive method is an exceedingly difficult method to psychologically tolerate. Most retail traders hate the uncertainty with trend following. Unlike predictive methods that attempt to take a controlling stance in dictating our fortunes and pounce on immediate opportunities, we must wait until a trending opportunity arises and then jump into the prevailing trending current and float along with existing price action.

Most of the time we are wrong in our trade entry decision and consequently continuously feel the sting of small losses (referred to as whipsaws) which slowly contribute to building drawdowns, however we participate in this game because during those rare instances we are right, the market can deliver us a windfall that makes the patient wait so worthwhile.

We make our fortunes from the outlier as opposed to the everyday churn of the market and therefore we need to literally invert our thinking to alternative predictive trading techniques that attempt to exploit a repetitive feature of price action.

Now hunting for the outlier sounds like looking for a need in a haystack. To be able to capitalise on these infrequent unpredictable events necessitates that a trend follower needs to be in the right place at the right time. Now while this sounds like a statement of luck, a trend follower adopts a particular technique that turns this apparent statement of luck into a rules-based process that attracts those traders seeking long term sustainable returns.

You see being in the ‘right place’ necessitates that trend followers deploy extensive diversification as a key weapon to participate in trends wherever they may occur…… and being at the ‘right time’ requires that trend followers adopt systematic methods that are available to participate in trending conditions 24/5.

This introduction has hopefully opened your eyes to this peculiar world of diversified systematic trend following. As simple as the technique sounds of simply following price, the practical application of a diversified systematic trend following technique is quite an endeavor, but in this Primer series we are going to give our readers a step-by-step guide to how to adopt this very robust trading technique that is applied by many of the most successful traders in the world.

So, sit back and relax as we take you on this journey of discovery that hopefully will change the way you decide to interact with these markets and set you on the road to long term financial wealth.

Chapter 1: Care Less about Trend Form and More about Bias

When we begin our trend following journey it is easy to be drawn to the idea that we need to classify trends and be prescriptive in the form of trend that we wish to trade. Unfortunately, such a narrow focus can mean that we filter out a vast array of potential trending conditions that may have an enduring directional bias impregnated within them.

Furthermore, can we really rely on our brain as the arbiter of truth? Our brains are engineered to find causality in visual patterns. This can lead us to false conclusions when simply using visual clues alone as a basis to infer meaning into trend form. The host of different interpretations surrounding the context of different forms of trending condition requires us to adopt systematic rules-based interpretations using the data itself to convey any gross statement of statistical meaning such as its directional bias and overall volatility.

We therefore need to understand that while tempting, we should not focus our attention overly on trend form itself. There is little to be gleaned for a trending price series that can help us ‘trade the trend’ apart from the overall bias and the volatility.

Fortunately, we do not need to pay that much attention to trend form as when we are trading on the right-hand edge of the chart, the ultimate form a trend takes can only be determined in hindsight.  Without a crystal ball, the actual form a trend takes to its conclusion is an academic exercise and does not really matter.

While we say that we are trend followers, what we are trying to achieve is to use trends as simply a method to enter a market with our ultimate destination unknown. By only entering a trade during trending market conditions we are filtering our entry method to avoid random or mean reverting market environments. Also, by only focusing on the most significant trends and catching them late in their evolution, we are basing our expectation on a principle found in liquid markets that the more fully formed trend has a greater probability of containing serial correlation within it. With serial correlation embedded within a trend, then perhaps there is a chance that this trend that we are riding will turn into an exceptional outlier.

The Many Forms of Trend

Trends come in many forms. There are countless definitions of a trend found throughout technical analysis literature from formal definitions requiring a few consecutive higher highs and lower lows, to trends defined by moving averages of a price series and so on and so forth.

While trends appear to be easy to visually spot, when it gets to their specific definition, it is not such an easy task.

In Chart 1 below, we can see 16 highly liquid markets spanning several asset classes such as Forex, Indices, Soft Commodities and Metals charted on the monthly timeframe. Visually the trends appear obvious but is extremely difficult to actually define the ‘trendiness’ of markets in categorical terms that all trend followers would agree on.

Chart 1: Market watch for the Monthly Time-frame

Now while many may say in recent times that ‘The Trend is Dead’, Chart 1 clearly refutes this claim stating that at least from a visual perspective trends are alive and well. So, what gives?

While trends clearly persist right up to the present day, the degree by which trend followers can take advantage of these trending environments has become far less simple. The ‘trends of today’ are not as clear cut as the ‘trends of yester-year’. Trends today are more volatile with considerable variation of form. Market participation has changed, and trading behaviour has become more complex. A myriad of speculators now apply a vast array of trading behaviours which considerably alter the nature of trending condition.

Remember that complex markets are adaptive in nature. Trends are emergent structures of financial markets representative of market disequilibrium (periods of market transition) which cannot be categorized into a single definition. Unlike predictive oscillations of price action, the trends that we want to catch (aka the outlier) are inherently ‘uncertain’ in nature and can be found in the fat tails of the market distribution of returns (more on this later).

As a modern day diversified systematic trend follower, our aim is therefore to be less definitive in our opinion of what constitutes a trend as we may find that such an approach encourages us to miss many trends that ultimately manifest as outliers.

Furthermore, with our approach to diversification, we want our systems to capture a variety of trend forms and segments to provide much needed correlation benefits at the portfolio level. Having a diversified portfolio of trend following systems allows us to attack trends across different markets and timeframes significantly increasing the variety of trends that we can participate in.

The ultimate form that trends can take are spread across a spectrum. They may include:

• parabolic trends that start with an initial humble trajectory and then take flight with an exponentially rising or falling directional move of arbitrary duration and degree;
• to perfectly formed trends that closely approximate a linear regression and are easy to trade to;
• slowly evolving and volatile trends that are particularly hard to trade; and
• any combination thereof.

Now as seasoned trading veterans are aware, the systems we use to capture trending market conditions comprise constraints that dictate the type of trend or trend segment that the system can capture. The rules applied to the entry condition, the initial stop and the trailing stop can be thought of as constraints (like a container) that filter the types of price action that can survive while existing within these constraints. As soon as the constraints are breached, then as per the definitional constraints of that system, the trend is said to end.

While the system rules define what trends can be traded, this does not mean that a trend condition is over. It may elect to resume its prior trending behaviour, or take a breather or often, go counter-trend to the preceding prevailing trend direction  for a period of time.

Clearly the system you apply is scaled to a particular form of trend or trend segment whose definition is dictated by the system. So, for a diversified systematic trend follower, they adopt a diversified suite of trend following systems targeted to capture a myriad of forms of different trend.

The Brain likes to assign Causality to Visual Patterns

One of the major issues surrounding trend following is that people like to infer meaning into trend form. This is understandable as our brains have evolved in a natural environment where an edge has been conferred by natural selection through our ability to ‘predict’ outcomes from sensory clues. This ‘precognition’ appears to have given us a distinctive survival edge against those species which have simply ‘reacted’ to environmental threats. The competitive landscape of our natural environment is a hostile one. Any small edge which confers a survival advantage is received by the human species with welcome.

We need to always remember that our brains are encased in our skulls and may not faithfully render the reality that is ‘actually out there’. All it can do is use the initial stream of sensory data inputs to ‘best’ match (aka predict) those inputs against past experience or learning held in memory. It interprets pressure vibrations of molecules as sound and photons received on a retina as sight. If the prediction is right, then our mental models need no adjustment, but once the data stream has been fully received, we can then test the efficacy of that prediction against the actual data stream itself. If there is a predictive error, then the ‘missing’ additional data can be added to our memory for future reference. This process is something we call learning.

Now while this advantage has been extremely useful for our species survival, the predisposition for our brains ‘to predict’ can create problems for interpreting other less familiar complex systems within which we have not evolved.

We must continually remind ourselves that our natural environment and financial markets are different complex adaptive systems. Unlike our natural environment where competition is played out between different species, the financial markets comprise a single species where ‘intelligence’ is par for the course. In such a different environmental context, being a successful predictor does not necessarily convey the same advantage to survival.

Let us have a look at how the predictive nature of our brain may not necessarily convey the same advantage that we have obtained within our natural environment. Look at the results of how our brains interpret the following examples. Now the issues associated with interpreting these examples has been made clear, but just imagine how without applied attention we could easily run into problems if we relied on our brains to interpret market data.

Here are some simple examples to mess with your brain.

Photo 1: Rabbit or Crow?

Photo 2: Troxler Fading (The Lilac Chaser)

Photo 3: Wash your Mouth out with Soap

Having an understanding of the inherent tendency of our brains to predict ‘causality’ into form should create concern when we rely on the brain to visualize chart patterns. The vast amount of price data may or may not possess a causative influence in it, but it is exceedingly unlikely that we will be able infer great meaning from these patterns using our brains.

What we need is an objective way to assess trends which at least we all can agree on. While most trend traders have their own personal ‘definition of trend’ it is not clear whether these definitions have any more meaning than a random plot of market data. It would therefore be very handy if we could apply a principle that standardizes our method of analyzing trending price data. So in the spirit of this agenda, I have had a go at creating my own method of objectively analyzing trends using objective rules based processes that we all can agree on.

Here is how I do it. Using a standard deviation channel, it is quite easy to define the overall bias of the closing price of a time series bar chart by extending the Standard Deviation Channel (SDC) from the lowest lows (swing lows) and highest highs (swing highs) of the series. I can then extend the standard deviation to enclose the maximum adverse volatility excursion of the trending series and provide a context to understand the volatility of the trend. Note that in a downtrend, the adverse volatility is to the long-side and in an uptrend, the adverse volatility is to the downside.

Chart 2: An Objective ‘Rules Based’ Method to Analyse a Trending Price Series

Having such an objective rules-based process to define the directional bias and volatility of a trending price series, we can now revisit Chart 1 and apply this process to observe the significant variation in overall characteristics (refer to Chart 2 above). Using this rules-based process we therefore do not run into the ‘classic’ problems arising from a trend followers’ interpretation of a trend that may be riddled with visual biases.

Chart 3: Market watch for the Monthly Time-frame using a Rules Based Approach for Trend Interpretation

We can now revisit our statements regarding the plethora of various trend definitional forms. We can see from our rules-based application that the bias (slope) and the volatility of trends and trend segments (defined by the width of the SDC) vary considerably across these different markets. Given the plethora of form, there is not a ‘one size fits all’ solution to this riddle.

Having a single definitional approach to trend following significantly hampers our trend following objectives. What we find is that our overly prescriptive approach to what defines a trend significantly reduces our ability to participate in clearly trending market conditions.

So, let us drop the categorisation scheme and go for the simplest statement that catches all. A trend is simply a directionally biased price series that can take a broad number of different forms. Having such a simple definition ensures that our systematic rules-based approach adopts very simple system designs that allow us to have the greatest chance of capturing the greatest variety of trend forms and/or segments.

While I have provided a simple method to objectively define a trending condition, in our world of systematic application, we adopt systematic rules-based methods that avoid the temptation for our brains to impart biases into any form of price action. Each of our trend following systems are used to objectively assess price action without the possibility of imparting any form of discretionary bias. It is through this application of quantitative rules based approaches which allows us to therefore conduct back-tests over market data that are repeatable and free from discretionary judgement.

Do we care about trend form when Trading the Right-Hand Side of the Chart?

Now while we have been discussing the many forms of trend, does this really matter when we are trading the right-hand side of the chart. At the point of entering a trend, all we can see ahead of us is uncertainty. We do not have a crystal ball to tell us the exact form this trend will take. To do so is an exercise in hind-site navel gazing.

You see we do not care about the different form of trends at all. All we really care about is whether the current trending condition we decide to ride may have endurance.

As soon as we jump on board a trending market, the ultimate trending form could take ‘any form’. The reality is that we are not looking for a particular form of trend at all, but rather a trending condition that might have a kernel of serial correlation embedded within it which may drive future directional momentum. It is then up to us to have a system that is sufficiently robust and non-prescriptive to give us breathing room while we ride this ‘uncertain’ trending condition.

When considering what we as trend followers are seeking it is sometimes useful to think in metaphors. So, let us equate trend trading to longboard surfing.

Under our analogy we sit out past the breakers and wait for the swells to arrive before we take the plunge on our boards with the hope that these ‘special’ waves might have some enduring momentum cooked up in them. We are all seeking auto-correlation (aka momentum) in our waves, but we have different designs (aka surfboards) to capture elements of it.

We could have many surfboards in our collection in which we need to choose the conditions in which we use it or we could opt for simple designs that work across a variety of different wave forms.

Now do we focus on surfing the near-shore break or do we swim out past the breakers waiting for the sets of ocean swells to arrive? As experienced surfers we know that it is the swells which are our target as opposed to the normal beach break. Each beach has characteristic wave forms associated with the local conditions and geography of the coastline and they rarely have enduring persistence, but ocean swells have broader features and enduring momentum, and the local terrain has far less influence on their overall form.

While we watch for the waves, we are watching for the ocean swells that may occur in autocorrelated sets of waves. Under this arrangement we can be pretty sure that these wave forms have the necessary enduring momentum to give us a great ride.

Rather than catching all waves we are patient long board riders sitting far outside the normal break waiting for the exceptional before we dip our toes in. We do not know the exact form this potential wave event will take, but we have a buoyant simply designed relatively large board that allow us to ride a variety of possible forms. We replace skill associated with the ‘known’ with a design that accommodates for uncertainty.

Chart 4: Surfing the Trend

It is the Bias in the Time Series that Really Matters

Ok, so having equated trend following to surfing let us get back to business in understanding what we really are seeking to capture in our trend following world.

Despite the variety of different forms which can be confusing, what we are seeking in our trend following pursuit is to trade those trends with serial correlation embedded in them. We can therefore adopt a more valid form of categorisation treatment for trending environments. Under this method of colloquial categorisation, we can classify trends into two broad categories of either:

• Random trends; or
• Real Trends.

The difference between the two is based on whether there is an underlying bias or serial correlation present in the price series where a price earlier on in the price series has an impact on a price point later in the series, and the two price points are defined as being correlated. This underlying bias can be thought of as a causative driver that gives a momentum nudge to the price direction.

So here is an example of a trend constructed from random price data constructed from 10,000 price points.

Chart 5: Plot of a Random Price Series of 10,000 data points

For the trader who views these price behaviours all the time, the price series of Chart 5 would be conferred to have a bias in the price series that has led to this trending price behaviour. In other words, the trader may see this price series as being a ‘real’ trending price series dependent on how they define what a trend is.

The problem however is that this was constructed through random price points and by definition, this trending series will have no trending persistence into the future. This price distribution could be plotted within a Normal distribution of market returns.

Chart 6: Plot of the Same Random Price Series with applied Bias (Serial Correlation)

To continue with this example, we now insert a degree of serial correlation into this random price series to obtain Chart 6. Visually this would also be conferred as a trend.

However, unlike the Trend in Chart 5, the bias has persistence and if it persists will have a ‘biasing’ effect on all future price data. We can see that the impact of this bias on the price series creates a ‘drift’ in the series. This is clearly the trend that we want to trade…but it is not the form of the trend that is important to us….but rather the bias (or the drift) that exists in the underlying price series that has causative power.

On a random price series, the impact of a persistent bias can result in many different forms of trending pattern. We cannot be choosy about the trend we want to change by nature of its form as a myriad of forms are possible with bias impregnated within them.

However, for the trader, it is visually impossible to deduce if the price series has a bias in it. This can only be deduced through hindsight statistical treatment. When trading on the right edge of the chart into the unknown, the ability to distinguish between both forms is impossible. For example, if we compare both charts in hindsight in Chart 7, we can clearly see the differences….yet our ability to distinguish between the two trends by visual clues alone is impossible as the bias is represented by the difference between the price series of Chart 5 and the Price Series of Chart 6 and not the form of each trend. It is the bias itself that illustrates the causality of the correlated price data.

Chart 7: Comparison Between the Random Price Series (Chart 5) and the Random Price Series with Bias (Chart 6)

So that puts us in a dilemma as Trend Traders. Do we trade all trends and run victim to ‘false’ trends?

A way we mitigate against the propensity to be caught by the illusion of a random trend is to only participate in trends when they are well formed and have a greater probability of being non-random in nature. We wait for more extreme trending conditions before we elect to participate in them as we need the price series in our data sample to build to sufficient levels where randomness has a lesser say and any enduring auto-correlation or bias in the series has a greater chance of expressing itself.

Unfortunately, there comes a time, despite our deliberate delay, where we need to participate in all trends that meet our criteria of being sufficiently material. If we do not, then we can be subject to what is referred to as a Type 2 Error. Namely, we miss the opportunity to participate in an outlier.

Given the infrequency of major outlier events, to miss one of these significant windfalls can be the difference between success and failure for us trend followers. We need these major windfalls to pay for all those losses we have incurred along the way. So, we must catch all of them. The way we do this is through diversification and enduring patience waiting for these potential events to unfold.

While there is no guarantee that the trend we ride has any causative substance, as we may just be riding an unusually material random trend, by delaying our trend entry to more extreme levels (on the verge of what we refer later in this Primer series as the “Gaussian Envelope”), then there is a greater chance that we have avoided the randomness (or noise) associated with random price series.

Chapter 2: Divergence, Convergence and Noise

So far in this Primer we have discussed that when considering market trends, it is wise to not get too attached to the notion of trend form. We have also introduced you to the notion of ‘real trends’ or ‘fake trends’, however there is another powerful way to also consider how price action can be applied to trading which offers considerable insight.

A divergent or convergent time series provides a quantitative way to classify any trading strategy seeking to exploit serial correlation and each definition lies at the spectral ends of price action. This therefore includes our trend following approach and a host of other trading techniques that seek serial correlation in a time series.

In simplest terms convergence means coming together, while divergence means moving apart and these definitions provide a very powerful way to classify trading styles into two broad camps that relate to the region of the distribution of market terms from which alpha is derived (more on this later however Chart 8 below introduces you to this concept).

Chart 8: Where an Edge is Derived from the Market Distribution of Returns

Chart 8 above highlights how convergent trading methods focus on the non-randomness that exists in the peak of the Distribution of Market Returns, whereas divergent methods focus on the non-randomness that exists in the left or right Tails of the Distribution of Market Returns.

An understanding of divergence and convergence is essential, as the methods we deploy to capture the elusive edge that exist in these zones of opportunity are, the antithesis of each other. By confusing the issue and treating both convergence and divergence in the same way within a diversified portfolio of trading solutions, you will quickly find that each method detracts from each other.  It is possible to trade both methods with different processes, but it is very unwise to include both in the same portfolio.

Now while a speculator seeks the arbitrage that exists within a convergent or divergent price series, a ‘noisy’ time series provides a problem for all traders. Unlike a convergent or divergent time series, a noisy time series lies in the middle of these spectral extremes of price action where there is no observed bias in the price series. As a result, a noisy time series is deemed to hold no arbitrage potential.

While we use the term random to express a noisy time series, this is not used in the same strict definition as that you may find in a physics textbook. Rather, a noisy or random time series in financial market terms is just a trading pattern that holds no arbitrage potential for the traders that call them ‘noisy’. There are different scales used to quantify and trade market data and what may be termed random or noisy in the longer term, may have arbitrage potential and be deemed non-random in the shorter term…..so don’t get too hung up on the term ‘noisy’. It is just a useful way to define a price series that holds no arbitrage potential for the quantitative trader throwing around the ‘name’.

Quantitative Techniques

Unlike many other forms of trading style that seek to understand the causative reasons for why price decides to move in a particular direction, trend followers choose to simply refer to the price action itself. This therefore places a diversified systematic trend follower into the definitional category of technical analysis and more specifically quantitative traders….. as while we trade based on signals generated from price action, we deploy statistical methods to interpret price data.

As discussed previously we do so to eliminate the propensity of our brains to ‘fool us’ and rely on the data itself to tell us an objective story that can be agreed on my all trend traders who adopt a quantitative method under the same assumptions. We can also return to our assumptions in a back-test and repeat these experiments with fidelity and thus use the back-test as a method to ‘objectively’ test these assumptions.

Many other forms of trading technique rely first on identifying the causal reasons for why price moves, then armed with this information, decide how best to capitalise on this causative driver. So for example fundamental investors first identify the causal factors for why price moves such as a particular market being overvalued or undervalued, and then they evaluate when those conditions are present. If these conditions are found, then a fundamental investor places a trade that takes advantage of the potential future price move associated with this assumed causal driver. The timing of this possible future move is however unknown. Furthermore, our ability to test these assumptions in a repeatable back-test is a fools errand as the process employed is highly discretionary.

Another example of a discretionary trading technique that first assigns causation to the move before striking are News Event Traders. This type of trader specialises in trading News Events where they closely monitor the news feeds and undertake scenario tests to evaluate their importance and possible impact on price behaviour. Having undertaken their analysis prior to the anticipated News Event they focus on the News release data and determine which scenario they will apply based on the analysis.

These prior popular trading methods (particularly abundant with discretionary processes) are examples of predictive methods that seek causal reasons for why price moves so they can predict a future outcome.

While we strongly agree that the primary causative driver for trends are economic fundamental factors, we avoid the need to speculate which of the drivers are being applied and the timing of those impacts to create business cycles and different trending conditions.

We however as quantitative trend traders simply take all our trading clues from the statistics derived from historical price data action alone without any heed of any possible underlying causal factors. We understand that ‘all roads lead to Rome’ so rather than seeking to understand the possible causative reasons for price behaviour which can be vast and ‘fleeting in nature’, we focus on the primary (as opposed to the possibly derivative) information source for our technique……that magic clue called ‘price action itself’.

You see many alternative methods of trading are based on a philosophy that may or may not have a bearing on price ‘in the moment’. As a result, many alternative forms of technique that use derivative causative drivers of price may be appropriate “for a period of time”, but may not necessarily “stand the test of all time”.

However, unlike these derivative methods, by trading the primary agent (namely price action itself), such a canvas will never alter. So long as financial markets offer potential for arbitrage, then that record will be recorded in price itself. We actually do not need to know the ‘why’ at all in trading it.

Quantitative Camps

Fortunately, despite there being a plethora of different trading styles, quantitative techniques can be bundled into two discreet camps. Those who target convergent opportunities (such as mean reversion or counter-trend methods also referred to colloquially as short volatility methods) and those who target divergent opportunities (such as trend following and momentum methods referred to as long volatility methods).

It is possible for a trader to adopt both broad styles in their trading repertoire, but each method is the antithesis of each other and requires diametrically opposed treatment.

In attempting to understand why trend traders don’t consider that they ‘predict’ future price moves, it is useful to think of the reasons in ‘divergent or convergent’ terms.

A convergent trader is regarded as backwards looking. They use historic price action as a basis to evaluate a price level which is regarded as ‘an intrinsic’ level that price historically has converged towards. Once this ‘intrinsic price level’ has been calculated, then a predictive stance is assigned to all future price action on the understanding that a trading opportunity exists when price takes an excursion far from this intrinsic level and is therefore ‘predicted’ to eventually revert back to this intrinsic level.

A divergent trader on the other hand is regarded as forwards looking. They use historic price action simply as a basis to understand where price is ‘unlikely’ to be in the future if prices diverge away from their historic state. Now while a trend follower is one such divergent trader, we simply do not know where this future price will diverge to or the possible direction of this divergent move. We just understand that divergence implies that change is ubiquitous. The lack of any historic benchmarks to use in this philosophy of divergence makes trend followers ‘risk managers’ as opposed to ‘profit seekers’ as the only means they have to possibly take a controlling state in the uncertain future state of price is to mitigate risk events but leave the possibilities open for unbounded divergence. We consequently do not use back-tests as a basis to predict future returns but rather, we use back-tests as a basis to test the sensitivity of our risk management assumptions.

Given the vast disparity between the two different schools of application we avoid convergent methods like the plague as they can undermine our ambition of wealth building if not applied judiciously. This is despite the attractive ‘ephemeral’ lure of what convergent markets can deliver from time to time. We will spend a great deal of time in this Primer Series explaining the ‘why’ for this particular stance taken on avoiding convergent methods.

Divergence

Trend followers like to categorise their opportunities as arising from financial markets that are in disequilibrium (or a state of transition). This tendency towards transition can lead to tail events when the transitions are large. We tend to think that trending markets exhibit this state when previous equilibrium levels are no longer respected. During this period of transition, price takes a directional journey either higher or lower to ultimately settle on a new equilibrium level for a period of unknown duration. The extent of the journey between equilibrium levels varies and periods of transition can be of any arbitrary duration and extent.

Given the tendency during transition for price to diverge away from previous equilibrium levels, we refer to this price action as being divergent in nature.  For those traders that capitalise on divergence (eg. trend followers and momentum traders), we refer to them as ‘divergent traders’. Divergent traders can never predict when and for how long market conditions remain in transition and they simply work off the principle that during these uncertain market regimes, it is a wise tactic to simply float with the market and participate in the journey.

Given their non-predictive stance and uncertainty regarding the possible extent of the directional move, these traders simply cut their losses short and let their profits run without any profit targets deployed. They also do not pay a great deal of attention to their trade entries, but they do have a lot of them as provided that an existing trend is in place, they simply jump on board for the ride. By having a variety of different entry techniques to catch a trend, this provides much needed diversification that ensures that no trends of substance are missed. Furthermore, diversification of entry is a much-needed method to reduce correlations between return streams at the global portfolio level.

Trend followers need to wait until a trend is formed before they are attracted to participate in the price action. As a result, they never catch the beginnings of the trend and are always late to the party. Furthermore, the stance taken of letting profits run mean that trend followers always leave the trend when it ‘bends’ and as a result never are able to exit the trade at the optimal point of the trending condition. Given the late entry and late exit, trend followers acknowledge that they never can exploit the full length of the trend and they therefore focus their attention on catching the ‘meat of the trend’. The infrequent nature of trending events combined with their potentially unbounded nature in duration and extent naturally leads to trend following strategies having low trade sample sizes per system deployed. It is through diversification at the global portfolio level where we up the ante in trade frequency but is never a case at the level of the individual return stream.

The price paid by this form of trading style is that we can never know for how long a divergent condition will persist and there can be many false leads provided by trends that are randomly constructed without possessing any ‘causative’ bias in the trend construction. As a result, most of the time, traders that participate in apparent trending market conditions get it wrong and price action turns in the other direction as soon as you jump on board. It is only in the quite rare instance that a market will continue to trend given a causative bias embedded in the price series and therefore many researchers refer to a trending price series as being a market ‘anomaly’.

Typically, these methods stack up over a large trade sample size simply due to a handful of ‘outlier’ trades that have delivered sufficient profit to pay for all the losses along the way….and as a trading method, divergent trading methods can rarely deliver a sustainable cashflow over time. These methods are tailored for wealth building as opposed to cashflow generation.

Most retail traders simply do not possess the patience or required level of capital to participate in this form of wealth generation.

Convergence

At other times when markets are undergoing periods of stability, price can exhibit repeatable price behaviour patterns around an equilibrium level and during these periods of ‘stability’ trend followers need to avoid these conditions like the plague. These are the market conditions that suit methods that attack the tendency of price to move towards these equilibrium conditions. Traders that work off the principle that price will converge towards a historic equilibrium are therefore referred to as ‘convergent traders’.

Most retail traders who trade price action prefer to tackle this form of market behaviour as they can exercise control in predicting where price at any level is likely to land in the future. The relative frequency of this repeatable cyclic condition also leads to high trade frequencies that attempt to match the rhythm of the cycle and if done well, can lead to high win rates while the condition exists. Who wouldn’t be attracted to convergence? Well us trend followers for a start that are sceptical of a markets ability to retain a repeatable pattern over time.

During periods of market stability, predictable oscillations about an equilibrium can be very stable and endure for significant periods of time. Given the obvious nature of a repeatable market action this style of trading requires preciseness of entry and exit to take full advantage of the enduring repeatable pattern.

Convergent traders therefore place a great deal of emphasis on both the entry on the exit and apply profit targets that target the equilibrium levels around which these patterns oscillate.

Unlike their divergent cousins, convergent traders focus on the commencement and end of trend turning points. Having an understanding of the markets rhythmic oscillations allow this style of trader to be specific in pin-pointing areas that are most likely for price to commence their mean reverting move back towards equilibrium. These traders are not looking for diversified methods in targeting these opportunities. In fact, to be diversified using convergent methods significantly dilutes the premise that these traders can predate on a repeatable market condition. It pays to be precise when targeting convergence and a diversified approach is an obstacle for this required ‘preciseness’.

It is during predictable oscillating market conditions where convergent styles have particularly high win rates and can support many trades that are curve fit to the enduring pattern of price behaviour.

During extended periods of repeatable price behaviour, traders can participate in ‘easy money’ derived from the spoils of predictability and these methods are cash flow generating in nature and very appealing to retail traders. In fact, a repeatable market condition over an extended timeframe often leads to traders upping the ante and increasing their position size to capitalise on that opportunity and make hay while the sun shines. It is this temptation you need to strictly avoid when trading convergent methods as, unlike divergent methods that place their emphasis on managing risk by always cutting losses short, convergent traders often forget the fact that market conditions never persist. When market conditions change, the over-leveraged position sizes then lead to a string of very large losses that are sufficient to wipe out your trading capital in a quick series of unfavourable trades.

Unfortunately, the price paid by adopting ‘convergent trading principles’ is that you must attack the current condition while it lasts, and there is never any guarantee in the enduring nature of that condition. As the opportunity is exploited, markets then start to move to new equilibrium levels and the repeatable pattern no longer exists.

Convergent traders continuously need to monitor their performance and be prepared to jump ship when drawdowns arrive. It is exceedingly difficult to evaluate whether the drawdown is a temporary phenomenon where market behaviour will once again resume its predictable oscillation, or whether that predictable market condition will ever occur again.

So, while cashflow is easily obtained during periods of market stability, when markets alter their behaviour once the condition has been exploited, convergent traders need to be on the ball and quickly stop their trading systems before they encounter large string of unfavourable losses that destroy the small pool of cash flow that has been previously generated.

Convergent traders need to ‘strategy hop’ as market conditions change continuously searching for the next exploitable pattern of stability and repeatability.

What many retail traders do not understand is that a sustainable trading life is as much about being able to endure unfavourable market conditions as it is about being able to exploit a single market condition. While many new entrants enjoy the spoils of an enduring market condition, these spoils are frequently totally lost when market conditions become unfavourable.

The attraction of a high win rate, high trade frequency and the prospects of a regular cashflow is what typically attracts retail traders to the trading game. Unfortunately, the lure is a false promise given the non-stationery nature of financial markets. While you may have luck on your side for a short period or you are able to effectively exploit a ‘convergent market condition’, there is no guarantee that you will be able to survive during those times when markets offer unfavourable market conditions.

Noise

While the market spends much of its time converging around an equilibrium and far less of its time in dis-equilibrium diverging away to a new area of stability, most of the time markets, in their efficient state, exhibit noisy market behaviour. During these market conditions, there is no arbitrage potential from any directional stance to exploit and gross price behaviour simply adopts random directional price moves. While many traders can achieve profits during these ‘noisy random conditions’ these spoils are derived from luck alone. These conditions are very like the conditions you face when playing games of chance at the casino. Your success is a factor of luck alone. There is no directional ‘bias’ in the market in which a trader with an actual edge can participate in.

In fact, when including the costs of trading which apply a slight frictional drag of bias to overall trading results, over the Law of Large numbers your luck is likely to slowly deteriorate to a point where you are left with no trading capital if you never are fortunate in finding an edge derived from a ‘real’ bias in the underlying market data.

Chart 9: Thirty Random Equity Curves resulting from trading a purely random price series – 500 Trades

Have a close look at Chart 9 above. This chart reflects 30 possible equity curves resulting from 500 trades undertaken by a trader when market conditions display no actual edge. Each trade adopted a position size of 2% of equity per trade. Starting with an initial equity position of $1,000 you will see that some lucky traders lifted their equity to $2,000 having undertaken 500 discreet trades.

Now luck swings both ways as well as you will also note that some ‘unlucky traders’ saw their equity rapidly diminish to a few hundred dollars at the end of the exercise.

Now you might find it hard to see, but there is a slight ‘frictional drag’ present in these equity curves arising from the impost of trading costs on the series.

So let’s see how the lucky traders fare when we extend the data sample with this frictional drag for say 10,000 trades.

Chart 10: Thirty Random Equity Curves resulting from trading a purely random price series  -10,000 trades

As Chart 10 above shows, it is a sad story for every participant in this story of randomness. The slight negative bias imposed by the trading costs ensures that over the ‘Law of Large’ numbers that small bias with compounding will diminish any trader’s account.

The Interplay between Divergence, Convergence and Noise

Now while the preceding definitions create an impression that these different market states of divergence, convergence and noise appear to represent distinctive market conditions, the reality is that the market state can exhibit various combinations of these price behaviours at any point in time. The result is that a complex market can exhibit a myriad of different market states.

For example, when we see a clear trending series representing a divergent market condition, we may also find a degree of convergence within that overall directional behaviour. This leads to the classic overshoots and undershoots along a trend cycle that frequently is interpreted as wave behaviour. We can also find that trending price series frequently stall in their trajectory within defined congestion zones where convergence and noise can dominate.

Sometimes the divergent condition is so pronounced that alternative market behaviour of noise and convergence is suppressed to such a degree that a trending condition never retraces and exponentially accelerates producing a parabolic time series.

Some traders who believe that they are trend followers will wait until a retracement of the trend to then jump on board the overall trend direction seeking a high reward to risk with a larger position size than more traditional ‘breakout’ trend traders. This style of trading is predictive in nature as the assumption surrounding this tactic is based on a prediction that a retracing price will ultimately revert back to the trending condition.

Given that more traditional trend traders adopt a non-predictive posture, they tend to operate off breakout entries that simply directly enter a trending move. The advantage of a ‘non-predictive’ breakout entry is that all trends are captured this way. Waiting for a retracement entry before entering a trend can result in a missed trending opportunity. While this predictive class of trend trader can achieve some great reward to risk results from their ‘sniper entries’ into a divergent trend that also possesses a convergent signature, they tend to miss the trends that never retrace….such as the noted parabolic moves seen with some cryptocurrencies and stocks such as Tesla. The parabolic move is a major windfall for the trend follower as these can lead to gargantuan trends, and you don’t want to miss any of them.

Now the plethora of different conditions a market can take based on the relative mix of divergent, convergent, and noisy states makes the trading game a very difficult one over the long term.

Many traders fall under the assumption that a particular market state is here to stay and while they may make hay while that sun shines, their prospects of wealth building returns over the very long term is an unlikely one.

We all are drawn to the lure of profits trading these markets, but to survive in this game we need to flip that mindset. This game is a game of survival and not a ticket to instant riches.

While the complex systems of our natural environment and the financial markets are different beasts, there is a common theme running through all complex systems. The winners are the participants that survive and can live to tell the tale to their offspring, and these are few and far between. The fossil record and the trading record have a very large graveyard of unsuccessful profiteers.

Chapter 3: Revealing Non-Randomness through the Distribution of Market Returns

For any liquid market we can plot a distribution of returns over a large time interval which through its expansive coverage incorporates many different market regimes. What we find in this plot is that the nature of the market distribution is typically not normally distributed. What this means is that the serial correlation present in the time series, provides a non-random bias to the time series leading to the conclusion that markets are not purely efficient or purely random in nature.

If markets were 100% efficient, then this implies by definition that speculative practices are a fools game. Having no non-random bias in the price series implies for example that price at t=10 has no relationship to price at say t=1. In other words, price at each time interval is deemed to be independent in the history of price action.

So, if we interpret that each price interval in a time series is truly independent, then when plotting a random distribution of price over large data samples, we should find that the histogram of market returns plots in accordance with a normal distribution. This is a characteristic distribution that signifies that there is no edge available in this market and that participants are effectively gamblers at the casino.

This does not mean that there will not be winners and losers in this game just as we find at the casino….but what it does mean is that the outcome for a participant is solely dictated by luck or random chance. You may be lucky and have large strings of wins just as you may be unlucky and have large strings of losses….or you may breakeven.

Under a normal distribution we can describe the plot with only 2 values. The mean and the standard deviation. This assumes a stationary market where conditions can be described perfectly by these two values. Having these two values, we can then plot a histogram that adopts a bell-shaped curve symmetrically distributed about the mean. In quantitative circles we refer to this normal distribution plot as the Gaussian envelope.

Now clearly, as speculators, we want to believe that a Normal Distribution does not faithfully describe the market state over a large data sample and that we can exploit some ‘real’ underlying bias in the data. We want to trade an uneven playing field where we can exploit the edge we find in that arena and turn a game of luck into one of skill which, under the Law of Large numbers, has enduring permanence.

Fortunately, when we do this exercise and plot the market distribution of returns of large samples of real (as opposed to theoretical) data, we find that it is not quite a level playing field. We can run this exercise over any large data sample of market data for liquid markets and we inevitably find that these markets are ‘not quite efficient’.

Chart 11: Histogram of Daily Returns across a long time interval for Soybeans, Crude Oil, Spot Gold, DAX, S&P500 and EURUSD

For example, Chart 11 above demonstrates how real market data when using long term time horizons clearly have kurtosis where real market data strays outside the Gaussian envelope (normal distribution) into the peak and tails of the distribution.  These areas in the plot that lie outside the normal distribution…. are the areas from a quantitative viewpoint where a real exploitable bias in the data exists.

In other words, real market data exhibits a plot that represents some other type of distribution such as a Cauchy distribution. Debating the actual form of distribution is another fierce academic exercise in quant world which we actually can avoid as we are hopefully traders as opposed to academics, but hope is offered in that real market data is clearly not purely random (or efficient).

For any speculator seeking to exploit an edge, all that matters is that the distribution is not randomly distributed. Where the distribution departs for a normal distribution is where this edge resides.

What we find when you plot the market distribution of returns over a very large data sample is that the plot of the distribution does not fit snugly within the Gaussian envelope. In fact there are two clear zones where an edge resides in the distribution. Namely the peak of the distribution and the tails of the distribution. When we examine this entire distribution closely we find that it is a composite of different distributions reflective of a complex systems emergent and adaptive nature.

A more realistic interpretation of the distribution is that it has multiple means (averages) and multiple standard deviations that reflect different market regimes. The peak of the distribution can broadly extend across a zone and may not be ‘sharp peaked’ having no centralised defined locus and the tails of the distribution can be fat or thin signifying a range of different standard deviations associated with multiple means.

These two discreet zones lying outside the Gaussian envelope reflect the zones that define the edge held by convergent methods and the edge held in divergent methods.

Chart 12: Histogram Plot of the Monthly Market returns of S&P500 over a 70 year data sample

Source: https://towardsdatascience.com/are-stock-returns-normally-distributed-e0388d71267e

Now Chart 12 displays the distribution of monthly returns for the S&P500 over a 70 year timeframe. Visually you can see that the shape of the distribution can significantly depart from a Normal Distribution. There is variation in this departure  for every liquid market demonstrating that different markets do have slightly different ‘distribution signatures’ but the material differences between liquid markets are rarely material. This allows us to generalize and conclude that markets while ‘mostly efficient’ are not perfectly efficient.

Now the zones that lie outside the normal distribution define the zones where an edge resides and conveniently provide a basis to categorise whether you are a convergent or a divergent trader. As discussed previously both methods seek to exploit serial correlation in a price series, but at different zones within the distribution of market returns.

Convergent methods seek the edge associated with the bias in the daily returns at the peak of the distribution, and divergent methods seek the edge associated with the bias in the data towards the left and right tails of the distribution.

Convergent methods exploit the zones around a markets central tendency (equilibrium) or the mean of a historical price distribution. The frequency of this tendency towards equilibrium is large represented by the high values of the probability density function (refer to Chart 12), yet the extent of the directional movement exhibited by the daily returns is compressed within a tight zone around the peak. This leads to the characteristic high Pwin% and low Reward to Risk (R:R) of convergent approaches.

Divergent approaches exploit the zones of the tails of the distribution which are rare occurrences observed by a very low probability density function (refer to Chart 12) but have a very wide range of possible values of any arbitrary extent. This leads to the characteristic low Pwin% but high R:R of these approaches.

We also need to note that the entire distribution can shift from left to right and is not necessarily placed symmetrically around “0” which indicates a degree of skew in the overall data. For example, the long bias in equities associated with the method of index construction shifts the distribution to the right causing a skewed distribution leading to longer left tails than right tails. For markets such as currency pairs that have no preferred bias, the distribution is frequently more symmetrically distributed with equivalent left and right tails.

Having an understanding of the market distribution of returns and its variation for different liquid markets is very helpful for the speculator.

Firstly, it gives us confidence that markets while generally efficient offer opportunities for exploitation and secondly it tells us that the available edge in a liquid market is only slight. By far the majority of price action cannot be discerned as having any exploitable bias. This should therefore make us ‘wise risk managers’ as for much of our trading activities, we are going to be ‘fooled by randomness’.

Finally, it tells us what distinguishes divergent methods from convergent methods and that to participate in either form of opportunity requires totally different trading behaviours.

Convergent methods harvest the bias that concentrates around the peak of the distribution of returns whereas divergent methods harvest the bias the concentrates in the tails of the distribution.

Therefore, for convergent methods, intense effort needs to be placed on identifying the mean (or equilibrium point) around which this bias persists and requires a predictive mindset to exploit it. Having an understanding of what constitutes the equilibrium allows you take advantage of that opportunity with as many trades as possible for as long as it lasts.

With repetitive cycles comes a mindset that starts to lock in that predictability. Many convergent traders start compromising on risk management by progressively pushing there stops further out or eliminating stops altogether in preference to performance exits. While there is no question that stops do actively introduce drawdowns into system performance, the potential for large losses creates a negatively skewed obstacle for convergent traders when conditions cease. Under negative skew, the average profits are many but much smaller than the few large losses….but this only applies when conditions are favourable. When conditions become unfavourable, the large losses become a large consecutive string of large losses and the convergent strategy can plummet towards total risk of ruin.

Chart 13: RIP Long Term Capital Management

Source: Reddit

This is the real obstacle for convergent trading. Convergent techniques tend to warehouse risk (more on this term later in the series). Unfortunately, you do not find this feature stressed enough in the risk metrics adopted by industry….yet trend followers are fully aware of this little nugget of wisdom and is a major reason for why we avoid these techniques. The message is this….”stay away from strategies with negative skew”. It is akin to not having a release valve on a pressure cooker. They are risk time bombs waiting to explode.

Successful convergent techniques also typically deploy profit targets centered around this central tendency designed to exploit the short-term nature of the convergent move. The high frequency of returns based around this central tendency tells convergent traders that they will be right most of the time. This can lead to an associated bias that reinforces predictive behaviour such as a willingness to apply more leverage to the strategy using enhanced position sizing….however the associated hubris ‘of being right much of the time’ can lead to extreme losses for the confident predictive trader when they are wrong and the equilibrium dissolves as markets transition and start to diverge.

Divergent methods that harvest the bias in the tails must drop any predictive notion as these events are unpredictable in nature and very infrequent, but through systematic application of a rules-based process again and again under diversification, can take advantage of these anomalies if and when they occur. Unlike convergent methods, divergent methods that target the tails of the distribution have no profit target in mind as the uncertainty associated with the extremity of these moves cannot be anticipated. Now the penalty clause for divergence is not nearly as severe as it is for convergent methods. While markets are noisy or during convergent phases, most of the time, divergent traders simply do not participate in trade events, but if they get it wrong, which may happen many times in succession, the risk management practices used by this technique simply leads to many small losses. These can and do pile up over time, but they can be managed, and control exercised over capital preservation. Ideally, divergent traders define the Gaussian envelope itself and use this envelope as a basis to decide when to trade. As prices start to extend to the limits of the Gaussian envelope, divergent traders ‘lock and load’ in anticipation of an imminent trade event to ‘destination unknown’.

The very infrequent nature of the tail event makes the divergent trader very uncertain and forces them to diversify and adopt strict risk mitigation measures for the times that they are frequently wrong and need to preserve capital. All trades apply equal small $risk bets as there is never any guarantee if a trade will result in a possible windfall.  However, the rare times we are right, the market can deliver significant wealth building returns despite the small bet sizes given the non-linear moves that are capable within the left and right tails.

The dichotomy between both approaches requires totally different methods and trading behaviours, yet both can yield alpha….however given the vast differences and its impact on the psychology, they don’t mix well together for the psyche. We strongly recommend that you pick your poison and stick with it. Playing the game from both angles inevitably leads to tears and you frequently find that one approach will detract from the other.

What needs to be noted at this stage is that assuming an edge can be found at both loci (the peak and the tails), then a failure in one approach is the gain for the alternative approach. This is why we state in this zero-sum game that trend followers operate in that zone where ‘predictive model’s break-down’.

You may be familiar with the movie “Margin Call”. In that movie, the Hedge Fund experienced a risk event where market conditions strayed outside the tolerable thresholds of their modelling resulting in the company recognizing that they needed to quickly offload the risk they were holding to avoid a corporate collapse. This style of hedge fund adopted convergent processes and warehoused risk within their models. Sitting on the other side of a trade in events like these are the trend followers who capitalise on those prey whose ‘models break down’ in these extreme market conditions.

Scene from Margin Call – The Movie (with Poetic Licence)

Apart from the difficulty faced by convergent models in managing risk another issue commonly faced in convergence is that there is considerable room at the peak of the distribution for many different central tendencies. Choosing the right equilibrium level is a task that requires considerable statistical expertise in application and there is no guarantee in how long that condition will last until a new equilibrium level is found. The markets’ ability to select numerous equilibrium levels over its lifetime requires you to stay on your game and correctly identify when conditions have changed, and it is this issue of convergence which results in many account blow-ups. These methods focus on chasing returns with little emphasis placed on risk management.

However, for Divergent traders, risk management is our primary goal. To stay alive until an exotic directional anomaly occurs whereby our systems with their rules-based approach latch onto the outlier which the market can deliver from time to time. If we can stay outside the bounds of the normal distribution and only participate at those times where we are at the edge of the distribution near the tails, then we can avoid the incessant noise and mean reverting tendency of the market. All we must do is trade both the left and right tails of the distribution with tight stops to prevent us from ever entering the dreaded fat tailed environment in the opposite direction to our chosen trade direction.

What gives us faith in our philosophy is that we are realists who accept that markets are complex adaptive systems as opposed to theoretical constructs that obey statistical principles riddled with assumptions.

Real world events induce abrupt change to otherwise well-behaved markets creating extreme variation within these otherwise mean reverting or random conditions leading to a significant proportion of overall price variance.

Chapter 4: Characteristics of Complex Adaptive Markets

Throughout this Primer we have been repeatably referring to the financial markets as being complex adaptive systems, and in this Primer, we would like to briefly understand this ‘complex and adaptive’ aspect of financial markets to demonstrate the important role it should play in shaping our philosophical approach  to trading.

Now before we start, I need to stress that complexity science is a huge subject, so doing justice to the cause in this brief Primer is simply ‘the tip of the iceberg’.

Furthermore our ability as traders to adopt useful tools or statistical metrics arising from complexity science is a vain endeavor as the terms “complex and adaptive” mean that we are dealing with shifting goal posts all the time… There is no single metric that adequately describes it.

But the reason for our interest in examining complex systems is to demonstrate that powers of prediction within this arena have significant limitations. As a result we need to reign in our hubris and beat our swelling ‘all knowing’ ego into submission. It is in our interests as traders to become humble observers in the markets who ‘go with the flow’ as opposed to self-assured predictive thinkers that feel they know how to beat these markets.

The reality is that Financial Markets do not actually care what we think. It beats to the tune of its own drum and we are only privy to part of that beat.

Unfortunately, our brains believe that we can crack this riddle with useful workarounds and simplified predictive methodologies, yet by understanding how complex systems behave, our brains might want to think again. We need to think outside the ‘predictive box’ and in doing so we see that the standard toolbox we have been using to address these markets are a disaster waiting to happen. By understanding the limitations of statistical science, we become better traders as our appreciation of risk and uncertainty rises. Simply by appreciating that financial markets are complex systems is enough to make us better ‘risk managers’.

One of the central take-outs from this Primer is that being reductionists, humans like to be able to reduce complex problems into their constituent parts to better understand them, but in doing so, apply simplifying assumptions leading to conclusions that are ‘Almost Right’ but miss the fine print that actually make them ‘Not-Even Wrong’.

The complexity surrounding complex systems lies in their relational connectedness seen through their non-linear and emergent nature. They are scale independent and grow (typically fractally) from a seed into emergent form, where the emergent features themselves are not enduring in nature. These ‘ephemeral’ emergent features then act as scaffolds for future emergent form.

While offering a false allure of stability for indeterminate periods of time, the fine detail (frequently assigned by the statistician to rounding errors or having no material consequence) conspire to make complex systems continuously adapt over time. They cannot be faithfully described by any permanent description and are fundamentally non-stationery in nature. This essence of complex systems is not captured by conventional simple statistical treatment where we like to view systems as ‘being constructed’ from assembled bits and pieces of permanent and enduring nature. There is no scale in complex systems as they have no rulers or internal clocks. The method that complex systems grow is simply by non-linear fractal development using extremely simply rules. We do not necessarily see all these fractals, but rather see the emergent form arising from their rules-based construction.

You see, we find complex systems all around us wherever we look. In fact, you would be hard-pressed to find any ‘entity’ or ‘thing’ at all that is not actually a complex system at its heart.  A brain, an atom, an elephant, a ship, a rainforest, a cloud, a city, a financial market, are all examples of complex systems, so it is worthwhile digging in to understand the central features of these amazing domains.

Having this understanding allows us to see the forest for the trees and able to recognise that statistical treatment is rarely a precise tool. The errors we find between theory and reality are sufficient to overturn entire theories, so in understanding complex systems better, we can build our powers in skepticism which is a very useful trait for an ‘uncertain’ trader seeking an enduring edge. Being uncertain is a gift that leads to enquiry, whereas being certain is a ticking time bomb waiting to explode.

Understanding complex systems allows us to possibly remove the blinkers and see some of the heinous errors we have made through our simplistic rendition of human theoretical constructs such as:

• treating quantified risk measures as viable tools to assess uncertainty;
• where we infer system mechanics from probability sampling;
• where we assume system stability;
• where we adopt the baggage associated with the Law of Large Numbers to assume the nature of real distributions, and
• in our conventional risk metrics applied to risk management….to name a few.

The complexity surrounding ‘complex systems’ leaves simplistic predictive models tailored for a static domain in the dust. We need to recognise the significant limitations of standard statistical treatment in addressing them. In fact, having an understanding of the limitations of predictive modelling creates an ‘enduring edge’ for the wiser divergent trader accustomed to ‘uncertainty’. Our edge is the fallout that occurs when predictive models go wrong.

Complex Systems

Despite our awareness that almost everywhere we look, we observe a system at work, our understanding of complex systems has unfortunately been left wanting, partly due to the confounding intractability of interrogating complexity but largely attributed to the simplistic reductionist methods of enquiry we have traditionally used in attempting to understand them. It is this reductionist process that simply fails to account for the reasons for system complexity, namely the inter-relationships and dependencies created between constituent parts as opposed to the constituent parts themselves.

Common to all complex systems is a multiplicity of many parts in which there is an absence of a central control element, either internal or external. Rather it is the system itself that progressively becomes more self-sustaining and efficient through the exchange of internal resources via interactions between system participants, allowing for nested sub-systems and system structure to emerge that enhance system stability and robustness.

A key property that arises from complex systems is emergent structure and behaviour and a property of a system is ‘emergent’ only if we genuinely have a new feature that cannot be explained through a more detailed fundamental description of its constituent parts. This is where a new ‘thing’ in naive terms is constructed within a system context that can only be explained by the relationships that exist in the entire system as opposed to its constituent parts. A ship, an elephant, an artwork, a building, an organisation, a thought. The range of potential emergent properties of a system are endless.

There are never actually any things when you look hard enough….there are only processes. ‘Things’ arise from a limited biased mindset that is unaware of the deeper relationships that exist and their importance. Western Cultures like to use nouns to create permanence to ‘a thing’ as their approach to understanding things has been through reduction and simplification but if you dig deep enough into a complex system, there actually no things at all, only processes. In fact, you find that it is nested processes all the way down to create processes within processes…within processes…… all the way down the rabbit hole. Systems thrive from their complexity. Simplification fails to understand their devilish character.

Also common to all complex systems is that the re-use of system elements for multi-functional purposes allows non-linear power laws to come into effect where relationships between things evolve from a one-to-one to a one-to many relationship. The result allows for progressive system efficiency and durability making the composite more resilient to perturbations…….however due to the strong coupling that arises between system elements, a failure in one or more relationships can lead to cascading failures making the entire system progressively more vulnerable to catastrophic risk. The reason for the italics….is because this is what excites the price follower in these financial markets. These rare moments of decoupling when markets decide to transition.

Price Followers have a strong respect for complex systems and recognise the inherent levels of uncertainty that exist in them given their complex nature. They never assume they know everything but rather go with the flow of it all and do not try to fight it. Predictors however like to be certain about certainty….but price followers are very uncertain about certainty…….the two philosophies lie on either end of a spectrum from certain to uncertain about how a complex system behaves. The difference in opinion therefore shapes very different trading techniques in tackling any edge present in these markets.

Financial Markets

Markets are complex systems and we as participants are only privy to part of that complexity. We need to look from inside out when trading on the right edge as opposed to Gods who look from outside in, or from a perspective of  hind-sight.

You see, when you look from inside out as opposed to outside in, you immediately understand that you are just a curious emergent form residing in a domain of knowledge that always expands. The ability of complex systems to use their emergent structure to act as scaffolds for future form mean that our quest to understand them is simply never-ending. Once we understand an emergent feature which we describe within a particular domain, we then find that this domain operates within a larger domain of enquiry….and so on deeper down the rabbit hole we go. This is symptomatic of an embedded observer seeking to understand the system they reside in. It is a continuous quest of knowledge in an uncertain domain whose knowledge frontier expands.

This frontier between our knowledge of the market state and the uncertainty of the state that resides outside our bounds of competence is what distinguishes risk versus uncertainty. Our assumed knowledge regarding the state of the domain we understand leads us to many false conclusions and makes us blind to ‘fat tailed events’. Only in hindsight do we understand that the domain we trusted was just too narrow in extent. Unfortunately, as humans limited to our knowledge, we assign certainty to the known knows and wrap this certainty up with financial metrics that define ‘risk’.

So given our penchant for ‘certainty and predictability’ we like to encapsulate our risk metrics under Gaussian assumptions that theoretically take priority over the reality of ‘complex systems’ and use metrics such as the Sharpe ratio or standard deviation as a basis to assess trading risk. All the time, the divergent price follower who simply observes how complex systems actually behave shout “NO!, This is the wrong way to treat these complex systems. They deserve more respect”.

What we fail to address however is the uncertainty that lies outside our knowledge bubble that sends us to the graveyard early due to the inevitable inter-dependencies that exist between our limited domain of knowledge and the relationships that extend into this uncertain ‘uncharted domain’.

The multiplicity of agents in a financial system which is characteristic of complex systems, are the system participants. These include bankers, brokers, traders, hedgers, speculators, and gamblers. The way they interact create system complexity and the ‘gross expression’ of this myriad of interactions is price action. These participants have different opinions, but they transfer that opinion into the market through their entry and exit. Now the thoughts that shroud these opinions extend well beyond the financial markets themselves into a broader domain called ‘life’ and this is where the uncertainty residing in the left and right tails of the distribution of market returns germinates. The real reason for why the hedge fund decided to offload its risk may forever be shrouded by ‘uncertainty’.

Markets effectively represent huge computers that efficiently compile this opinion into a gross statement represented by price. Now as efficient as the market is in processing opinion represented by trading behaviour, we have seen in preceding Primers that arbitrage opportunities reside in this price action.

The innumerable transaction events of different scales (of entries and exits) arising from participant behaviour consolidate into either random features with no collective impact on overall price action (aka noise) or non-random emergent features whose collective impact place a bias on overall form. Under this interpretation we treat price action as an emergent signature arising from collective participant trading behaviour.

We can only interpret the very broad characteristics of this complex interacting behaviour through our reliance on gross statistical measures arising from price behaviour…just the same as we can interpret the gross features of a cloud by its temperature and pressure arising from the mechanical behaviour of its composite water molecules.  By summating these interactions, the overall behaviour of the system itself can be summarized through the mechanics of price action and this behaviour  can either be abrupt and fleeting in in nature or sustained over a period of time.

Now having an understanding that price action is representative of an emergent form arising from participant interactions, we can use metaphors that might help us interpret these systems better. Unfortunately, the complexity surrounding ‘complex systems’ renders traditional statistical treatment hopelessly deficient so using metaphors is often a very useful way to interpret these vast riddles.

It is helpful to consider price action as an emergent statement arising from the myriad behaviours of participants who exert an influence in overall form. Under such an analogy, we could group participants by common behaviours and consider the mix of participant grouping and their overall influence on price action. Recognising that participants change over time; we can there clearly see how the mechanics of price action is never stationery. It really is a moving feast. The ever-changing dynamics of participant mix insist that we cannot predict the future market state with fidelity.

There is no certainty in our financial markets. There are simply periods of stability in an otherwise uncertain state. It is uncertainty that is ubiquitous in a complex system and it is certainty that should be viewed as the anomaly. While we as price followers nod our heads to the statisticians and treat a trend as an anomaly, we secretly ‘flip the bird’ to the statisticians out there with our understanding where the anomaly really resides.

Expand the Mind to Experience Wider Domains

Before we leave this brief introduction to Complex Systems, I will leave you with a few examples that will tease the mind and help to shed new light on the way that you view complex systems. Unfortunately, the study of complex systems is a huge topic and cannot be conveyed in a single Primer but at least these examples may get the reader to appreciate what complexity science is about and why us trend followers pay lots of attention to observing processes at work.

Example 1: Metronome Synchronisation

In Example 1 we undertake an experiment where we observe the behaviour of 32 metronomes supposedly operating independently to each other in a random chaotic manner. This is of course an assumption from simplifying the experiment through reductionist logic. What we actually find is that there is a slight bias imparted by the table itself that synchronises the metronomes to create emergent order throughout the entire domain. The entire setup needs to be viewed as one complex system.

It is the processes that matter. Not the things. Look for the relationships between things to identify processes at work. Sometimes the market will be noisy…….and sometimes the market will be coordinated. Don’t predict it, just follow it. The table with the metronomes holds the key to the connection between them. Observe the entire experimental setup, as without it, you will start trying to make sense of the things and lose sight of the process. Just having this understanding empowers you to see opportunity and risk and then simply make better decisions.

The reason that this toy seems mystical is the way you have approached the problem. A table, a metronome, many metronomes. All “nouns” of convenient human categorisation that take form in our language and then become permanent ‘objects’ of your thoughts leaving us blind to the underlying process .  All based on a limited understanding of a system forgetting the importance of the relationships between different emergent aspects of the system.  You need to approach this problem using verbs to find the processes at work and eradicate those nouns of perceived bias.

Example 2: Not random but rather Deterministically Complex

In classical systems there is order in every system state. Financial markets are not quantum systems and they obey classical rules. There is no ‘actual’ fundamental randomness in classical systems, there is just emergent complexity. We assign terms such as randomness or non-randomness to describe complex systems, but they are not the reality, only human constructs that allow us to understand them. Example 2 highlights this principle.

The physicist David Bohm coined the term the ‘Explicate versus the Implicate order’ to wrestle with this notion. What changes is the nature of the state. In complexity lies deep simplicity. Very simple things that are configured in different arrangements.

The grander system state moves to different states through gradual transition but the impact of this transition is not linear throughout. What makes the fat tails is that you are only observing a small sub section of the entire state.

Observation of the processes is the key as opposed to small sample prediction of the things. If you first observe the greater system then you can choose the optimal path by going with the ‘flow of it all’.

Methods of prediction without knowledge of the entire processes at work can be disruptive to the system state. When you think you know it all from your sample statistics, then backtrack and think again as markets are adaptive and constantly changing. Do not ever deal in certainties.

Knowledge is an adaptive process that allows you to ‘keep up’, not a cup that is either empty or full. There is great value in using statistics. The problem however lies in its naive application.

Example 3: Prediction versus Uncertainty

The two pendulums below are used as a metaphor to describe convergent behaviour and divergent behaviour seen within complex adaptive systems.

Convergent trading systems (as exemplified by the single pendulum with a single mean) are based on the principle that price will converge to an estimated value whether that be an estimate of intrinsic value, or an estimated historic mean. They are therefore predictive and linear in nature given this assumptive stance and are couched in terms of assumed stationery condition.

Divergent trading systems (as exemplified by the dual pendulum with multiple means) are based on the principle that price will diverge away from an estimated value. They are non-predictive and non-linear in nature and relate to market transitions as opposed to periods of market stability.

Markets like any complex system tend to exhibit quasi stability (false equilibria) punctuated by unpredictable periods of non-stationarity. They therefore exhibit two types of arbitrage opportunity. That arbitrage associated with predictable stationery conditions (eg. convergent trading styles) and that arbitrage associated with unpredictable moves away from stationarity (eg. divergent trading styles).

A single pendulum exhibits predictable oscillatory behaviour suited to convergent trading styles such as mean reversion or value investing and the double pendulum which exhibits rich dynamic behaviour exhibits divergent behaviour suited to price following.

Chapter 5: The Search for Sustainable Trading Models

In our Primer so far, we have touched on some of the central arguments for why we need to pay respect to these amazing complex systems we call financial markets and adopt a very humble stance, as opposed to a controlling stance, in how we interact with them.

We are active participants in that system and not Gods looking from outside-in, who have no role in influencing outcomes. We actually shape the outcomes.

Each of us have our own knowledge cone that extends into the market domain and base our actions on what we interpret to be the reality. This knowledge cone is always limited in extent but it defines what we infer is ‘the reality’. Based on our limited access to this reality we impart our knowledge into the market and slightly expand this domain further.

From a trend followers viewpoint, we are philosophical relativists who are sceptical in assigning any form of objectivity to the ‘things’ we observe. We suspect there is more to it than the state of our knowledge.

We believe that our understanding is a two-way relationship between the observer and the observed and the extent of the domain (or context of this relationship) is what appears to us as cause-and-effect. This however is just a single perspective or symmetry of a much wider viewpoint taken by every participant. The way we interact with the system shapes our perceptions of it.

This tendency we like to have as humans to assign ‘objectivity’ to this narrow ‘knowledge’ domain makes us ignore or suppress the small things we deem immaterial whose relationships may extend out into a wider possible domain which in a wider nested emergent context, may have pivotal consequences to the illusory domain we think we understand. This can lead to catastrophic failure in our ‘predictive conclusions’.

While trend followers adopt a similar approach to the Efficient Market Hypothesis in the way we interpret information being ‘injected into an all-knowing market’. We drop the notion of ‘rational participants’ and replace that notion with ‘limited participants’ whose knowledge comprises simply what is ‘known’ about the system we interact in. Our limited knowledge results in us taking action which may only be a half truth about the state of the ultimate reality.  This has dramatic consequences to the conclusions made by the EMH. It means that there is arbitrage associated with the ‘error’ that exists in the information we inject into the market through our trade actions. This is where the edge resides for trend followers in this zero-sum game. The edge arising from simply adopting a rules-based process that ‘over-rides our desire to predict’.

This illusory reality we are forced to trade within therefore makes us very uncertain traders.  We assign our understanding of things to simply a limited viewpoint surrounding our propensity to like to classify and objectify stuff. We understand the reality of a complex system is far deeper than the constituent parts and relate to the way these apparent ‘things’ (which are really just emergent processes) are deeply connected.

How we Approach Risk

We adopt the viewpoint that conventional risk treatment and assessment adopted by industry is severely limited and does not truly express the total path of possible market states in which risk is known to lurk. As a result, we have the opinion that the subject of risk needs to be broken down into the risk that can be quantified through our ‘state of knowledge’ and the risk that cannot be quantified and relates to that broader domain for which our current state of knowledge is yet to expand into.….called uncertainty.

I remember when I was cutting my teeth as a ‘Fund Compliance Manager’ in the 1990’s and was presented with a risk matrix used to assign risk in quantifiable terms to those Investment Managers (IMs) seeking solace under our ‘Responsible Entity’ structure in Australia. The matrix was a template that encouraged the assessment of IM risk in quantifiable terms by weighting the ‘likelihood and consequence’ of risk events derived from reviewing their processes. It used to make me laugh. Was I a ‘God’ that had this amazing foresight and ability to quantify the unknown? Unfortunately, I was bound to the rules of Regulatory Compliance and was forced to complete the matrix….and while at least the process forced me to ‘consider’ possible risks it more frighteningly made me acutely aware of the industries inability to measure risk. I could not help but feel that lurking within this process was a ‘failure event’ waiting to happen.

By viewing risk in limited ‘quantifiable terms’ using proxy metrics such as Risk Matrices and the Sharpe ratio, or indeed any risk metric that is riddled frequently with Gaussian bias, trend followers feel that we are grossly understating the possible risk that lurks in a trading technique.

We do not assign blame to this viewpoint but rather treat it is an overly simplistic interpretation of a wider reality which is a natural consequence in the way we use a limited brain to understand complex systems,

We are swayed by the powerful inductive reasoning of mathematicians and philosophers who challenge conventional reasoning such as Mandelbrot and Taleb whose words deal in non-linear terms.

We need to pay heed to the risks of uncertainty and not simply sweep them under the carpet as being, too hard to fathom. These events are the real windfalls or catastrophic failures we see in the history of complex systems. These non-predictable punctuated events of a material nature, are the real stories of the history of complex systems. These fat tailed events change the emergent structure of the complex system permanently and are the major transitions we either need to worry about or have fantasies about. They are the real game changers that quickly decide who the losers and the winners are in this game.

As trend followers looking for the  ‘big prizes’, we believe that the change associated with ‘predictable’ events arising from an interpretation of what is ‘known’, are small pennies when compared to the possible massive change associated with events attributed to the unknown (or the fat tailed event). The small change arising from normal day to day price variation are an inevitable perturbation that we as ‘risk managers’ already have a firm grip on. We need to embrace this change as this is an inevitable consequence arising from striving to make opportunities in that domain of the unknown where big changes occur and where our fortunes await. If we worried about that inevitable volatility that resides in the ‘known’ and sought perfectly linear equity curves for each return stream in our portfolio, then we would never leave ourselves open to the massive windfall that can arise outside this known zone of perturbation in the unknown domain. That zone where the outliers reside.

In fact, our whole modus operandi is couched in terms of trading the risk events associated with uncertainty and avoiding those risk events associated within this narrow domain we ‘believe that we know’. We embrace the ‘volatility of the known’ and lie on the verge of the Gaussian envelope waiting to pounce on opportunities in the unknown domain.

We view the success of trading strategies in terms of the way they have responded to a myriad of large and small risk events and simply let profits arise from letting financial markets do what they do. We disregard the illusory stability of nice ascending linear equity curves and the future promises of high Sharpe ratios. You see, we believe that such narrow-minded treatment of risk is the fundamental issue that ‘biases’ the way we as humans select our preferred trading strategies or allocate capital to portfolios.

It’s All About Survival (Robustness)

We prefer to think that we, as humans, actually have it ‘mostly’ wrong so in evaluating the Managers who are less wrong than others, we use the long-term track record as our basis of distinction. There is no better way to assess the risk of a trading strategy than through the long-term track record itself. Being a survivor is the key metric we use as a basis to assess the relative strengths of a trader. Being a survivor is also a legitimate way to assess how a trader has been able to adapt to changing market conditions which are an inherent feature of any complex system.

‘Adapt or die’ you might hear being used by traders in these markets. Well surviving is all about the ability to manage known and unknown risk events and still be alive to tell the tale and we feel it is the only real method to assess the sustainability of a trading strategy. We do not need any fancy risk metric that is used as a proxy to assess the ‘riskiness’ of a trading technique and only assesses risk in Gaussian terms. We prefer to use the performance track record itself as our most valid risk management metric or in its absence, the long term back-test. It is the entire track record itself that exposes risk weaknesses that are not necessarily ‘seen’ by simple risk metrics. We pay more attention to ‘maximum single points of weakness’ such as the maximum drawdown than single holistic statistical risk measures such as the Sharpe ratio, Sortino ration etc. which give us a more subdued version of the possible risk weaknesses that lie in a return stream or a portfolio.

This track record and the long term backtest provides a ‘warts and all’ story to assess how a manager has fared against the known knowns, the known unknowns and the unknown unknowns. Thank you Mr Rumsfeld.

The secret to a long trading career is based on a notion of ‘survivability’. While many traders can experience periods of hedonistic bliss over a short time duration when markets are behaving according to their trading plan, no trader can prosper using a single technique across all possible market conditions.

There are inevitably those times when markets simply do not behave according to your plan and your success as a trader is defined by how you protect your hard-earned capital during these unfavourable market conditions.

Being a market survivor means that you can survive the unfavorable but are always able to participate in the market when conditions become favorable. The surest way to kill a trader’s prospects of generating wealth from the markets is to prevent them from being a participator. So, a key goal for any trader who enters this game is in learning how to manage risk and protect your finite trading capital. There are simply too many ways for you to lose your capital in these highly efficient, competitive, and non-stationery financial markets.

It is this principle of survivability, also referred to as robustness,  that allows you to stay in the game for the long haul and be able to capitalise on those opportunities that are favourable to your trading plan.

Markets are prone to change their nature and profiteering is simply a small part of the bigger problem when markets change their state. The true mastery of the markets lies in the way we manage all possible risks.

Now many readers might infer we are ‘scaredy cats’ who jump at shadows and should be concerned with everyday risk and not worry about ‘all possible’ risks as they are impossible to quantify, but we have a novel way of dealing with ‘all risks’ that ensures we are never left straggling in the adverse tail of the distribution of market returns. We always cut losses short. This simple technique is not a method to quantify risk, but rather a process that ensures we are never blind-sided by it.

Of course, the reality is that in our application we apply initial stops and trailing stops to all our divergent trading strategies and only apply small risk bets to each solution in our diversified portfolios, but this is not for the reason that most traders think. Such measures do not guarantee protection when the market gets angry and simply ignores these paltry tactics. During major tail events, liquidity becomes a major issue and frequently we find that our risk mitigation methods are simply ignored.

The actual reason we apply these measures is to ensure that at the global portfolio level, the warehoused risk that lies in every return stream is ‘released’. This means that no return stream is likely to bring down the whole portfolio due to some unforeseen risk event. By continuously ‘releasing risk’ from the portfolio through trade exits we don’t hang onto to risk. We let it go. This means that our portfolio is always ‘optimally configured for ‘Future risk events’ and is not holding onto risk from the past. This is what gives us robustness and allows us to fight the uncertain future with confidence. It is not the risk of the past that should worry us, it is the uncertainty of the future. We discuss this in a later Primer in terms of a Portfolios ‘Responsiveness to Change’ which is a key element of what it means to have a robust portfolio.

Having risk mitigation mechanisms does not guarantee that we can address all risk, but it certainly ensures that we are more robust than alternative methods that disregard these mechanisms. This is what survivability is all about.

Under uncertainty, it is not the profits we need to be concerned with as they are a natural consequence arising from how the markets behave. Rather, being a survivor is always about being able to participate in these ‘profitable events’. To achieve that, we need to be around to participate in them.

To have a long term track record is the way to choose between the ‘scams’ and the legitimate prospects. There are many ways that we as humans can conceal the underlying risk that lurks in all trading strategies.

We deploy devilish methods to conceal risk such as the use of averaging down into losers, avoiding mark to market valuations, showcasing equity curves during only favourable market conditions, providing curve fit backtests that respond to a single market condition and applying no protective risk stops or trailing stops to trades. All used to provide a false sense of security to investors or indeed ourselves that capital is being prudently managed and that we are ‘good traders’.

There are a myriad of clever ways to obfuscate the inherent risk that lies within all trading strategies and we refer to these techniques as methods to ‘warehouse risk’.

Unfortunately, you do not hear this term bandied about industry that much, as it implies that there is something about risk that we do not know about. Indeed, there is. There is always the risk of uncertainty.

Who are the long-term Survivors that we can learn from?

To save yourself from the classic trap that meets many novice traders, just like entering any profession, it is very wise to undertake extensive due diligence and learn from the successful traders with a validated long term track record. These are the ones who are around to tell the tales of the events surrounding past storms of uncertainty.

By undertaking this trading game from the start by first learning from the top traders in this game, you can save yourselves many years of wasted time and lost opportunity. Having an appreciation of the sustainable returns that are enjoyed by professional fund managers with a long-term track record will not only allow you to jump onto the shoulders of the giants in the industry but it will also  allow you to establish more realistic benchmarks that you can work within with your own trading endeavours.

It is very unlikely that a retail trader will outperform these top ranked long-standing funds in the world, so it provides a useful benchmark to establish ‘realistic’ expectations that a trader can work under without over-extending yourself and making yourself vulnerable to catastrophic risk of ruin.

For the vast majority of retail traders, most are very unlikely to survive more than a few years and even for those that manage to escape the lessons of the market with a track record greater than a few years, there is no guarantee that it was not luck that allowed them to survive the day. You see there is so much room for luck in an efficient market that it might take many years for a trader to realise that they never actually possessed any ‘real trading edge’ in the first place.

An examination of the successful Fund Managers with a long term track record allows us to identify trading methods that have a better chance of standing the test of time and provide much needed clues for a retail trader seeking a sustainable track record in trading these financial markets.

Now an examination of the literature is a bit deceptive as I simply could not find a listing of ‘the best long term hedge fund managers of the world’. There were a few listings of some solid performers over some time horizons, but a listing of the best long-term performers was peculiarly absent. I suspect that the reason for this absence in the literature lies partly in the ‘secrecy’ of these cherished institutions but also suspect that there are actually only a very small number of FM’s that can boast such a long-term track record.

There is no point considering those FM’s that simply seek to outperform an Index and do not focus on ‘absolute returns’. Evaluating these FM’s does not bring you any closer in understanding what survivability is about. It simply kicks the can down the road as you then need to compile these ‘Index Performing’ funds into a portfolio to see how it might be possible to generate ‘absolute returns’ from this amalgam.

I did find a useful Appendix in Greg Zuckerman’s excellent book “The Man who Solved the Markets” that has made a bold attempt of creating a listing of the best of the best in terms of absolute returns, but the information was limited.

Table 1: Absolute Return Comparison of some of the Most Successful Hedge Fund Managers in the World (after fees)

Source: Gregory Zuckerman: “The Man Who Solved the Market – How Jim Simons Launched the Quant Revolution” – Appendix 2

Now as limited as the information of Table 1 is, it should at least make you question the % returns that are bandied about in the retail world.

These top performers are not simply one-man bands. They have entire legions of quants, mathematicians, programmers, and physicists in their composition. Are we, as retail traders, seriously considering that we might out-perform them? Time to see your local psychologist if you think that is the case.

If the best in the world offers a net return (or Compound Annual Growth Rate) of 39.1% over a 30 year plus track record, then this should at least create the upper level of realistic expectations within which we as Retail Traders can work within. While chants of 100% per month or year can be heard in the retail trading forums of the world,  this may be possible, as nothing is impossible,  but reality makes us conclude that even if such returns are achieved, luck with excessive leverage forms the basis of them. They simply are not sustainable metrics and are more likely to crash and burn when conditions become unfavourable.

As traders rather than gamblers, some of us want a sustainable career for life. If that is your desire, then pay attention to the industry benchmarks, learn from the best in the world and work within realistic expectations.

Now I would like to go farther in my due diligence of the best FM’s in the world and in particular look at aspects of risk as opposed to mere performance returns. Even though this list boasts FM’s with a long-term track record, there is still a chance for chance and over-leverage to have a say in the matter.

What I would like to find is not simply a handful of ‘the best players’ who use different approaches to target alpha in the market, but rather a whole class of very successful long term traders who apply a similar trading philosophy and clearly can be seen to manage risk within my risk tolerances.

I found an article on my due diligence search by Daniels Trading in April 2010 that really caught my attention. It was titled “Beating Warren Buffett – Can your Investment Manager Beat Warren Buffett’s Berkshire Hathaway’s Stock Performance over the past 10 years?”

Now everybody knows Warren Buffett, and I thought to myself, what a great question? There is only one Warren Buffett, but imagine if we could find an entire class of FM’s that produce comparative performance who I could mimic in my trading endeavors. Isn’t that an easier path, jumping on the shoulders of giants, than trying to invest the wheel?

As I read on in this article I wasn’t disappointed. Table 2 below was presented in that research.

Table 2: Comparative Performance between Berkshire Hathaway versus Trend Following CTA’s

Have a close look at Table 2, You will not only see how Buffett over the 10 year period has been outclassed by a large number of CTA’s, but that these CTAs are trend following firms that apply traditional trend following methods that are the subject of this Primer series.

But there’s more. Have a look at the risk metrics that are included in Table 3. Not only were performance returns exceeded, but the drawdowns incurred to achieve those returns were far lower on average by the CTA’s.

Table 3: Comparative Performance and Risk Metrics between Berkshire Hathaway versus Trend Following CTA’s

Let’s now update this paper and undertake our own assessment since 2000 to the current day, using available data from Nilsson Hedge and Yahoo Finance to see if this performance track record continues to today.

Table 4: Comparative Performance Metrics between Buffett and the Trend Following CTAs 1^st Jan 2000 to 31^st March 2021

Chart 14: Comparative Equity Curves between Buffett and the Trend Following CTAs 1^st Jan 2000 to 31^st March 2021

Table 4 and Chart 14 continue to demonstrate that a large majority of the top performing trend following CTAs are comparable to and even outperform Berkshire Hathaway in terms of CAGR and more importantly continue to offer far better risk adjusted metrics using the MAR ratio. This ratio is a reflection of the annualised return versus the maximum drawdown experienced by the Fund over this time horizon. You can see from Table 4 that Berkshire Hathaway had a maximum drawdown of 44.5% over this time horizon whereas all CTA’s produced comparative returns with far  lower drawdowns.

But you want more don’t you? After all, there are a whole host of well-established long-term trend following firms to choose from that deliver superior risk adjusted returns to their clients without significant ulcers. How about some of our favourites that you may find on Twitter offering their sage advice such as remnants of the league of surviving “Turtle Traders” such as Jerry Parker’s Chesapeake Capital and also other Twitter favorites such as Niels Kaastrup-Laarsen of Dunn Capital fame. It really pays to listen to them diligently. They speak from years and years of experience surviving these markets. Unlike the more secretive ‘black box hedge funds’, these guys are more than willing to share their sage advice. Isn’t it just logical as aspiring retail traders that we should want to learn from them?

Table 5: Comparative Performance Metrics between Buffett and the Trend Following Favourites 1^st Jan 2000 to 31^st March 2021

Chart 15: Comparative Equity Curves between Buffett and the Trend Following Favourites 1^st Jan 2000 to 31^st March 2021

It tells us that we do not need to be rocket scientists or have a legion of mathematicians and physicists at our disposal to build wealth building returns. We just need a diversified systematic trend following approach to trading these markets and enduring patience in this robust method.

We do not need to reinvent the wheel at all in our trading approach. We can learn from the best and simply mimic their techniques….but we do need realistic expectations and a firm grip on what it takes to be a survivor in these fickle markets.

When it comes to wealth building over the long-term, a long-term track record of performance success is perhaps your most reliable guide for an uncertain future. No-one knows what the future will bring…but at least we know from the validated track records of these long-standing managers that they are no strangers to taking uncertainty by the horns and delivering absolute investment returns for their clients.

Chapter 6: The Need for an Enduring Edge

In our previous Chapter we pointed out how we, as trend followers, are very uncertain about the degree of confidence placed by the current state of the industry in ‘understanding these financial markets’. This makes us very sceptical traders who view all traditional methods of interpreting them as possibly suspect. Our reticence to trust traditional doctrine makes it appear that we have no guiding quantitative tools to work with in navigating these markets as we are unlikely to trust any of them, however that is not the case. We don’t need these elaborate tools that seek to simplify or act as proxies for interpreting the reality. We simply need to apply critical thinking and powers of observation to develop our own tailored tools that we use in that dreaded zone that lies in the tails of the market distribution.

Nearly all trading metrics and tools you see are sculpted with ‘predictability’ in mind and are targeted towards the peak of the distribution of returns. Trend Followers however are targeting the tails of the distribution and require a totally different set of metrics and tools to deal with them.

To demonstrate that trend followers can perform or even outperform alternative trading philosophies, we then took you on a brief introduction to some of the best performing funds of the world to demonstrate that trend following is perhaps one of the most effective processes we can apply to trading these markets without requiring us to invest in an army of quants, mathematicians and physicists to achieve our wealth ambitions.

We then discussed how trend followers use the term ‘robustness’ seen from the ‘warts and all story’ of an entire trading record as a basis to assess whether or not their systems have an edge. Once again we prefer the entire story provided by a back-test or validated track record that spans across a broad array of different market regimes before we award that system with the phrase ‘to possess a demonstrated ‘edge’. We prefer to understand this entire story of performance as opposed to using more traditional ‘shortcut’ measures such as the term Positive Expectancy which in their summarised efficiency of a single result, frequently require the use of assumptions that make their application suspect.

But nevertheless ‘an edge’ is an essential term that all speculators (as opposed to gamblers) need to digest, as without it, you are left floating down a river ‘without a paddle’.

A system with an edge is defined in the literature as a system with overall positive expectancy over its trade history…..and the greater the trade sample size, the greater confidence we have that the system displays an actual edge.

We use a proxy called ‘Expectancy’ to quantify this edge which is defined by how much money, on average, we can expect to make for every dollar we risk and can be calculated using the Expectancy formula E = (Pwin x Avg$ Win) – (PLoss x Avg$ Loss) over the trade history of that system.

• E = Expectancy or Expected Return
• Pwin = Probability of a win
• Avg$ Win = Average win in dollars over the trade history
• PLoss = Probability of a loss (or 1-Pwin)
• Avg$ Loss = Average Loss in dollars over the trade history.

But be sceptical as soon as you see a static formula applied to adaptive markets. The formula above leaves the impression that Expectancy can be used as a basis to compare different systems, however the devil is always in the details.

As soon as we apply a probability in a formula to achieve a ‘single summarised result’, we immediately turn a non-stationery background environment into a stationery one. We have a single value to represent an entire trade history. Given that markets are adaptive, then trading performance, being a derivative also needs to be variable. But a single statistic implies constancy across the time series. Many robust systems experience lower expectancy during unfavourable conditions and higher expectancy during favourable conditions, so a single statistic is not going to tell us this information and is therefore not a very useful statement about the performance of the  overall system or portfolio. Treat expectancy as a general guide to assess an edge but do not treat it in absolute terms.

Furthermore, no mention is made in the formula above about the sample size required to make this assessment. Can an edge be calculated on 10 trades or 10,000 trades, and what does this infer? Is an edge calculated on 10,000 trades in a high frequency trading environment that might span a few weeks, the same as the edge calculated on 10,000 trades in a long term trend trading portfolio over 20 years?

You see, by inferring that the statistical formula has meaning, we automatically include associated biases of reasoning into it.

Ideally what we want to infer in our assessment of a system with an edge is whether there is a profitable bias in the trade results of a system that is beyond what a random distribution of trades would imply. A large sample size of trades is therefore preferred to a small sample size of trades in our use of the expectancy formula….but perhaps more important than sample size are the number of various market conditions that the strategy has performed under.

While sample size is a useful proxy, and large sample sizes are more likely to have encountered a broader range of different market conditions, perhaps an even better measure of the term ‘edge’ is the relationship between system profitability which includes the number of different market conditions encountered by the system. While this would be nice information to have, our ability to derive such a formula is problematic due to market complexity. So we are forced to use generalizations.

We can of course use the expectancy formula as an idealized generalization which has some degree of merit in application to assess whether a strategy is profitable but watch out for the assumptions embedded within it.

Now many retail traders are under the misapprehension that an edge relates to the Probability of a Win (Pwin%) alone and therefore seek systems that can generate winning percentages that exceed 50%, however it is quite easy to guarantee this outcome by adopting profit targets that are half the size of the Initial Stop loss or by developing systems that deploy averaging down techniques or Martingale progressions that hold onto losers and never realise the loss until the total loss is recovered or in many cases the account blows up. Such manipulative tricks are commonly applied by charlatans that prey on the retail community to make it appear that they have found the ‘Holy Grail’ of trading systems that never incurs a loss.

While an illusion ‘of safety’ is created with a strategy that offers a 90% win rate, such illusions only offer psychological comfort. Warehoused risk lies in these solutions as it does in any trading strategy we generate. There is no way that we can avoid the risk inherent in a trading system through clever tricks. If we want to strive for higher returns, then we need to accommodate greater risk in our strategies to achieve it. What we will find in a later Primer is that trend followers use diversification of non-correlated and co-integrated return streams as a far more powerful method to achieve this ambition.

What is forgotten by many Retail traders seeking a high win rate is that Expectancy needs to include both the win rate and the reward to risk relationship for each trade. That is why the formula includes measures of the win rate and also the average win and the average loss.

For trend followers that are more often wrong in their trade decision, the Pwin% is usually in the order of between 20%-40%. Where we turn this game into one of positive expectancy is through our reward to risk relationship. By always cutting losses short and always letting profits run our average wins far exceed our average losses by multiples.

For example, let us apply the following results based on a trade history sample to our Positive Expectancy equation.

• Pwin% = 25%
• Avge Win$ = $3,250.00
• Avge Loss $ = $500.00

Therefore:

E = (0.25 x $3,250) – (0.75 x $500) = $437.50

As we can see from the example above, despite the poor Win% of only 25%, the system produces a positive expectancy of $437.50 per trade over the data history. If we have a large trade history (sample size) then we can be confident that our system displays a real demonstrable edge, subject of course to the limitations of  the equations assumptions.

But just having an edge does not guarantee a winning result. An edge plays out over an extended trade sample size. In every trade undertaken, there is a degree of luck incorporated into that trade event and as seen in our previous Primers dealing with random chance, luck alone may have a significant say in a small sample of trades undertaken.

A real edge as the well-known Australian trend trader ‘Nick Radge’ states only shows up over the next few thousand trades.

For example, chart 16 below demonstrates a comparison between a random set of equity curves and an equity curve (in red) from a trend following system with a real trading edge.

Chart 16: Thirty Random Equity Curves Including a Trend Following System with an edge (in red) -500 trades

On close examination of the chart above, you simply cannot detect that the trend following system has any defined edge over the first 500 trades. In fact, there are many random equity curves that have outperformed this trend trading system over a small trade sample size. Starting out with $1,000 over the first 500 trades the system with an edge managed to barely break-even.

However, let’s see how this slight edge plays out over 10,000 trades. You might say….’C,mon, are you seriously saying we need 10,000 trades under our belt before we can conclude that a system has an edge?”. Well we have a neat trick up our sleeves that significantly increases our number of trades undertaken to more quickly assess whether a system has an edge or not. It is called a diversified portfolio. Under a systematic diversified portfolio which comprises hundreds or thousands or discreet trading systems, your trade frequency quickly reaches levels well beyond what a simple discretionary trader can manage.

Chart 17: Thirty Random Equity Curves Including a Trend Following System with an edge (in red) -10,000 trades

Now if you closely look at the chart above, you will be hard pressed to see any random equity curves at all. They have all deteriorated to a very small equity balance by about the 5000th trade. However, now look at the performance of the trend following system with an edge over the 10,000 trades. It is now exceedingly obvious that something was driving this performance apart from lady luck.

Starting out with $1,000 we can now see how the next 10,000 trades with compounding have magnified the result to $350,000.

This is what we mean when we refer to trend following as a wealth builder as opposed to a cashflow generator. There is simply no way that we can guarantee that our weak edge will be reflected in short term performance. Random chance has a significant say in the matter over the short term. We can however be far more confident in the long term that there will be a sufficient trade sample to allow the edge to play out and benefit from compounding.

Of course, to allow that to happen we need to strictly manage risk along the way. The large effect of luck in the short term can significantly impede our wealth pursuits if it is not strictly managed.

This is how an edge plays out. It can only be demonstrated with a large trade sample size. Therefore, you need to be careful when you hear people refer to a trading system having an edge. You need a large sample of trades to back up this statement.

To see an edge play out, you really need to focus on the next few thousand trades. A single trade result is inconsequential to systems with a real edge. Luck plays a very large role in the short-term These financial markets are truly efficient most of the time…however you only need a small enduring edge to take advantage of compounding and the ultimate wealth it can generate.

Now this leads us to a problem when assessing the edge in either convergent or divergent systems.

Under a convergent methodology that is predictive in nature and responds to a ‘current’ market condition in the now by attacking the edge that exists in the peak of the market distribution, we need to see the edge immediately imparted in the trade results. We therefore do not want to see our equity curve immediately deteriorate when we move from the In Sample environment where we developed our trading models to the Out of Sample environment where we apply these models in a live market environment.

However, for a divergent methodology that is non-predictive in nature and target the edge in the tails of the distribution, we need to allow for enduring patience before we can assume that our systems are not performing as per their design. Having designed our systems within the In Sample environment, it may take thousands of trades before we can infer anything about their performance efficacy.

The Need for an Adaptive Edge

For the vast majority of retail traders, most are very unlikely to survive more than a few years and even for those that manage to escape the lessons of the market with a track record greater than a few years, there is no guarantee that it was not luck alone that allowed them to survive the day. You see there is so much room for luck in an efficient market that it might take many years for a trader to realise that they never possessed any real trading edge in the first place.

Now given that liquid financial markets are non-stationery in nature, your edge also needs to adapt over time. The application of a simple trading strategy with no changes over an extended timeframe of say 10 years or so duration is unlikely to hold its edge. So, you must have a process embedded within your trading plan that allows you to adapt your technique over time to respond to changing market conditions.

Many readers will be familiar with the Turtle trading experiment of two successful professional traders Richard Dennis and Bill Eckhardt. Richard and Bill launched the ‘Turtle Traders Experiment’ where they advertised for traders in the Wall Street journal and inducted their trainees in a Trend Following breakout recipe called the ‘Turtle Strategy’. The strategy was wildly successful in the 1980’s generating a fortune for the group of traders. From the success of the program many of the Trainees then later started their own trend following funds and have been successful to this day….however the recipe has changed substantively.

What worked in the 1980’s no longer works in its prescribed form today. While the overall breakout technique is still deployed today, the parameter variables have needed to respond to adapting trending conditions.

Such is the adaptive nature of these markets.

An Edge is a Pre-requisite Required for Trading Success

Despite the limitations of the Expectancy Equation in assessing the profitability of a strategy in adaptive market conditions and the embedded assumption of stationarity in its method of calculation, a real edge or “profitability bias” in the trade history is an essential prerequisite that must be achieved before we can apply fancy money management tricks or compounding techniques to magnify the overall result.

For example, if your system does not have an edge, there is literally no way that money management techniques designed to optimise your profitability can take effect. Many retail traders get stuck in ‘the rut of pink unicorns’ where they believe a Martingale strategy or some complex progression method can come to their rescue and deliver a no-loss strategy over the long term without requiring a definitive edge in the underlying system. They spend their entire lives caught up in this fiction, and leave their photos on the fridges of their homes to remind their partner that they still exist as they devise their Masterplan in their trading rooms.

This is akin to believing in a perpetual motion machine or some arcane form of alchemy. If you want to progress down this path, then be warned….you will require a padded cell by the end of it.

Another popular trading myth derived from the ingenious George Costanza principle in the Seinfeld TV series declares that if you flip a strategy that loses money, we can turn a losing system into a strategy with an edge. Once again this is a ‘Fools errand’.

What many do not realise is that most of the time a losing strategy is simply a strategy with no edge at all. It produces a random series of trade results. If we impart trading costs into the equation then the negative drag of the costs result in a losing strategy over the longer term.

When you flip a strategy with no edge, once again you need to apply the negative drag of frictional costs and you then find that the flipped version is also a losing strategy.

You see a flipped strategy actually requires correlation in the time series of trades to offer any possibility of providing a ‘George Costanza solution’. So, say we have a losing strategy that has a clear autocorrelated bias in the trade series that sends it directly on the path to hell. When you flip this strategy, there may be sufficient autocorrelated bias in the series to exceed the negative drag of costs and turn into a strategy with an edge…..but like a strategy with an edge, a legitimate ‘Costanza Strategy’ is as rare as hens teeth.

Most retail traders simply trade random strategies with no edge at all and may or may not be successful dependent on how the chips fall. The only way you have a chance of turning trading into a sustainable career is by finding strategies with an elusive edge.

Without an edge, then psychology, money management, risk management and diversification methods won’t save you.

More importantly however, an edge is the first step towards something far more powerful. The principle of compounding which we will be visiting in the next Chapter.

Chapter 7: Compounding, Path Dependence and Positive Skew

So far in this Primer we have steered you towards at least appreciating how we, as Trend Followers, tend to think. We are a strange and sceptical mob who spend a lot of their time just observing the way processes unfold and philosophizing.

Now you may say that philosophy is all well and good but who needs to be a philosopher if we simply want to trade the markets? Well go ahead. Trade what you see or what you think, but don’t come to me crying when you get your ass whipped and then find in your disaster debrief that ‘what you thought’ and ‘how you saw’ was the actual reason you got your ass handed to you on a plate.

A belief system is so very important to your trading endeavours, as it manifests in your behaviour and thus the way you interact with these markets. Financial markets are far too important to leave in the hands of the mere mathematicians and quants. It needs philosophers to tell us a different story and open up new ways to think about trading.

Yes, we are philosophers…and proud of it. You see our whole raison d’etre for our philosophical focus lies in the fact that the Financial Industry to-date, has been woefully inadequate in providing us with the requisite tool kit to apply our trade in the tails of the market distribution and trade the divergent nature of this condition.

A one-size fits all mentality has been applied by industry under the assumption that we all trade the peak of the market distribution of returns and like to predict. Without the correct tools to use, and a litany of ‘convergent baggage’ lying in current statistical science, we just have to be philosophers.

Now while our philosophical viewpoint biases us towards divergent techniques as our preferred method of wealth creation, the real wealth creator responsible for most of the heavy lifting over a successful trading career is not the edge that lies in the technique itself (of trend following) but the path of the return stream that accompanies this trend following edge. It is the path taken which can either assist or detract the compounding effect.

Now many will say that it is the linear ascending equity curve which is the optimal path for compounding to take positive effect over the trade series, but we beg to differ. It is the stepped equity curve that is characteristic of the trend follower that offers greater benefit which includes material favourable outlier impacts embedded in the signature, but never strays too far into adverse drawdowns. The style of equity curve that is delivered by trend following methods at the global portfolio level.

You may say, “well that is not how I understand trend following. The method is inherently volatile, as are all methods with positive skew”, however once again you haven’t understood the ‘complete story’ about path dependence and compounding. Compounding takes time to achieve, but can be accelerated when including the beneficial impact of outliers in the equity curve. Compounded equity is not something delivered in a few hundred trades. It only manifests over thousands of trades, but an accelerated uplift early on it the time series from the benefits received from a few positive outliers, turns a wealth story into a rags to riches tale.

Chart 18: The Volatile but Powerful Risk Signature of Transtrend BV: Enhanced Risk Program

Table 6: Monthly Performance Returns of Transtrend BV: Enhanced Risk Program

Have a look at the equity curve and monthly returns of Transtrend above, a respected trend follower in the industry. Notice how the high returns between 2000 to 2009 have vaulted this Program way above the performance of a buy and hold in the S&P500TR Index. The reason for the accelerated growth profile can be attributed to the number of outliers it caught in its early years, whose benefit via the principle of compounding still are felt today. Notice the volatile nature of its signature particularly between 2015 to 2020 yet it has managed to keep its maximum drawdown low to 16% which has preserved it’s stellar long term performance.

Now, name me a Fund Manager (apart from a handful of ones already listed in Part 6 of this Primer series) that devote their efforts towards convergence as opposed to divergence. Here is a hint, you just won’t find them. The illusory stability of convergence, with its nice linear equity curves at the level of the individual return stream, are examples of approaches that conceal ‘warehoused risk’.

Despite the attractive lure that convergence plays, it is a short term ephemeral signature associated with temporary market stability. A symptom associated with playing around with strategies that have negative skew. The real track record for the convergent player, when viewed over the long term (such as a trading lifetime) is either a story of fantastic blow ups or at best a far bumpier ride than what divergence delivers. Most convergent players only last a few years before they hit the trading graveyard. They never experience the fruits of what compounding can deliver over the long term.

Not to say that we can’t dabble with a bit of convergence if we really know what we are doing, as I know a lot of clever guys who can pull of this stuff off when wisely applied, but without extensive experience in managing the inherent risk of negative skew, the process can compromise the ability to deliver paths that can benefit from the compounding effect.

So clearly, ‘time in the game matters’, and a prerequisite for that condition is to have a philosophy that can maintain an edge over an extensive array of different market conditions, but the edge only needs to be weak. Just sufficient to deliver a suitable path of returns that can then benefit from the heavy lifting of compounding. If we are performance chasers only looking for strategies that ‘currently’ work, or deliver illusory returns offering the false promise of a strong edge for a single market condition, then we are likely to select ‘convergent methods’ or worse still convergent methods that are excessively leveraged.

Once again for our trend following world we need to ‘flip the mindset’, and look for solutions that have stood the test of time, offer a small edge that realistically is present in a ‘mostly’ efficient market, and do not stray too far into the adverse left-hand tail of the distribution of trade returns (not to be confused with the distribution of market returns).

Dangers await in the dreaded zone of the left tail of trade returns. As trend followers, who trade both long or short, our trade distribution results convey our characteristic signature of ‘cutting losses short and letting profits run’. We never let our trade results stray into large losses associated with the left tail where danger resides, but leave ourselves open to the bounty available in the right tail of the distribution of trade results.

Without an edge as a pre-requisite condition in our portfolio development process, the ‘two-edged sword of compounding’ can send our system or portfolio into accelerated oblivion. We discussed what an edge means in our previous Primer.

Most Retail Traders do not consider this little nugget of advice and are only concerned with either the win rate or the positive expectancy of their trading solutions. This is where they typically get left behind by the professionals. They think wealth is tied to profit factor, and forget totally about compounding. Sure, profit factor is nice, as is a system with an edge, but having a return path configured for wealth building takes your wealth to a whole new level.

So, when it comes down to wealth building, it is the principle of compounding that takes the primary role and the compounding effect is intricately connected to path dependence.

We have visited this ‘path dependent’ statement before in this Primer series where we referred to an auto-correlated time series of market data as having path dependence and why we are seeking this serial correlation as trend followers. However now we are referring to the serial correlation that is present in a trade series, as opposed to a market data time series.

Given that the bulk of the heavy lifting in wealth creation can be attributed to compounding, attention then needs to be placed above and beyond the mere edge of the strategy, to now focus on the actual path taken by the returns of a strategy over time which is exhibited by the equity curve of the strategy.

No matter which style of trading strategy you deploy, the method is just the means to achieve an edge, which then allows you to use compounding over time to achieve a far greater end. What we are seeking is a method that offers both a weak edge AND can provide us with the best return over time that provides an optimal path for compounding to work its magic.

So, with both convergent and divergent styles to consider that can offer an edge, which is better suited for compounding? To be able to answer this question, we need to understand what a geometric return is.

A geometric return differs from an arithmetic return by considering the impact of serial correlation in an equity curve and is primarily used for investments that are compounded. It is expressed as:

Serial correlation is a path dependent concept where a time series possess variables that are correlated across time. For example, an equity curve is a representation of the performance of a trading strategy over a particular time horizon. If there is an upward overall bias in the equity curve over time which lead to progressively higher equity, then the equity curve is said to possess positive serial correlation.

Where serial correlation in the equity curve changes direction and becomes negative in direction leading to auto-correlated losses or the serial correlation ceases to exist leaving only an independent series of random returns remaining, an equity curve then either enters drawdown or stagnates.

An equity curve with overall positive serial correlation possesses a bias towards positive equity, with only a few smaller excursions into negative equity. Hopefully, the serial correlation in the trade history is spread throughout the trade history leading to less ulcers for the trader, than those equity curves where serial correlation ‘clusters’ and is haphazardly dispersed through the trade history. Where and when the serial correlation resides, and its direction, are the dominant culprits for the volatility of an equity curve.

Under a serially correlated path, the impact of compounding over that time series at regular intervals is exponentially magnified in accordance with the degree of positive serial correlation present in that series and where it lies (aka its distribution) in the series.

When positive returns that are generated by the strategy are reinvested back into the equity at regular intervals with a positive bias, we hopefully achieve an upwardly rising equity curve. Of course, the trajectory of the equity curve relates to our overall performance. If we have an edge over time (which is a serially correlated positive bias in the series), then the equity curve will rise over time. If however we have negative expectancy, our equity curve will decline in value as we have no profits to reinvest back into the curve.

The intent of profit reinvestment is to progressively magnify the equity which is then periodically compounded at reinvestment intervals. This path dependent feature of investment is critical to wealth building as we want the principles of compounding to do most of the work in wealth creation.

If you consider the compounding process and how it applies at equal intervals over a time series, you could imagine that if an equity curve is rising, then the impact of progressive compounding through reinvesting profits) increases. For example, let’s assume we risk 1% of equity per trade for a particular strategy. As equity builds over time through the reinvestment of profits, if serial correlation persists, then the 1% trade will act on progressively building equity leading to an exponentially rising equity curve over time.

Conversely, if the strategy has a long unfavourable drawdown where much of the time is spend in negative equity with equity declining, there obviously is little profit to reinvest. The 1% trade risk application where any profits are reinvested does not have the ability to generate as much profit from the compounding effect than a favourable equity curve continuously reaching its high watermark.

So, let us have a look at a few different equity curves arising from non-compounded strategies and observe the impact on this series when profits are reinvested in the strategy (aka we turn the series into a compounded series). The variation in Net Wealth between different strategies attributed to compounding will astound you. You will then clearly see how the path of the equity curve over time is critical in the outcome with respect to long term net wealth.

Having this understanding then allows you to compare the equity curves of different trading strategies and identify those that are optimally configured for compounding.

Chart 19: Non-Compounded Equity Curves of Three Separate Systems with an Edge

Chart 19 above displays 3 separate Equity curves derived from 3 separate trading systems that all possess an edge. The non-compounded result of each trading system is very similar where each strategy commences with a single dollar in equity and at the end of 1135 trades all possess similar ending equity balances of approximately $1,144.

To achieve this non-compounded result, all we do is apply an equivalent fixed $ risk per trade over the duration of the time series. All trades are treated equally in terms of their allocated $ risk.

Despite the similar result, each take a different path to achieve this result.

• System 1 is a steady performer over the entire history of trades with no significant drawdown or acceleration in returns.
• System 2 has a slow start with subdued performance up until trade 1,045 (Label C) where it then experiences very strong equity growth.
• System 3 has a very strong start easily outperforming both System 1 and System 2 until trade 1,045 where it starts to enter a steep drawdown.

So now let us have a look at the impact of these 3 different paths when we compound the result. All we are doing is applying a trade risk % to each trade based on the level of equity at the time of placing a trade. This therefore means that we apply a fixed % risk per trade of Equity as opposed to a fixed $ risk per trade which achieves the non-compounded result.

Chart 20: Compounded Equity Curves of the Same Three Separate Systems with an Edge

The compounded result displayed in Chart 20 possesses a totally different performance result to that achieved through the Non-Compounded Result. Clearly, we can see that in terms of overall Net Wealth at the end of the time series, System 1 produces the superior result. This all attributed to the heavy lifting of the compounding affect and the equity path taken by each separate system.

Let us examine why?

The initial trajectory of the System 1 uncompounded path is subdued (Point A), compared to both System 2 and System 3. As a result, when the principles of compounding are applied after each trade to these subdued equity levels, we do not have much growth in the equity curve in the early stages of the trade history (up to Point B). We cannot see this subdued effect early in the time series in the Compounded Chart takes time for compounding to exercise a bias to the time series.

However, when we reach trade 1,045 (Point C) we start to see compounding take hold in the time series with a rapid acceleration in equity. Given the late stage where compounding takes hold, despite the rapid rise in equity from Point C, it does not catch up to the levels of equity achieved by System 2 and System 3.

Unlike System 1, System 3 is a very strong early performer, and you can quickly see how compounding takes effect to lift the equity curve well beyond the alternate strategies. However, we start to see the acceleration diminish from Point B with a rapid deceleration into negative equity by Point C. This under-performance is very detrimental to a compounded equity curve and quickly reduces overall equity of System 3.

System 2 however is a very steady performer and as a result, the compounding affect can accelerate the equity curve along the entire trajectory. Clearly it is therefore the consistent more stable equity curve we are after to deliver superior compounded returns.

However as discussed earlier, an even better path for compounding is one which has significant lifts in the equity curve early on in the time series, yet never suffers any significant adverse volatility that compromises the compounding effect. Having the early-step-ups means that compounding is accelerated and time is provided to turn that acceleration into an exponential growth curve from the get go.

What significantly compromises the compounding effect of an equity curve is a progression into negative equity. Compounding, like leverage, is a two-edged sword. When equity is building, compounding magnifies the $ gains but when the equity curve is entering drawdowns, compounding magnifies the $ losses.

From a close examination of the 3 systems described above you will notice the path taken is critical for the compounding effect to deliver long term wealth. While the compounding bias is slight, it plays a progressively more dominant role over the long term with a larger trade sample size. Given the short-term lifespan of a retail trader, the powerful effects of compounding are rarely experienced. This is why trend following is considered a long-term game of wealth building. Convergent players may achieve fast returns immediately but we are after far superior returns for the long term.

Clearly, if we want to play this trading game and achieve wealth over our lifetimes, then we need to take compounding seriously. The only way we can participate in the fruits of compounding is to be a survivor that can boast a long-term trace record with a persistent edge.

If that edge is compromised by significant extended drawdowns, or if our equity curves are too ‘negatively’ volatile leading to very sharp drops in equity, we will severely compromise our wealth ambition.

While trend following spends a lot of its time in drawdown, the extent and speed of these drawdowns are limited when compared to other methods over the long-term time horizon. This is what makes their return streams ideal for compounding.

Our aim with our Trend Following method is to achieve a portfolio result that embraces positive volatility yet avoids negative volatility. If, at the portfolio level, we can smooth the equity curve to avoid extended periods under-water, and better still retain the characteristic step-up signature, then we can use the miracle of compounding to lift our equity to the heavens.

It is not the volatility of the Equity Curve that bothers a trend follower, as favourable volatility (which results in an equity uplift) is exactly what we want to take advantage of with compounding. What we want to avoid is adverse volatility that detrimentally effects equity levels and diminishes the ability of compounding to magnify the equity of the time series.

Now the long-term sustainability of progressively rising equity curves is significantly hampered by the trading technique you wish to deploy. Some equity curves delivered by convergent methods over the short term, have a deceptive linear ascent while favourable market conditions persist, however this short-term feature is compromised when conditions change and become unfavourable. The linear ascending equity curve is now compromised by significant and fast drawdowns in equity which significantly impede the compounding effect.

This style of equity curve comprising smooth ascents over favourable conditions and then dramatic drawdowns during unfavorable regimes which may also lead to risk of ruin if the favorable condition does not resume is associated with predictive systems such as mean reversion methods.

Fortunately for diversified systematic trend followers, we adopt extensive diversification amongst a myriad of different equity curves that when compiled into a portfolio provides a relatively smooth ascent of portfolio equity including step-ups with far lower and less aggressive drawdowns over the long term.

This feature of diversified trend following systems makes them an ideal candidate for generating exception long term wealth aided by the principles of compounding.

We need to drill down more into how a diversified trend follower manages the volatility present in their portfolio’s equity curve, and to understand this, we need to investigate the concept referred to as Positive Skew.

Positive Skew

Now that we understand that is it downside volatility as opposed to upside volatility that we need to be concerned with in delivering the optimal path to wealth building we can throw out the traditional ‘Gaussian’  tools used by industry for risk management such as the Sharpe Ratio which are ambivalent in their treatment of the direction of volatility, and other risk management measures that are not path dependent methods such as the Sortino ratio.

Remember that we operate in the non-linear land of the fat tails where standard statistical measures based on an assumption of ‘Efficient Markets’ no longer applies.

We need to look at risk management tools that explicitly deal with asymmetrical ‘exotic’ market conditions associated with fat tailed behaviour, as opposed to symmetrical risk management tools associated with markets in equilibrium. Our trade performance as trend followers is dictated by our ability to catch outliers and prevent ourselves from wandering into the left tail of the distribution of trade returns. This leads us to the characteristic ‘stepped equity curve’ that is naturally configured to offer the greatest geometric returns through compounding.

One of our biggest weapons we use to allow for an optimal geometric return (or path) is our preference for only trading systems with positive skew.

When discussing positive skew, we are referring to the skewness of the distribution of our trade results. This should not be confused with skewness associated with market data. Given that our divergent methods have a low win rate but a high reward to risk ratio, we find that if we plot our trade results in a histogram, only a handful of successful ‘outlier’ trades are the real drivers of our profitability (as we let profits run).

By far the majority of our trades are either small losers (as we cut losses short) or are breakeven trades or small wins representative of the fact that the vast majority of our trades are just random results. There are only a handful of trades that have been able to catch a trend with endurance backing it.

Chart 21: Distribution of Trade Returns of a Typical Trend Following System with Positive Skew

The result for a plot of our distribution of trade results typically follows the histogram above in Chart 21. If you closely examine the chart, you will see that there are no or very few significant material losses. The vast majority of losses cluster with the range of 0 to -6 R. In this example R is simply a way to express risk$ in equivalent terms. So, a risk of 10R is 5x larger than a risk of 2R. There is only 3 trade events that were worse than -6R attributed to system application errors.

You will also note that the vast majority of winners cluster in the range between 0 to +12R. The total distribution between -13R to +12R leads to a breakeven result for this model. However, it is the few winners that exceed +12R all the way to +45R that lead to overall net wealth of the system.

What you will note from this distribution is that is it constrained in its left tail. This means it is exceedingly difficult to achieve significant material losses that lead to fast and excessive drawdowns.

In statistics, a positively skewed (or right-skewed) distribution is a type of distribution in which most values are clustered around the left tail of the distribution while the right tail of the distribution is longer.

Skewness refers to an asymmetry that deviates from the standard bell curve (or normal distribution) in data (in this case trade results).

Distributions can range between having positive skew to zero skew to negative skew. In essence this can be stated in the following manner.

If your trade distribution has much larger average $wins than average $losses, then the distribution will be positively skewed. If your trade distribution has equivalent $wins to $losses, then the distribution will have zero skew and follow a normal distribution. However, if your trade distribution has smaller average $wins than average $losses then the distribution will have negative skew.

In practical terms a trader with negative skew, provided they have a high win rate, will achieve positive expectancy and the equity curve will grow, however if market conditions change and the Pwin% drops, they become excessively exposed to the left tail of the distribution of trade returns. A sequence of large $losses (which is a symptom of negative skewed strategies) leads to fast drawdowns and potential risk of ruin.

A trader with positive skew however can survive far longer during unfavourable market regimes as the small losses do not build drawdowns to the same material degree as their negative skewed counterparts. Furthermore, the system risk constraints that prevent the trader from entering the left tail make this style of system surprisingly robust and capable of weathering uncertainty.

The speed and nature of drawdown associated with the skewness of the trading system is a major feature that compromises geometric returns.

The upside volatility associated with the large average $wins of a divergent trader contributes to Geometric Returns whereas the large average $losses of a convergent trader can compromise long term wealth ambitions.

Let us have a look at how we can possibly identify positive or negatively skewed trading systems or portfolios by looking at their equity curve.

Have a look at the non-compounded equity curve below in Chart 22. The blue  line represents the equity balance from all closed trades and the green line represents the unrealised balance or the ‘mark to market’ value of equity at all points in time. This green line represents the value of all closed trades and the unrealised profits or losses on all currently open trades.

Chart 22: Non-Compounded Equity Curve of a Positively Skewed System

You will notice that there were never any points where the green line wandered significantly below the blue realised balance. This means that over the course of the trade history, there were no points in time where the unrealised risk held by the strategy got out of hand. The strategy simply always cut losses short. There were however times where unrealised equity was higher than the realised balance which is a consequence of leaving some profit on the table. Unfortunately, this is the nemesis of the trend trader. We can never time the end of trends and unfortunately always leave some profit on the table when the trend ends.

Now while we had to leave some profit on the table, we never allowed the strategy to stray towards the left ‘fat tailed’ side of the distribution of trade returns.  In practical terms this means that we never exposed ourselves unduly to future risk. We managed risk in the strategy across the entire duration of the equity curve never letting it get out of hand.

The profile of the equity curve however now lends itself to compounding. There are no significant material drawdowns that compromise the compounding effect and there are many upward jumps in the equity curve which accelerate the compounding effect on the series.

Now let us have a look at an equity curve with symptoms of negative skew (Chart 23).

Chart 23: Non-Compounded Equity Curve of a Negatively Skewed System

In this extreme example which applies to a Martingale Technique that is a classic example of an extreme negatively skewed system you will notice that the unrealised equity curve (green) can sit well below the realised balance. We refer to this profile as a strategy that is ‘warehousing risk’. Other common examples include strategies that warehouse risk using averaging down principles.

Convergent strategies such as mean reversion techniques frequently fall into this class given their need to apply tight profit targets and more lenient stop losses (if used at all) to catch the falling knives of mean reversion.

In Chart 23 above, without a regular mark to market (appraisal of unrealised equity) it appears (using the blue line) that the strategy is performing exceedingly well with a steadily rising equity curve and few apparent losses. This strategy however has been closing all trades in profit and holding onto trades in a loss and never realising the risk.

The real story of this strategy is that it possesses extreme negative skew. Many small losses with only a handful of losses that can lead to complete risk of ruin.

Now for the purposes of compounding you can clearly see the result of the fast severe drawdowns and its impact on equity. Such dramatic drawdowns have dire consequences to the compounding effect which is dependent on the path taken by the return stream.

So, what we have learned in this brief introduction to Skew is that the compounding principle favours those systems that display positive skew.

Compounding with Leverage

Now we understand the pivotal role that path dependence plays in delivering wealth through the compounding effect, we can now turn to the impact of when we get too aggressive with our leveraged trend following solutions.

Remember that compounding is a two edged sword. If the impact of compounding over the long term progressively exponentially magnifies the result, we can be left with a death defying ‘pogo stick ride’ with some of the FM’s that play with higher leverage. I refer to these style of highly volatile trend following funds as “Formula One Grand Prix racers with speed wobbles”.

Chart 24: Mulvaney and the pogo stick

Now the Trend Following community love Paul Mulvaney but that is just because we understand his philosophy that goes for the ‘king hits’, however for those that get ulcers when big drawdowns arrive, then they may get concerned. There is no doubt that the Mulvaney Program is one of the top performing Trend Following Programs in the world, but if you want a piece of this extreme action, then do not look at the monthly performance.

Paul clearly knows his stuff given his very long track record and recognises that with high volatility comes very high returns, but this can turn a Trend Following method into one in which encourages investors to time their entry into this style of Program which you want to avoid.

Now as long-term as a Trend Following Program might live, it won’t last forever and there comes a time where we are tempted to just let the reigns go, loosen our risk  belts and go for the king hit, like Paul. You see if we put on our racing  gloves, we can choose to go-for-broke to demonstrate the real power of these Programs under leverage, however for those investors seeking high Net Wealth in say 50 years time from now, then such a volatile offering may not be the solution as timing entry into the equity curve becomes important.

Other Trend Following Managers deliberately turn down their potential returns by reducing their leverage to cater for a long-term investor seeking a smooth ride where compounding can take the full effect.

So now that we understand that the profile of the equity curve (or more specifically the path of returns) is the major feature that enables compounding to accelerate wealth, we can now explore a central feature that Trend traders focus on to achieve an optimal path of returns. This is what is referred to as a Risk Adjusted approach to maximise geometric returns which we will be exploring in our next Chapter.

Chapter 8: A Risk Adjusted Approach to Maximise Geometric Returns

In the prior Chapter, we discussed that while an edge is an essential prerequisite for a trend following portfolio, the real game changer that is responsible for long term net wealth is the principle of compounding over the long-term.

Having this understanding, we then demonstrated how the path of returns delivered by a portfolio, is responsible for the degree of compounding (aka the geometric return) that can be achieved over a time series.

Better geometric returns are achieved when adverse volatility is mitigated, and beneficial volatility is embraced. Trend followers place a great deal of emphasis on managing risk in the portfolio to ensure that adverse drawdowns do not materially and detrimentally compromise wealth building objectives but they also leave their profit potential open (both long and short) to allow for unbridled step-ups in our equity curve through targeting outliers that lie in the tails of the market distribution of returns. It is this latter perspective that most of us don’t pay enough attention to.

In fact, when compared to alternative trading methods, the classic equity curve of a trend following global portfolio over the long-term is configured to generate superior wealth building returns. This is because the path of returns generated from trend following methods offers superior risk-adjusted metrics to alternative approaches.

Notice how I have focused on two separate aspects of a trend followers modus operandi when discussing how a trend follower’s process naturally produces a path of returns suited to compounding.

Most traders only consider the adverse volatility side of this statement and not the positive volatility side, and believe that a nice straight ascending equity curve (popular with convergent systems that possess negative skew) is what is required to produce the best path. But this is a problem that applies to traders with a linear mindset. We are non-linear guys who recognise that the ‘when and where’ non-linear things happen in a time series such as outliers, have exponential effects on the overall end result of that series.

Trend followers target the outlier (a non linear feature), and manage adverse tail risk at all times to produce a superior ‘stepped’ path to their cousins. In fact, as we will see in this Primer, we use diversification of non-correlated trend following return streams to further enhance the path of returns. We understand that, unlike traditional mantra suggests that “there are limits to the benefits that can be achieved by diversification”, there actually are no such limits for trend followers who trade the fat tails.

You see once again the devil is in the details. We are not simply using diversification as a method to manage adverse risk. If so, then we could agree with conventional wisdom that there was a point where no further marginal benefit would be received by additional diversification.  However, this is not the case for trend followers as they use diversification to achieve two objectives:

• To mitigate adverse volatility which detracts from the benefits of compounding; AND
• To target the tails of the distribution of returns where the unpredictable outlier resides. This additional nugget vastly accelerates the compounding effect over a time series.

It is the latter point that makes us understand that there are no limits to the benefits of diversification when trading the fat tails. Given that outliers are an unpredictable event, then the more diversified you are in targeting them, the greater the chance you have of ‘riding them’. When playing convergence on the other hand, by restricting the ability for your solution to capture outliers by using profit targets, you vastly underestime what compounding can achieve though its heavy lifting potential.

You see, through our method of collating diversified return streams into a global portfolio, we deliberately select return streams that when collated, provide offsets to reduce adverse excursions of the portfolio equity curve AND then we pile in as many uncorrelated return streams into the portfolio as we can muster, to target outliers and receive correlation benefits which accelerate the compounding effect.

Let us provide an example of what we mean by a risk-adjusted approach.

Below in Table 7 are performance metrics and non-compounded equity curves for the currency pairs AUDJPY and EURUSD that adopt a trend following system using a classic trend following approach.

Notice that the Maximum Drawdown of AUDJPY is 56.15%, the CAGR is 7.65% and the MAR ratio (which is the CAGR/Max Draw%) is 0.14. By visual reference you will notice periods of time where significant step ups in the equity curve were experienced. For example, at around 1986, 2009, 2014, 2017 and 2021 (refer to yellow ellipses)

Also notice that the Maximum Drawdown of EURUSD is 28.27%, the CAGR is 7.40% and the MAR ratio (which is the CAGR%/Max Draw%) is 0.26. By visual reference you will also notice periods of time where significant step ups in the equity curve were experienced by this market. For example, 1981, 1988, 1998, 2001, 2009 and 2015.

Both return streams have volatile signatures, however adverse volatility (expressed by Drawdowns) are relatively slow to build and not material in nature, but outlier impacts are steep, fast and major in extent.

Table 7: Non-Compounded Return Streams of AUDJPY and EURUSD using a Trend Following System 1^st Jan 1980 to 25 Mar 2021

Now let us compound these trend following results to see what can be produced in isolation and then as a composite.

Let us start this process by looking at compounding in isolation, where we apply a trade risk percentage of equity as opposed to a single lot per trade starting for AUDJPY only. We are simply looking at how each of these non-compounded path’s fare when we allow compounding to have a say in the matter.

Table 8: Non-Compounded Vs Compounded Return Streams of AUDJPY and EURUSD using a Trend Following System 1^st Jan 1980 to 25 Mar 2021

Table 8 above demonstrates several points of note.

• The uncompounded solution produces a CAGR of 7.65% and a Maximum Drawdown of 56.15%. The relationship between the return and the adverse drawdown is a MAR ratio of 0.14. With a Max Draw of 56.15% it is unlikely that many traders would possess this risk appetite when returns on offer are low.
• The compounded solution using low leverage of 1% trade equity using the same underlying non-compounded return stream produces a comparable CAGR of 7.79% but with a far lower drawdown of 31.28% and a far higher MAR of 0.25. Given the lower ulcers delivered by this compounded solution for the trader, it is far more likely that this strategy would be preferred to the non-compounded solution….and yet the fundamental basis for both return streams are the same?

This is what we mean by path dependency when dealing with compounded solutions. The shape of the return series created by classic trend following solutions is configured to allow for compounding to do heavy lifting in accelerating returns and reducing adverse volatility.

The third compounded solution, using high leverage of 5% trade equity is just used to demonstrate what these trend following systems are capable of if we unleash the tiger in them. In this extreme example, the solution delivered a powerful CAGR of 18.37% but you would have had to tolerate a maximum drawdown that nearly wiped the entire account of 87.64%. Now I am not saying any of us would tolerate this, but this is what a trend following return stream can generate if we want them to.

However, note the following with this ‘adrenalin pumping Formula One Grand Prix tiger’, despite the higher returns, look at what has happened to the MAR ratio. It has declined to 0.21. This is what happens under compounding where volatility of the return stream starts to compromise the compounding effect.

What we are therefore actually after, is the most efficient balance between risk and return expressed by the MAR ratio. Unlike other risk metrics such as Sharpe and Sortino, the MAR ratio (or for the advanced, the Serenity Ratio) is our preferred metric as it is a path dependent ratio that tells us the most efficient balance we can achieve for compounding to then do its heavy lifting. It is not about the risk inherent in the return stream itself that matters to us. We actually like volatile equity curves provided that the volatility is capped on the downside but left unhindered for the upside.

Note to readers: The Serenity Ratio is a complex path dependent risk adjusted metric that only marginally outperforms the MAR ratio. For purposes of simplicity, we only discuss the MAR ratio in this series. For interested readers, the following article from KeyQuant provides an excellent introduction into this powerful path dependent metric. 

In fact, when you look at conventional wisdom, it is suggested that statistical practice should exclude the impact of outliers in any assessment, as their ‘magnitude’ overwhelms the average volatility of the every day return series.  We of course beg to differ. When it is the extreme non-linear anomaly that is responsible for the greatest overall change across a time series, then it is essential that we pay specific attention to them.

Worrying about the ‘average’ drawdown of a series, or the average volatility of a series is simply not paying justice to the material effect that outliers play in determining a traders future successes or failures.

What really matters to us is the path or returns that are generated by our systems that gives us greatest bang for risk buck when we apply compounding to the series.

This is why we treat single measure ‘exotic’ points in the time series seriously such as the maximum drawdown, or the impact of a few beneficial outliers in that series. We pay close attention to the maximum favorable and maximum unfavorable excursion of a time series which is pivotal to path dependent outcomes. Water down the impact of these non-linear monsters at your peril. You will achieve at best moderate absolute returns over your lifetime, or at worst an account blow up, if you don’t take heed of this warning. Trend followers know this stuff and we strive for exceptional absolute risk-adjusted returns.

What importance is an average to the following string series of returns?

0.5, 1,-2,3,0.7,4,3,8,10,5,4,7,9,105,0000

The average of a series like this is inconsequential to us. What really matters to us is the impact of the 105,0000 in that series….or the anomaly that resides in the series. If you don’t want anomalies, then don’t trade the tails. Focus on the predictive game that resides in the peak of the distribution of returns….but we won’t be holding our breath that we will see you alive and well in the future.

Now to get back on track, rather than look at all the compounding options for our next example of EURUSD, I will simply show you the 1% compounded solution when compared against the non-compounded solution. This will once again ram home the point that path dependency really matters in our pursuit of absolute return over the long haul.

Table 9: Non-Compounded Vs Compounded return Streams of EURUSD using a Trend Following System 1^st Jan 1980 to 25 Mar 2021

The MAR of the compounded return stream exceeds the MAR of the non-compounded return stream in Table 9, simply due to the path dependence of the series. Once again trend following comes to the rescue and optimally configures a return path that allows compounding to work its magic. This is exactly what you need for the long road to wealth building.

We have not had to incorporate tail risk hedges into the portfolio return stream through clever option practices, as the trend following systematic processes have already embedded these principles within the technique.

So now let’s see what happens when we adopt an additional layer of secret sauce to this recipe through adding principles of diversification. We are not simply satisfied that the trend following method itself incorporates the fruits of path dependence into the return series. We then stretch this logic further and go to another new level of risk-adjusted magic by incorporating different non-correlated return streams into the process. This injects further long term wealth into the process.

So now we combine the 1% compounded AUDJPY and EURUSD series together at the global portfolio level, as we know that the location of  drawdowns and the location of beneficial outliers in the series are not likely to exist at the same points in time for each separate return stream.

Table 10 below demonstrates where the outliers reside for each series. Notice that some occur at the same location contributing to the windfall at the same time, whereas others occur at different point of the time series. This applies equally to drawdowns. A Max Drawdown in one series will be offset by a lesser drawdown or even a high-water mark in the other series if they are uncorrelated.

Table 10: Visual Chart highlighting the locations of Outliers in both Series (AUDJPY and EURUSD) using a Trend Following System 1^st Jan 1980 to 25 Mar 2021

So, using this principle of outlier and drawdown offsets to produce better risk adjusted returns at the global portfolio level we would rather compile two separate lesser correlated series of 1% each than simply doubling down with a 2% application to a single return series.

So here we go. Let us see if the MAR ratio and the compounding effect is improved through this risk-adjusted process.

Table 11: A Simple Portfolio Comprising two Markets (AUDJPY and EURUSD) using a Trend Following System 1^st Jan 1980 to 25 Mar 2021

Table 11 highlights how under diversification, the risk adjusted process undertaken produces a superior result with a CAGR of 10.13% and a Maximum Drawdown of 30.56% and a more efficient MAR ratio of 0.33. We have a far better relationship between return and risk under this compounded solution than what could be achieved without diversification.

So what this Chapter has shown you is that the reason trend followers always talk in terms of a ‘risk-adjusted’ process is that they know that ‘the path taken by the return stream’ is an essential criterion of wealth building. Having a weak edge is only the first step consideration for a Trend Follower. The real challenge then lies in how we shape the equity curve itself to allow compounding to do its magic.

To achieve this outcome:

• we must think long term as we can only see the work of compounding delivering wealth over a significant trade sample;
• we should not sacrifice our geometric returns through excessive leverage and be satisfied with trading a weak edge. That is all that is required for compounding to do its work and not be compromised in the process.;
• we must pay attention to the maximum favourable and unfavourable excursions of the equity curve and not be side-tracked by the balance of volatility that lies in the curve. The favourable outlier cannot be excluded from our assessment. A mere focus on drawdowns is not sufficient as outliers accelerate our ambitions;
• we must cut our losses short at all times and just let our profits run to embed the best path into the trade series; and
• we must diversify to not only reduce adverse volatility of the path of returns, but also significantly increase the chances of bumping into a favourable outlier or two, or three……

In our next Chapter we will take a deeper dive into Diversification as this is a very valuable tool for the trend follower that does provides a free lunch in its treatment of adverse risk, but also just as important, increases our ability to capitalise on outliers. This is rarely mentioned as most trading methods focus on the peak of the distribution of returns. We don’t. We focus on the tails where the non-linear extraordinary event matters.

So much attention is placed on maximum drawdowns with so little attention placed on the outliers in delivering a trend followers windfall. We intend to address that discrepancy in this Primer.

Chapter 9: Diversification is Never Enough….. for Trend Followers

When dealing with diversification, it is tempting to adopt conventional wisdom and talk in terms of diversification as a method to reduce overall portfolio risk exposure, but for trend followers, diversification is so much more. It provides a method to probe the markets and capture the elusive outlier, it allows a trend follower to take advantage of correlated markets that are beneficial to their cause and it allows a trend follower to capture the many forms of trend form that exist in the land of the fat tailed environment.

For that reason, the title of this Chapter suggests that there can never be enough diversification for trend followers. Here we will be exploring the conventional aspects of diversification and then also explain the many additional benefits that diversification can bring for the diversified systematic trend follower.

Principles of diversification relate to how we can obtain additional benefits from a portfolio above and beyond what can be achieved by its component return streams. In effect this statement is saying that ‘the whole is greater than the sum of its parts’. It provides an essential overlay for all systematic trend followers that we simply cannot go without.

Some trend followers think that diversification holds the mystery to the ‘secret sauce’ of trend following and why the method is so robust as a trading technique. I am happy to leave people with that conclusion. For diversification is not just a ‘free lunch’ it is a ‘free banquet’ for the entire Trend Following community.

The power of diversification for trend followers is one of the most important aspects of our approach to trading these markets and so we will be spending a little more time digging into the nuances of diversification in this Chapter.

Some of the Many Benefits of Diversification for Trend Followers

In our previous Chapter, we saw how diversification is a tool used by trend followers to sculpt an optimal path of returns for the purposes of compounding. This is the primary reason for why trend followers dedicate their efforts towards diversification. After all, delivering absolute returns and significant windfalls in net wealth over the long term is what this game is all about.

Risk management through diversification however is simply one aspect of the optimal path for compounding that reduces the adverse volatility of the portfolio equity curve, but we should never forget the other important aspect of the optimal path which is ‘beneficial volatility’ and its impact on compounded growth.

In terms of this beneficial aspect of volatility for trend following portfolios, diversification is our preferred method of increasing our chances of catching outlier moves and capturing any beneficial correlations that exist across return streams when markets get ugly and highly correlated.  This feature of diversification exponentially magnifies geometric returns.

But there are many more benefits as well.

The list below includes some of the many powerful benefits of diversification that are specific to trend followers who target the tails of the distribution of market returns. Some are applicable to other alternative trading and investment methods, but many are specific for our trading approach. You can therefore understand why a trend follower can never be diversified enough and that our efforts are placed on continuously improving our systematic diversified portfolios:

• Diversification is a method to balance the adverse volatility present in a portfolio equity curve using correlation offsets, to produce higher risk adjusted returns than could otherwise be achieved through its constituent parts;
• Diversification is a method to take advantage of beneficial volatility in a portfolio equity curve either by increasing our chances of riding outliers, or by summating positively correlated ‘beneficial aspects’ of return streams to accelerate the compounding effect;
• Diversification is a method to increase trade frequency for a trend follower who need to patiently wait to enter ‘material trends’. More opportunities are provided under diversification to capture outliers; and
• Diversification is a method to capture many different forms of trend and trend segments known to occur outside the Gaussian envelope that are inherently unpredictable in form, and is a method that allows a Trend Follower to offer a distinct point of difference.

Diversification – A Method to Balance Risk in a Portfolio

At the broadest level from which most traders would be familiar, diversification seeks to reduce a trader’s exposure to the risk of any single return stream by spreading their trading capital across numerous return streams that are uncorrelated in nature. The basic premise is that when one return stream is under-performing and entering a drawdown, the other return streams in the portfolio are hopefully performing, and thereby offsetting the impact of this single drawdown in the portfolio.

The principles of diversification therefore seek to smooth adverse volatility of the portfolio equity curve. I refer to this diversification process as ‘weaving the most robust tapestry’. So, chart 25 below shows the loom which contains 25 various threads (return streams), and Chart 26 shows the robust tapestry produced after we weave the threads together.

Chart 25: Twenty-Five Return Streams Compiled into a Portfolio – “The Loom”

Chart 26: Twenty-Five Return Streams Compiled into a Portfolio – “The Tapestry”

Let us have a detailed look at how this principle works where we start with Chart 27 to observe this principle in action.

Chart 27: Adverse Risk of Return Streams which are Offset in a  Diversified Portfolio – Single Return Stream

Chart 27 reflects a single non-compounded return stream of GBPUSD which has a volatile signature. In this example we apply a $50,000 initial balance towards a single return stream ending in a final balance of $65,810 over the 21-year period.

Notice that the Reward to Risk relationship between the Return on Investment (ROI) and the Maximum Drawdown is a MAR ratio of 0.15. This provides an expression of the risk adjusted performance of this return series and its capacity to benefit from compounding through increased leverage over the time series.

Notice also how the Maximum drawdown was reached in July 2014. This clearly was an unfavourable environment for trending conditions in GBPUSD over this period. In fact, the unfavourable period extended for a 4-year period commencing in late 2009.

This maximum drawdown and its extent is a signature that we take note of. It is a sign of possible risk weakness for an overall portfolio if that return stream was included in a compilation of return streams. In this case, the Maximum Drawdown for GBPUSD was quite extreme relative to the overall path of the return stream it generated. This has led to a low MAR ratio which states that this return stream is unlikely to positively contribute to a portfolio and has risk weakness embedded within it.

So now that we understand what MAR tells us about a return stream, let us see what is produced when we compound this return stream in Chart 28.

Chart 28: Monthly Compounded Portfolio – 1 Return Stream

Chart 28 reflects that given the poor risk adjusted metrics of the single return stream, it was hardly worth the effort compounding this solution. The initial $50,000 balance ends up with only a very slight increase to $67,643. Now a Compound Annual Growth Rate (CAGR) of 1.43%

But what happens if we diversify the portfolio across 4 different return streams?

Have a close look now at how diversification works in Chart 29 where our initial $50,000 portfolio now includes 4 separate return streams under a non-compounded case which includes our GBPUSD example.

Chart 29: Adverse Risk of Return Streams which are Offset in a  Diversified Portfolio – Multiple Return Streams

Note: In this example, the return streams represent diversification across markets, but the principle is the same for return streams generated by diversification across timeframes and diversification across systems.

Now the first thing to notice in Chart 29 is that the final balance at the end of 21 years staring with an initial balance of $50,000 is now $144,358. Wow! We haven’t cherry picked these examples, but simply find that by diversifying across different return streams, some of those return streams have simply offered outstanding long term performance metrics. Thank goodness for our non-predictive logic.

When we were back in 2000 and allocating initially to GBPUSD, we had no expectation regarding future performance. However, back at 2000, if we now equally allocated across 4 different return then, without any form of selection bias, we have found that we significantly improve the non-compounded result. This is the same principle that venture capitalists use in their selection process that seeks to find a handful of possible winners from a large basket of future ‘false starts’.

So back to our risk management story under diversification for Chart 29. We have equally allocated risk across 4 return streams (not simply one return stream) using our trend following models, with no assumption made about whether they are extremely correlated or not. We just assumed that they would not be 100% correlated with each other across the entire length of the return stream.

So, recognising that the different return streams are unlikely to be perfectly correlated, we should receive adverse risk correlation offsets that will assist our endeavors.

When we examine Chart 29, notice how now the adverse maximum drawdown of the portfolio now resides at August 2012 and not the prior maximum drawdown observed for GBPUSD of July 2014. Under diversification, the principle of re-balancing risk across the portfolio has now been observed. The maximum drawdowns of individual return streams have now been offset.

The new maximum drawdown for the compiled portfolio is now 9.35%, which in this example is slightly higher than GBPUSD alone of 8.62% but in terms of its impact on geometric returns over the long-term, it must be expressed in relation to the return that the portfolio now delivers. So given this understanding, let us look at the ROI to see if any benefit has been received by the portfolio to compensate this slight increase in drawdown exposure.

Just look at what the ROI now is! For a small increase in Maximum Drawdown, we now achieve an uncompounded ROI of 5.04%. This is a massive increase in ROI from our GBPUSD series of 1.30%. In fact, we can see the improvement in the risk adjusted metric of MAR in that it has increased to 0.54 from 0.15 which is a 3.5x increase over the time series.

This means that this compiled portfolio will perform significantly better under monthly compounding.

So let us see what it produces when compounded.

Chart 30: Monthly Compounded Portfolio – 4 Return Streams

Kaboom, kapow!! Chart 30 reveals that under monthly compounding the initial $50,000 balance has now increased to $271,181. This is a $126,824 increase from the non-compounded final balance result of $144,358.  A CAGR of 8.04%.

Last words on Risk

While it is tempting to conclude with these examples that diversification is a method to reduce overall portfolio risk exposure, it does no such thing. Risk cannot be created or destroyed once a trade has been initiated in the marketplace. Only transferred. This is a strict Conservation Law for the Financial Markets. We can never eliminate risk, as risk is encountered as soon as we elect to undertake a trade event and participate in the markets, but we can manage it within ‘reasonable bounds’.

A portfolio is therefore a method we use to not reduce but rather ‘balance’ the risks associated with its constituent return streams by using principles of offset.

So, while a portfolio balances the risks in a portfolio, the total summated risk of all constituent return streams is still all there. We like to refer to a portfolio as a risk sponge that soaks up all risks and stores it in its architecture but relies on the fact that the constituent return streams can offset adverse risk by their uncorrelated nature.

The only way we can release risk from the portfolio is by removing a return stream from it. This is achieved by taking a trade exit and is not a feature of the portfolio itself but a feature of the trend following system that creates that return stream.

So, if you have a portfolio compiled of trend following systems where each system includes a stop and trailing stop condition, then by trade exit, the risk associated with a single return stream in a portfolio is released. We liken this to a release valve of a pressure cooker whereby risk (aka steam) is periodically released from the portfolio to lower the summated risk that resides within.

However, if you do not have risk release valves in your systems that make up your portfolio, then risk remains, and in a non-stationery market where correlations can alter and dilute the risk offsets, adverse risk can build to a point where ultimately, without risk release, we can have portfolio collapse.

Diversification – A Method to Take Advantage of Beneficial Volatility

Now we have already discussed this principle at length in our prior Primer so we can keep this brief.

Just as we use diversification to provide correlation offsets that reduce overall adverse volatility  across the return streams of a portfolio, we also use diversification to hunt for outliers and deliver correlation benefits that magnify geometric returns.

Now the reason for highlighting this little nugget and why it is only applicable to trend following is based on the design principle of our trend following solutions. You all know that we ‘cut losses short and let profits run’…but the deeper significance to this statement is it creates an asymmetrical reward to risk profile that allows us to capitalise on outliers without exposing ourselves greatly to adverse risk.

Under diversification, using our asymmetric systems, we can then hunt for outliers and also benefit from any positive correlation that may exist across return streams without exposing us to left tailed events. We can only do this when possessing asymmetric solutions that leave ourselves open to positive outliers while we cut our losses quickly in times where left tailed events can create adverse risk events.

Many traders shudder when we select return streams that clearly are correlated, but the reason for our confidence is that our system designs are going to save us from the tail risk present in trading a correlated series.

So, in addition to the benefits received by diversification in hunting for outliers over single return streams, we now have a further benefit not afforded to alternative trading mechanisms where we can be bold and trade quite highly correlated return streams. This is the reason for why trend following Fund Managers perform well during ‘crisis alpha’ periods where all markets start exhibiting correlated behaviour across asset classes. Trend followers can confidently trade within this highly correlated context, by virtue of the risk mitigation mechanisms they deploy within each return stream that exists in their portfolio.

Trends can arise from influences inside or outside the financial system. For example, when arising from inside the financial system, being diversified allows trend followers to capitalise on opportunities close to the source of the instability and then also benefit from the domino effect of that impact as it is spreads across the financial sector into other asset classes. This creates a favorable sequence of trending events that are easy to trade using diversified trend following systems.

Trends may also arise from outside the financial system such as the corona virus for example. In these instances all asset classes can feel this immediate impact, but having a diversified suite of short term systems allows many diversified trend followers to capitalise on this immediate impact. Trend followers with long term systems, can feel the brunt of such an aggressive impact across the sector, however their risk mitigation mechanisms allow them to reduce their overall exposure when compared to many alternative trading methods.

Remember that Trend Followers do not ‘warehouse risk’ in our return streams. We use risk mitigation mechanisms to ‘release risk’. As a result, we find that at the portfolio level we usually are optimally configured to take on additional future risk with our portfolios. So, when markets get nasty, and convergent methods are being ‘squeezed’, the trend followers can confidently take the other  side of the trades without exposing themselves in the same way to tail risk and in doing so make ‘crisis alpha’.

Outliers for any single return stream can lie in isolation or they can cluster during periods of ‘crisis alpha’ when the inter-related nature of markets all start to sing in tune. We can harness this power of angry markets to accelerate our geometric returns.

Diversification- A Method to Increase Trade Frequency when Targeting Tail Events

One of the issues surrounding trend following, when trading fat tailed environments is that  the trade sample size for each return stream is usually very low. We are very selective in when we decide to participate in a trending condition and when we latch onto an outlier, we could be riding it for many years. As a result, each return stream may only comprise a handful of trades.

To overcome this issue, and allow us to possibly participate in many more outliers of materiality, diversification across return streams allows us to achieve this outcome, provided that we are prepared to apply very small $risk bets to each solution.

Given our ability to trade small position sizes in these modern markets, and particularly when trading leveraged instruments where we only expose a fraction of our capital to gain exposure to these derivatives, then we can achieve extremely wide diversification under a diversified portfolio.

This therefore allows us to participate in many more tail risk events than you might think.

Applying Equal Sized Bets Across a Portfolio

Now given that we are targeting outliers that are non-linear in nature, and have the ability to dwarf the majority of returns in a trend following portfolio by their sheer magnitude, we tend to be more ambivalent in which markets we select as suitable candidates for trend following portfolios.

We simply require the following pre-requisites:

• That we only trade liquid markets;
• That we keep our bets small; and
• That we prefer a balance of market representation across asset classes of ‘different character’.

The first prerequisite is more a statement regarding risk management. With illiquid markets, we could find that a single return stream may be exposed if our risk mitigation mechanisms applied at the return stream level are not obeyed during chaotic market conditions. Sometimes, despite our best laid plans, markets can be sufficiently illiquid to fail to observe our stops and trailing stops from major gaps.

The second pre-requisite assists our first prerequisite ambition of risk mitigation, as a small bet means that even if our risk measures are not observed, it is unlikely to result in a material loss at the portfolio level. So for example during the Swiss Dollar de-peg in early 2015 where the Swiss currency pairs became highly illiquid resulting in sudden material moves, if we were incorrectly positioned for that move, then the adverse impact on the portfolio would have been large for a small bet, but not sufficient to cause total portfolio collapse.

The final pre-requisite is simply a recognition that even though we can go very far in diversification across liquid markets, there will always be capital limitations that finally put a cap on the degree to which we can diversify. Having an equal representation of return streams across asset classes is simply a statement that,  when faced with having to make a choice, we elect to not expose our portfolios to possible adverse risk events that may arise in a particular asset class.

Given our general ambivalence towards whether or not a liquid market is suitable for trend following, we apply equal risk bets in dollar terms for each return stream in our portfolio. We do not apply any predictive methodology that assigns a greater weighting to a particular class of return streams.

We feel that undertaking this process might make sense for a predictive mindset that sees some markets as being better trending markets than others…however as you will know by now…that kind of predictive talk for a trend follower are fighting words.

Non-linear fat tails can happen “anywhere and anywhence”. We are not seeking ‘normal trends’ that may arise more frequently in certain markets than others given each markets characteristic volatility. We are seeking material events of non-linearity that are known to occur in any liquid market when ‘sudden’ un-predictable events of transition just happen from time to time.

As we apply equal $ risk bets to each of our return streams, we therefore need to normalise our dollar risk that we apply to each return stream to reflect their normal volatility. We do this by using Average True Range (ATR) methods that ensure that when the bets are laid, an equal $risk is applied for each bet that we lay on  the table, and then we just watch and wait to see what happens.

The magnitude of an outlier can be significant, and can dwarf the average true range of volatility seen in the normal conditions of the market. The materiality of the event therefore means that we can get really small in our bet size for each return stream and still benefit from these windfalls when or where they occur due to their scale magnitude.

While we place thousands of very small bets through our method, as observed with our trade sample size for a diversified portfolio, you can see how the outlier trades that make the windfall are almost ‘scale independent’ to the rest of the trades undertaken. Wins of 25xR or even 60xR deliver king hits to the portfolio in terms of ‘step ups’ in the equity curve and we want as many of those ‘king hits’ as we can get our hands on.

Now going ‘hand in hand’ with this ability to achieve extreme diversification using many small bets applied to hundreds, or even thousands of return streams, we can only achieve this under systematic application.

Systematic versus Discretionary

Aside from our inability to back-test discretionary systems, one of the major issues surrounding discretionary trend following is our inability to achieve wide diversification under this principle.

There is no limit to the extent of diversification that we should strive for under a trend following philosophy given the added benefits that diversification brings to wealth building for trend following. Consequently, having thousands of separate return streams in a diversified portfolio is what we personally should be striving for if we really want to trade the tails.

But there is no way this ambition can be achieved under discretion. So this insists that if we want to take diversified trend following seriously, then we must be systematic in our processes.

Diversification – A Method to Capture Many Different Forms of Trend and Trend Segments and a Method to Offer a Unique Point of Difference

We have already discussed the many forms a trend can take outside the Gaussian envelope in Chapter 1 of this Primer, so there is no point laboring on about this principle here.

However we just simply need to realise that system diversification allows us to capitalise on the many forms of possible trending conditions that can exist in outliers of material significance. Having system diversification allows us to capture as much of the outlier that is possible in these exotic environments.

One of the beneficial aspects that we have not discussed however is that diversification also allows Trend Followers to have a point of difference between each other. If we were all doing the same thing, then the space would become quickly crowded. While Trend Followers are typically uncorrelated with other investment methods (such as Buy and Hold in Equities), there is considerable return dispersion with the grouping given the very broad diversification amongst our sector. This is a good thing for the long term sustainability of the Trend Following Fund Management industry.

Finale on Diversification

So here we are. Hopefully I have convinced you that under a trend following philosophy there really is no limit to the benefits that diversification can bring to wealth building. There is no such ‘marginal benefit’ limit in our view from diversification as we target the fat tails of the distribution of market returns and not the predictable peak of the market distribution. This makes our approach different to ‘predictive methods’. Diversification gives us added benefits in this ‘uncertain region’  and is an essential aspect to include in your workflow processes used to create your diversified systematic trend following portfolios.

Now you will hear in the literature about how Fund Managers use the correlation statistic as a basis to compile un-correlated portfolios, but in our next Primer I will discuss how many of us trend following folk feel that the correlation statistic is a very blunt tool that fails to address all the considerations we need to face in delivering diversification benefits to a portfolio.

We will look at the issues surrounding this blunt tool and provide alternative methods to achieve improved outcomes.

Chapter 10: Correlation Between Return Streams – Where all the Wiggling Matters

What a strange heading for a discussion on correlation. What the heck does ‘wiggling’ have to do with things? Well stick with us in this Chapter and you will find out.

In past Chapters, we have spoken at length about the power of diversification in balancing risk across a portfolio, and in allowing us to take full advantage of fat tailed environments to accelerate geometric returns. Well, the fundamental principle surrounding these diversification benefits lies in how the return streams of a portfolio move in relation to each other. How they co-vary or correlate as a composite.

If we visually think of return streams as wavelike phenomena whose volatile equity curves range between periods of building drawdowns inter-dispersed with periods of high-water-mark then we could imagine a physics experiment in the bath-tub whereby wave interference (between return streams) amplify or dampen the volatility of the entire portfolio.

Investigating Wave Dynamics in a Portfolio

In the text books you will find that correlation is a statistical measure that expresses the extent to which two variables are linearly related. The Pearson Correlation Coefficient is the popular statistic we use to assess the covariance of two variables (a description of how they more or co-vary together divided by the product of their standard deviations).

So in a nutshell, a single statistic is generated ranging between -1 to 1 that provides a generalised description of how related these movements are. Two variables are said to be perfectly negatively correlated (move in opposition) when the Coefficient is -1, uncorrelated (no related movement) when the Coefficient is 0 and perfectly positively correlated (move in tandem) when the Coefficient is +1.

So if we apply this to a comparison of return streams, two return streams can move in concert together (positively correlated eg. approaching +1), or be diametrically opposed (perfectly anti-correlated) in their relative movements (approaching -1) or be uncorrelated in nature (approaching 0) where the movement is not related.

Now while the degree of correlation is a sign that perhaps two return streams may have some fundamental connection between each other (aka they may be causally related), there is no guarantee that this is the case. Correlation does not imply causation. It may simply be the result of how the two different return streams play out with no causal connection between them. However when you have a subset of strong correlated series to work with, you can be pretty sure that there is some fundamental relationship between the return streams….but never 100% positive.

So let us have a look at this general principle in action. Chart 31 below describes the relationship or the covariance between two return streams, namely the Societe Generale Index (SG Index) and the S&P500TR Index. The SG Index is a composite of trend following programs and the S&P500TR Index is a useful way to describe a buy and hold Program in US Equities. Chart 31 is therefore used as a basis to showcase the very uncorrelated nature between a long only Buy and Hold approach in Equities versus a typical Trend Following Program.

Now the correlation statistic produced is very close to 0. In fact it is slightly negative at -0.09 so this implies that the two return streams are not related. Well this is the information imparted by this single statistics…but look again as the devil once again is in the detail.

Chart 31: Uncorrelated or is it?

When we drill down into the detail by comparing paths of each return stream over the same history, we find that at times the relationship is strongly positively correlated (blue ellipses), strongly negatively correlated (red ellipses) and at times, simply uncorrelated (yellow ellipses). It is the sum total of these discreet elements of time series that produces a low overall correlation statistic between the two return streams, but not a faithful description of how the two return streams actually move together.

It is therefore a blunt instrument and fails to tell us important information that is essential for our trend following pursuits. We need to know when a portfolio can possess risk weakness or when a portfolio can achieve accelerated returns or when a portfolio is simply stagnating  (spinning its wheels).

You see, we trend followers do not use a single correlation statistic to define a directional relationship between a single return stream and alternative return streams. A correlation statistic itself does not include any information about the adverse and beneficial relationships that exist across an entire time series. We therefore assess how each return stream assists or detracts in their inclusion into a portfolio across the time series by mapping the ‘moving relationship’ over the entire history of both equity curves.

By visual mapping we are looking for a number of features. Do equity curves rise or fall under differing market conditions. What we are looking for here is a direct causal relationship between favorable trending conditions and a rising equity curve or unfavorable market conditions and a falling equity curve. We also use a visual mapping process to stress test a portfolio.  Where is there a weakness in the portfolio, and where are the windfalls. A single statistic does not tell us this story. A detailed mapping process of the entire return stream of a portfolio and it’s component return streams does offer us this valuable information.

Correlations are rarely static and change over the course of time, but a single statistic will not convey that information. All the wiggles matter from a trend following perspective in delivering powerful geometric returns.

We have found previously where single statistics do not help our cause as trend followers, such as the simplified use of the positive expectancy equation in Chapter 6. When trading out in the fat tails, we need all the details. Our world is a different domain to most alternative trading methods and it is the fine detail that can lead to “Bloom or Gloom”.

In fact, there is a little used term which is frequently swept under the carpet that is even more powerful for a trend follower than correlation. This is a principle called co-integration. This principle is tied to how we design the systems within our trend following portfolios to give us a more static relationship between return streams. as opposed to a moving feast from correlation, upon which we can then generate superior geometric returns. But before we get to co-integration, lets understand correlation a bit better.

Correlation

Correlation describes best what can be achieved through diversifying into different markets and into different timeframes. There will be certain times when assets are more correlated with each other or not….but there are no guarantees of this persistent relationship. In fact, correlation when explored within non-linear market conditions needs to be very carefully considered.

The adage ‘correlation does not necessarily imply causation’ needs to be kept front of mind with adaptive emerging market conditions. The inter-relationship that exists within the nested systems of financial markets means that at times, causation can rear its ugly head leading to varying correlation relationships. A panic sell in one asset class frequently leads to liquidations in other asset classes to fund these shortfalls and before you know it, correlation goes to 1 across asset classes.

There is no better investment class to observe this feature as the equity markets. During long protracted bull runs, individual stocks tend over time to be less correlated, whereas during market crisis, this relationship evaporates where a sea of red across the entire market is attributed to a very high correlated relationship within the asset class itself, and its impact is also felt across other asset classes.

In terms of diversification at the timeframe level, we all know that trending markets are not simple linear ascents or descents and tend to exhibit a fractal wave nature of persistent overshoots and undershoots in its overall trajectory. The fractal-like nature of markets means that we can find outlier directional moves within any timeframe  and that ‘outliers’ are scale independent. Diversifying into different timeframes therefore allows the divergent trader to benefit from long or short trends arising in any timescale.

Of course, we prefer the higher timeframes (aka the ‘longer look-backs)’ than the shorter timeframes (or the ‘shorter look-backs’) to apply our trend trading process. It is not that there is an absence of outliers in any particular timeframe, but rather it is the influence of noise and mean reversion in the shorter time-scale that significant alters the form of the outlier and makes it very difficult to trade. So this is not a ‘markets don’t trend at shorter time intervals statement’….but rather that ‘trends within shorter time intervals are far more influenced in their general character by noise and mean reversion type of statement’.

Co-integration

Now as excited as we get about correlation , nothing picks our ears up more than the term co-integration. To obtain cointegration benefits in a portfolio is like the ‘Holy Grail’ for trend followers.

Not many traders talk about co-integration, but they should. There is a degree of confusion expressed by traders in understanding the difference between correlation and cointegration but here is a simple description that seeks to clarify the distinction.

• Correlation – If two stocks are correlated then if stock A has an up day then stock B will have an up day
• Cointegration – If two stocks are cointegrated then it is possible to form a stationary pair from some linear combination of stock A and B.

So here is an example that might allow us to see this difference better which is taken from the following source: gekkoquant.com

“A man leaves a pub to go home with his dog, the man is drunk and goes on a random walk, the dog also goes on a random walk. They approach a busy road and the man puts his dog on a lead, the man and the dog are now co-integrated. They can both go on random walks but the maximum distance they can move away from each other is fixed (ie. length of the lead).

So, in essence the distance/spread between the man and his dog is fixed, also note from the story that the man and dog are still on a random walk, there is nothing to say if their movements are correlated or uncorrelated.

With correlated stocks they will move in the same direction most of the time however the magnitude of the moves is unknown which means they can arbitrarily change and start to mean revert. This is in contrast to co-integration where we say the relationship is “fixed” and that if the relationship deviates from the “fixing” then it will mean revert.”

Now let us discuss ways we can deploy this principle under system diversification.

Think about what is achieved through a perfect hedge. For example, we buy 1 lot of a particular instrument and we sell 1 lot of the same instrument at the same time and hold for a defined period.

What we are achieving here is using a system-design principle to lock in a forced co-integrated relationship between the return streams of each position that never changes over the course of the trade event.

Of course, this example in real application is a ‘fool’s errand’ as the frictional costs associated with taking these trades lead to a negative overall sum game….however it is the design principle that is important to consider here.

So now let’s assume that we have a divergent trend following strategy that is not permanently ‘in the market’, but only enters the market associated with a particular trade signal. This strategy, in isolation, demonstrates a small edge over say 20 years or more to give us confidence in the method, but it has a 30% win rate with a risk/reward of say 1/3 to deliver positive expectancy.

Let us call this strategy a core strategy of a systematic portfolio.

The 30% win rate of the core strategy gives hope to a system designer that there “might” be an alternate ‘divergent’ strategy that could be designed which enters at exactly the same moment as the core strategy but with the opposite directional sign and also possesses a positive expectancy over the same data series.

Remember that a 30% win rate means that somewhere there is 70% of unsuccessful opportunities that might be available to an inverse strategy relationship.  But unlike the theoretical hedged position, you have not created a perfect hedged condition, as the open profit conditions of both systems vary but you have created a co-integrated anchor point at the entry with oscillation about this.

Imagine now how these two return streams co-vary together. There is a stronger causal connection between them and when one system is in drawdown the other will be reaching for the skies.

While the perfect symmetrical relationship is not theoretically possible with the inclusion of the frictional costs of trading, you will be surprised at what less than optimal design-solutions can achieve at the portfolio level.

It is through system design that you can go further in your quest to ‘force co-integration’ benefits into your portfolio with less emphasis applied to the vagaries that simple uncorrelated systems can produce.

Once we can understand this pearl of wisdom, we can now confidently state why we as trend followers insist that we adopt both long and short positions with all return streams that we compile into a portfolio.

It is not because we believe that all markets offer both long and short opportunities equally. We recognise they don’t, particularly with markets such as Equity Indices that have a long bias built into them.

The real reason we are prepared to sacrifice this ‘long bias’ and insist that we trade both long and short is to benefit from the co-integration that exists in a portfolio by this bi-directional stance.

Sufficient numbers of both long and short return streams in a portfolio guarantee a degree of co-integration in the portfolio which is far better that the moving feast that may exist under simple correlated relationships that change over time.

Well that is our argument for why we insist on trading long and short across all markets.

There will be a raised eyebrow or two relating to this statement in our camp, but we all have our different interpretations about trend following philosophy.

Now we are reaching the end of our philosophical musing about trend following and will shortly be putting the philosophy into action in this Primer where we demonstrate a workflow process used to design and diversified systematic portfolio from scratch, but before we do…I promised a chap on Twitter that I would pinch his term he uses to refer to Trend Following methods. The term called ‘Pain Arbitrage’.

In the next Chapter, we will have a look at what is meant by this, and why “pain can be a gain” for trend following folk.

Chapter 11: The Pain Arbitrage of Trend Following

We are a weird mob that just don’t seem to apply the same rules to trading as other forms of market participant.

When compared to predictive convergent trading methods, who base their success on instant gratification by targeting a current repetitive market condition, the ‘trend following way’ can be a real drain on the psyche.

We base our philosophy on a market principle that ‘trends can persist’ rather than a predictive pattern, and this distinctive point of difference can take a very long time to play out. Given the inherent tendency of markets to waver between periods of predictive stability only occasionally inter-dispersed with periods of trending transitions, we are wrong most of the time. So we simply switch our brains off in their predictive tendency and apply our price following processes. But, if you want pain, then you have come to the right place if you think with a ‘predictive mind’.

Now I know that you all want this equity curve below of a famous trend follower depicted in Chart 32…but I am going to tell you now, that you cannot have it. Not until you change the way you interpret pain, which means the way you think about how the market behaves. Pain is an illusion associated with a predictive mindset. Let us explore this further.

Chart 32: Dunn Capital Management World Monetary and Agriculture Program (WMA)

You see, what comes with this famous equity curve above is the very necessary pain that is required to achieve this outcome.

Look at Table 12 below. It is a pain summary for Dunn. I have highlighted the painful periods that are a necessary pre-requisite to achieve that glorious equity curve.  A blow-by-blow account of enduring pain experienced over long periods of time only periodically inter-dispersed with periods of jubilance.

Table 12: Dunn Capital Management WMA Monthly Returns

You can see from this example that the Trend Following way is therefore actually a lesson in pain. A lesson that can only be tolerated if you change the way you perceive pain. This can only be accompanied by changing the mind.

Changing the Mind to Embrace Pain

Our process flies in the face of human intuition and is an exceedingly counter-intuitive approach. But it works. In fact, it works so well that there are few rivals in our path to long term wealth but to achieve this ambition, we must learn to embrace pain. To achieve this, we need to change our minds by removing our very natural desire to predict the markets, and to simply go with the flow of it all. It really is as easy as this….but to achieve this, is an incredibly difficult task for an obstinate brain.

Whenever you meet something that is foreign to your mind, you experience pain. Pain is the brain’s way of telling tells you that something has gone wrong and that all is NOT well. Whether that is sticking your finger in a flame, losing a loved one, experiencing frustration, anger, etc……

But we also experience pain when trading these markets. Quite literally. Not the normal type of pain arising from the way we normally think about it, but the pain associated with anxiety, stress, impatience and financial setbacks. This is simply a small sample of some of the symptoms of pain we associate with trading the markets.  All this pain is associated with the pain arising from the brain getting it wrong. It simply screams out…”What is going on. This is the way I interpret these markets and they just aren’t behaving in accordance with my predictions”.

The pain all lies all in our brain and is deeply connected with our desire to predict as participants in a complex system where the complex system has the final say in the matter.

When you commence your trend following journey, be prepared to tolerate pain. Nearly all will fall by the wayside as they fail to take this bull by the horns, but a very few may survive. Not by confronting pain head on like an obstinate bull, but by learning to accept it as a natural consequence of trading the markets.

It requires a ‘mind reset and a re-wiring’ that over a long time with experience in the market will change the way you ‘feel about trading and these markets’ and indeed your life too.

With an altered mind you will no longer interpret drawdowns as ‘pain’ but a necessary consequence arising from hunting for outliers. You will no longer feel the wrath of being wrong most of the time as your risk bets will be naturally configured to trade small leading to ‘midge bites as opposed to massive failures’.

You will find that there is literally no point in listening to the news, hot gossip or staring at the screens as you have a systematic process now that avoids all that, forged from your intense research undertaken before you commit to your live trading experience, and you will find that under systematic application of boring repetition, you have available time to enjoy other more personal pursuits in  life that mean more to you, your partner and family.

You will undergo a metamorphosis as your brain is re-wired to start enjoying life as opposed to fighting it….and then, and only then will you call yourself a ‘trend follower’ as opposed to a trader.

That is the narrative of this Primer. If you want to experience pain, then be my guest and commence your journey into trend following, for it has pain in spades as interpreted from a predictive viewpoint. But in learning to deal with the pain, start to understand that pain is simply an illusion arising from a certain kind of mindset tailored for a different environment.

What is this Pain you Speak of Pilgrim?

There is no ‘fundamental pain in these markets’. It all lies in the mind. Learn to reinterpret this illusion called pain as a natural consequence arising from how we must trade these markets to obtain those glorious equity curves that currently exist just beyond our reach.

For once we learn how to change our mind from experience and learning , then you will find that the trend following processes we apply, no longer holds pain in the historical record. Pain is now medicated for and is viewed by the trend following mind as simply a very natural feature that allows us to achieve our expanded ambitions in both trading and in our lives outside of trading.

Pain Arbitrage

A dear Twitter friend the other day reminded me of a very useful expression that many traders apply to trend following. He referred to the “Pain Arbitrage of Trend Following” suggesting that the reason for the enduring nature of this robust technique, was explicitly connected with the pain that is endured by participants who apply the trend following method.

Such a useful way to describe trend following and so elegantly summed up. Trend Following appears to have such enduring arbitrage due to the pain associated with a certain type of mindset, namely a predictive mindset. The metaphor proposes that predictive methods deliberately avoid pain, whereas trend following methods actively exploit it. Indeed, he may be right.

In this zero-sum game, trend followers’ prey on fruits arising from the predictive mindset when their models break down. So, if predictive models are less painful than trend following methods, then it is natural to conclude that ‘for a predictive mindset’, trend following is thereby more painful.

We all fall victim to a Predictive Mindset

The premise of trend following lies in exploiting unpredictable phenomenon that are known to arise from time to time in the tails of the distribution of market returns.  So how do we exploit this unpredictability with a predictive mindset?

We need to change the way our mind works. This unpredictable element to trend following forces our approach to be the exact antithesis of the more appealing approaches that are preferred by a predictable mindset.

We all suffer from this predilection of prediction as our brains have been moulded over deep time to behave this way shaped by a different complex system, Namely, our natural environment. We therefore all fall victim to the ‘predictable mind’.

But the associated pain we experience when trading a different complex system, namely the financial markets, all arises from this mindset forged in a different domain.

To pinch an eloquent quote used by a Trading legend and Twitter friend who I deeply admire,

“Bubbles are the inevitable outcome of human nature. What part of human nature? We are the ape that imitates. We are the ape that seeks status. We are the ape that tells stories. We always prefer narratives versus facts.”

This whole Primer series that you may have been reading is a purpose-built narrative used to appeal to apes that like stories. It is not designed to appeal to the trend following mindset as it is preaching to the converted, but it is explicitly appealing to the predictive mindset.

It is providing a narrative that hopefully convinces the predictive mind that there may be a better way to approach being a participant in a complex system. A way to turn trading from a stress filled, ego-filled and anxiety ridden enterprise into a process of enjoyment, humility and acceptance.

The tortuous road that a predictive mind may have already experienced in reading this narrative has been a deliberate ploy adopted by me ,as the writer, to perhaps sow the seeds of change in your mindset. To possibly convince you of how ‘wrong’ we stupid apes can be in navigating these financial markets.

The markets are not a get rich quick scheme or cashflow generating solution. This is rarely the case. The ‘predictable moments’ of a stationery market are an ephemeral phenomenon that do not contribute to the greatest price movement. Indeed, the greatest price change arises from the anomalies arising from uncertainty itself that can only be exploited with a long-term wealth building objective.

Markets are exceptionally complex and adaptive systems comprising humans and their algorithmic creations whose partial knowledge embedded in a human mind make them open to large errors in predictive logic when navigating uncertainty.

We are all victims of our own minds. Without this necessary narrative, I cannot expect a reader to jump on board the trend following wagon. But I encourage you to expand the ‘blinkered and limited mind’ to see the trend following process for what it really is. A method that is purpose built to capitalise on market principles that exist in this foreign context from which our brains are not sufficiently equipped to deal with.

It is so hard to convince a human mind to jump on board this ‘gravy train’. Despite the exceptional long term track records for trend following, that everyone wants, when it comes to the very necessary ‘grind’ required to achieve these beautiful equity curves, no one wants to endure the associated pain to achieve it.

Our Brains are not Well Equipped for Trading the Markets

We want it all now, as we are very impatient apes and sucked in by a predictive mindset forged for a different environment. In that environment, our brains have evolved in the most efficient way to go out and get it now through predictive logic that beats the slow reaction time arising from having to wait to receive all required data before we make a decision.

Unfortunately, as traders we play in a very different complex system to the one, we ‘efficiently’ evolved in. This new system we find ourselves playing in is an ‘intelligent market’ full of predictive mindsets and not so simple creatures. If we simply want to go out there and get it now, we are no longer dealing with fruits, vegetables, and the occasional mammoth. We are now dealing with other humans who are just as smart or even much smarter than you. The different regime means that we now find ourselves in a battlefield in which our minds have not evolved to deal with.

We explored in Primer 2 how our brains have evolved inside a different complex system which has allowed us to use illusory heuristic shortcuts that have evolved to meet ‘fitness targets’ as opposed to the reality. We found in our exploration that there is no such fundamental reality of sight and sound. Sound is a symptom of translating air pressure in an eardrum, and sight is a narrow rendition of what our ‘blind’ brains interpret as the reality….What we actually experience is a heuristic illusion rendered between an observer and the observed designed to allow us to survive in a very complex system.

But our brains have never ‘evolved’ in this new complex adaptive system that we call the financial markets. We therefore experience ‘pain as a symptom of suffering by the brain’ in interpreting this foreign arena.

Life for a successful trend follower is characterised by the pain of persistent nagging drawdowns only periodically inter-dispersed with hedonistic joy associated with rapid accelerations to new high-water marks. These rare ‘joyful’ moments are jubilant times but only fleeting in nature…and then it is back to the grind. Day after day….after day…after day.

The Transformation of the Mind that is Necessary to Appreciate Trend Following

When you start of in this game of trend following with a predictive mindset you inevitably incur years of ‘mistakes in application’. This unfortunately is a prolonged but an essential side effect of trend following. It is through these mistakes, that a mind starts to re-wire and transform which allows you to drop any ingrained predictive logic and learn new ways to cope with this pain.

None of us are ‘born trend followers’. We must condition ourselves to become one, through the school of hard knocks. It is through these mistakes that we learn the nuances of our craft and develop the fortitude to keep on keeping on. Persistence and patience is essential but you must participate in the game. You cannot afford to bet big when you start this game. It is essential that you start with baby steps.

Never give the day job away until you can truly validate your track record and build your trading capital to a point that can endure your worst-case scenarios.

As the months and years lapse from this gruelling process, it just starts to get easier, slowly but surely as the brain re-wires.  Before you know it your hardened mental character now starts to view your approach as just a boring repetitive process that just ticks along. Your drawdowns now become just a means to a new high watermark. The boring pursuit then allows your altered mind to see things differently and wider.

With an epiphany you start to realise that trading is just a very small aspect of a bigger system called life. Where your now boring trading life can be filled with other more important pursuits such as spending more time with family and friends, taking up new hobbies and passions and leading a more fulfilling life. Trend following is a process that encourages you to diversify into other aspects of life.

At this point in time, after all the re-wiring, it then just all clicks Your brain literally metamorphoses turning you from a trader into a trend follower.

Trade the Approach that Suits Your Personality….What????

Now you might hear the sage advise from trading mentors that you should trade an approach that suits your personality. Well, I am here to tell you that such advice does not apply to us battle hardened trend following folk. Predictive methods yes, but trend following methods then no.

It begins as ‘pain central’ for us, but the only thing that makes us stick to the plan in the beginning is our understanding that with no pain, comes no gain.

It is only by conditioning through years of experience with this gruelling process that we can eventually come to accept it.

Given the extreme pain threshold that is a necessary pre-requisite to pave our way to glory, some view this sado-masochistic quality of trend followers as ‘pain arbitrage’, and the reason for why this method has stood the test of time without being arbitraged away through too much ‘crowding’ in this space.

Some Tips that Might Help to Rewire that Brain and lessen the Pain

So how do we tolerate this pain early on in our trend following careers? Here are a few pointers that at least work for me. Of course, every mind will be different and at different stages of development but some of this advice may resonate.

• Faith in the method. A faith that can only be bestowed through examining the long-term performance records of the Top FM’s in this field, listening to guidance from these experts and then rigorously testing the assumptions yourself through years of repeated back-test application before starting your foray into live trading with trainer wheels on using a small amount of risk capital.
• Test all assumptions yourself and never rely on hearsay. Be always sceptical and learn to discern between fact and fiction. Learn to trust your own capabilities and up-skill yourself to continuously enhance your skill sets and self-reliance.
• Embrace drawdowns as they will be an enduring feature of your chosen method and a necessary symptom for wealth building. Always refer to your back-tests as a road-map when things get tough.
• Never use a backtest as a basis to forecast future returns. A backtest is a method you use to assess risk weakness and define risk benchmarks to observe in the future. If you reach these thresholds, then stop trading and get back to the drawing board.
• Commitment and perseverance is essential in this game in the early years. Devour books and podcasts on the subject from trusted sources. Trend following is a persistent learning exercise but it does get easier over time.
• Keep a close eye on industry performance. Continuously assess your performance against industry benchmarks. You must accept that your performance will lie below the long-term metrics of the best professional FM’s in this game. Keep away from the marketers and charlatans that are rife in the industry by being able to discern between realistic expectations, pipe dreams and sales/exploitation tactics.
• Build a support base around you of like-minded traders and mentors who know this game and defer to them frequently for advice, troubleshooting and friendship and keep your nearest and dearest informed of your progress and stay fully transparent with them. Having a supportive environment is essential for trend following. As tough as you may feel you are, there will be times when your support group will be pulling you out of the mire.
• Only risk what you can afford to lose. Separate your risk capital from your livelihood. You will be beset with issues when you start this game so expect it. Experience is an essential pre-requisite for a trend following career.
• Continuously reinvest your profits back into this game. Compounding is the way we build our wealth. The systems we deploy and the edge we enjoy are just a means to this end.
• Systematic application. Your rules-based process must be automated so you can validate your live performance against your back-tests (your trading plan) and build trust in your systems.
• Meditative Methods, Sleep and Fitness. Practicing forms of meditative methods, keeping to a good sleep routine and staying fit and healthy to reduce anxiety.
• Commencing your trading journey with very small bet sizes that allow you to sleep easy at night and reduce the temptation of over-riding your rules-based process with discretionary judgement. It really helps to keep leverage small in the beginning that makes your method become a really boring process. Over time you may find that your pain threshold improves allowing you to progressively scale up your applied leverage….while still sleeping well at night; and
• Listening to Ed Seykotas ‘whipsaw song’ every night before bed.

If you Stand on the Shoulders of Giants, you Still Need to Change your Mindset

Of course, you can avoid a great deal of the pain by standing on the shoulders of the giants in the industry who have already endured this path of pain. But this assumes that your mind has already been pre-configured to accept this challenging path. This is not necessarily the case and is a central issue surrounding a trend following Fund Managers dilemma in writing their monthly reports to investors.

Investors, like traders who practice the trend following method need to also undergo a significant re-wiring to stay the course in this journey.

Concluding Remark

This Primer is specifically tailored to offer a way forward for traders and investors seeking to change the way they interpret these markets and turn us all into Trend Followers. Direct experience in applying trend following practices is the ultimate training ground to train a foreign mind, but at least a narrative provides a way for a predictive mind to say……

“Hey you guys are really a weird mob, I wonder if there is any substance to the words you say? Most of it is just gibberish”

Planting the seeds of discontent for an enquiring mind is the objective of this Primer series. It might make us see the forest for the trees. Embrace the pain to experience the gain, in all facets of life.

With an expanded mindset you may just see that that the garden of ‘painful’ thorns is actually full of beautiful smelling roses.

Well, that wraps up the philosophy side of trend following. I hope that I have paid it justice. There is just so much too it when you dig deep into the weeds. It is full of interpretational aspects that may not resonate with others that practice this approach, but we never said we all see eye-to-eye on these important matters relating to our craft. But we now need to get stuck into an applied process that transforms this theory into practice.

In the upcoming Chapters of this Primer series, we are therefore going to put the philosophy into action. The verbose narrative starts to dissolve into practical application where we develop a diversified trend following portfolio from ground up.

Chapter 12: Building a Diversified, Systematic, Trend Following Model

So here we are. In this trend following story so far, I have laid out a philosophy about how a  certain Trend Follower believes that a liquid market behaves, but now I am at the pointy end of this exercise where I need to now validate my ‘hypothesis’ using a rigorous systematic process that assesses what the Market thinks about this ‘crazy idea’.

I want to examine how close my trend following interpretation is to how the market might behave. While digging into the weeds using quantitative methods to assess the degree to which this interpretation may be correct, I need to keep in the back of my mind that the market does not care what I think. The market is far bigger than my interpretation. It is an emergent expression arising from what all participants think.

So, I need to face the ugly truth that I am never going to get it all right. A Model is just an interpretation as opposed to the Reality. I am just going to have a model that comprises a large rounding error. But through committing to a diligent scientific process, I plan to keep that rounding error between my model versus the reality as small as possible.

If I get a close match between my model and the reality, then I can be happy with my interpretation, and what is more, if this model can demonstrate that I can extract alpha from the market, then I could use this model to build my nest-egg.

So, to convert this process into one that is familiar for a trader, I am seeking to test the validity of my trend trading philosophy in a real market using market data.

Given that my philosophy extends to any liquid market, then the validity of the model needs to be tested under diversification. If I find through this process that my model has merit, when applied to historical market data, then I can convert this model into a diversified systematic trend following portfolio that at least has ‘fairly well’ described how that model has fared over history.

So here is the Hypothesis we will be assessing through this exercise.

“That we can construct a Diversified Systematic Trend Following Portfolio that, targets the tails of the distribution of market returns (both left and right), is adaptive in nature, and through its application across liquid markets can deliver powerful risk adjusted returns over the long term in a sustainable manner.”

Now remember that this hypothesis (model) is mute regarding the ability to extend this framework into an uncertain future as we have no future data to test this hypothesis on, so we are limited to historical data sets.

We therefore need to continuously undertake this iterative process again and again to continuously assess its validity with the addition of new market data, but at least this process ensures that we keep our models up to date….or our portfolios razor sharp.

In fact, this model for trend following is only part of a bigger model within which we apply constraints or model limitations, to narrow our experimental focus, so there is always room for making better models at many different scales. Revisions in one model, will require revisions in other models and so on and so forth.

So, you end up with a model that specifically responds to the narrow domain of Trend Following, which resides in a more comprehensive model described by complexity science.  It is just a complex nested system of models within models, and there is always room for improvement as we expand our domain of understanding.

The reality we face in interrogating this complex system we call the financial markets, is that this modelling we undertake is a never-ending exercise of ‘building better models’ that more faithfully describe the reality out there.

Just when you gift wrap your latest model that most faithfully interprets your reality, then it is back to this process driven exercise to do it all again with additional insights you have obtained from prior modelling.

In our complex adaptive markets, you need to continuously create better models. Our work is never done. But we strive to do the best that we can do.

Why the Scientific Method?

So, I now need to put this discretionary model of how I feel the market behaves to the test, and develop a diversified systematic trend following model that can evaluate ‘how suitable’ this interpretation may be in describing the reality out there.

I now need to describe a mathematical process called Lambda Calculus………..

I love being dramatic. We don’t have to go that far…..so come back all you mathematics haters.

But we do need to apply a rigorous systematic process under a ‘Sciencey’ method that can evaluate our hypothesis without letting that ‘pesky brain’ possibly bias what the Market wants to say about this model.

“What does Science have to do with trading?”, I hear from the bleachers. Well it provides a well worn method through a process driven approach that:

• outlines the central problem which the model seeks to address (or the hypothesis);
• provides a scope that limits the context within defined bounds (defines the assumptions); and
• then evaluates with rigor the validity of the hypothesis by putting grist to the mill and seeing what the data has to say about the matter.

Discretion is the Basis for All Models – As Much as We Think Otherwise

Now while our ‘hypothesis testing’ seeks to apply quantitative systematic rigor to the testing process, there is inevitably a degree of discretion in any method undertaken. The most obvious discretionary judgement is the hypothesis itself, which is a human interpretation of how a system behaves (a possible slice of the reality) and not the actual reality.

But there are many others that we will bump into along the way as we undertake our validation process of our hypothesis, such as our choice of universe to test (the limited selection of liquid markets we examine), the parameters we use in our system design etc. The list just goes on and on. It is ‘literally’ littered with discretionary decision making.

So as much as we try to objectify our method using systematic processes to try and avoid a propensity to ‘steer or nudge’ the testing result to agree with our hypothesis, we unfortunately can’t avoid it. This is a lesser problem than that faced by discretionary traders as systematic application does significantly reduce this propensity,  but in our method, we must ensure that we declare any discretion applied, by stating these decisions as assumptions in the method, so that we can revisit it another day if needed, to evaluate any bias that may have influenced the overall result.

There are also major issues when applying quantitative methods to data mining processes where we can easily ‘fit our conclusions’ to erroneous data. This can complicate our ability to validate our hypothesis. Worse still it can render our intensive validation completely worthless. We will be closely looking at these issues of ‘adverse curve fitting’ in a future Chapter and highlight the methods we use to reduce the impact of this curly problem for quantitative traders.

In our next Chapter we will introduce you to a Workflow Process that we use here at ATS as a methodology to validate our trend following hypothesis.

Chapter 13: A Systematic Workflow Process Applied to Data Mining

In our previous Chapter we outlined how we will be using the Scientific Method as a basis to validate our hypothesis (below) derived from how we interpret a trending ‘liquid’ market to behave.

The Hypothesis

“That we can construct a Diversified Systematic Trend Following Portfolio that, targets the tails of the distribution of market returns (both left and right), is adaptive in nature, and through its application across liquid markets can deliver powerful risk adjusted returns over the long term in a sustainable manner.”

We specifically narrowed our hypothesis to only assess:

• a particular market condition (fat tailed trending condition);
• for bi-directional trend following strategies;
• which are applicable to any liquid market (liquid and diversified); and
• that our trend model needs to be adaptive in nature (non-stationery).

Having such a narrow hypothesis helps us significantly reduce our data mining efforts and simplify our experiment, enabling us, subject to validation of our hypothesis,  to create sustainable adaptive portfolios.

Data Mining itself can be defined as a process of extracting information from large data sets from which we can then transform this information into a coherent structure for future use. This is exactly what we want to achieve through the testing of our hypothesis.

We want to mine historical data sets (a diversified array of liquid market data), to evaluate whether or not trades undertaken by adaptive systems designed to capture ‘fat tailed’ trending conditions have been able to generate sustainable returns.

This entire process that we undertake seeks to identify a possible causal relationship between the systems we deploy, which we define as trend following systems (but we could be wrong), and their ability to catch major directional anomalies (which we observe in the real markets).

The ‘proof statement of this hypothesis’ lies in our ability to generate an edge through our diversified portfolio of systems that is sufficient for compounding to then take over and then do its magic by generating long-term sustainable returns.

If we can demonstrate a ‘causal relationship’ beyond a mere random happenstance between our systems ability to extract an enduring edge which is suited to compounding, then we have validated our hypothesis. That is all there is to it.

Avoid the Tendency to Want More – It just ends in Tears

Our hypothesis is so simple and if we stick strictly to validating it, then no problems, however that is not what really happens at all when we get computer power behind us with our predisposition towards instant gratification.

We always want more.

We want the ‘best systems’ which boast stunning performance metrics. But answer me this….where in our hypothesis were we requesting the best outcome?

That was an extra layer that your brain instantly added onto this exercise and with it come associated problems. Namely unrealistic backtests and aggressive optimisation approaches that ‘murder the data’ simply to achieve the best results, not to test the validity of our hypothesis.

But wait…. there’s still more murder on the trading floor. Now that we have optimised and riddled our validation process with selection bias, we now suspect, rightly so, that we have murdered the data, so we then add further ‘data murdering tools’ like Monte Carlo approaches and Walk Forward techniques which attempt to remove the optimisation bias that we have now injected into our method. What a mess?

These unnecessary complications might give us what we want on paper, with a beautiful back-tested equity curve reaching for the heavens, but they are certainly not what the Market intends to deliver to us when we take these creations to the live environment. The Market just wants to give us a very painful lesson in how very ‘wrong’ we are.

Through this added layer of complexity that we have unnecessarily ‘forced into our thesis’ we now find ourselves so far departed from original hypothesis with which we wanted to validate.

We have explicitly stated in earlier Primers that we are only seeking a ‘weak edge’ as that is all that is needed for compounding to then take over and do the heavy lifting. Other more impressive quantitative scientists have already done the hard yards in the academic literature and have already demonstrated that a weak edge exists through their market studies.

But now given our ‘apelike tendencies where we want the banana’, we turn our backs on these empirical studies from our desires to obtain the ‘strongest edge’? Something smells like a “Charlatans trap” here.

Hopefully, you may now see the rut we fall into, when we steer away from our primary objective by letting our desires run rampant.

Avoid the tendency to want more than what the markets are prepared to deliver. It just ends in tears.

We are Seeking Proof of a Causal Relationship, Not a Statement about Future Expectancy

All we want is do is determine that there is a strong likelihood of a causal relationship between the trend following systems we use, honed from our philosophy about how the markets behave and that this causal relationship can be inferred from our portfolio’s long term performance record.

If so, then this correlated performance relationship between our systems/portfolio and historical market data has a better chance ,(than a random relationship), of enduring into an uncertain future. We can be sufficiently confident in deploying our portfolios into an uncertain future, where we suspect ‘fat tailed conditions’ MAY arise.

Future performance is primarily dictated by the Market and NOT our systems/portfolios. Our systems can of course turn this into a nightmare but optimal performance is capped by the market. Not an optimisation.

The best we can do is develop systems that respond to the Markets beck and call….and this relationship of causality between our systems and the market are not stationery and can vary over time.

The systems we deploy, must be able to adapt to respond to the dynamic relationship that exists between our systems and the market. Complex markets are constantly changing and adapting. So should our systems. The market trends that we have targeted over the historical record have changed. As a result, our workflow process must adopt an adaptive element into the procedural architecture.

So let’s get our helmets on and get back to work. Down t’ut pit we go.

It’s Off to Work we Go….with a Workflow

A workflow process can be defined as a sequence of tasks that process a set of data. It typically comprises a sequence of systematic processes that can be repeated and is arranged in a procedural step by step manner to process inputs into outputs, or raw data into processed data.

Figure 1 below describes a workflow that we use at ATS to validate our trend following models. Our process turns “in sample raw data into robust trend following portfolios with an edge”. The workflow sits snugly within our broader scientific method. Having passed through the workflow, we have a completed portfolio which we can then test over very long-range ‘unseen’ market data (referred to as Out of Sample Data OOS)

We try to keep our processes very simple.

I often have a giggle when trying to understand the quantitative mindset. They are very tough minds to understand as there is chaos in their process. While there is a general consensus out there in quant world, that it is best to keep systems simple, as this simplicity allows for robustness, the discussion then turns more ‘arcane’ about degrees of freedom and then before you know it the discussion heads into it the most complex and devious statistical add-ons to complicate the once simpler process.

In this confusing discussion about advanced statistical concepts and techniques, we are left scratching our heads in a quagmire of complex processes where huge gaps exist in the coherent train of thought which is meant to reside within our factory processes.

This is where a workflow comes to the rescue. It forces us to explicitly declare the processes we will be undertaking to test our hypothesis in a coherent logical manner, Step 1 leads to Step 2, leads to Step 3 in a logical procedural basis. If there are any queer complex steps in our process, we will see it in our workflow process. There just won’t be a coherent flow.

Like the processes of a steel producer, a workflow is a description of the processes that take the iron ore and process it into steel.

So if we look to Figure 1 below, we can see a diagrammatic representation of a workflow process we use at ATS.

You will see that it comprises 5 procedural steps that process the data to produce trend following portfolios. The processes include methods to develop and test for:

• system design and development
• system robustness;
• recency (adaptability);
• sub portfolio (market) long term historical performance; and
• portfolio long term historical performance.

Figure 1: Diagrammatic Representation of Testing Process Including Workflow

We will be using such a sequence of steps (a workflow) to “let the data speak for itself” with minimal discretional input, to play a dominant role in portfolio development.

Being able to repeat the procedures undertaken in our data mining exercise is an essential pre-requisite so that we can do it again if it is necessary and allow others to apply the same processes to validate the model. Furthermore, repeatability allows for the addition, editing or removal of steps at a later date to expand, amend or adjust the experimental process if required.

To assist in our strong preference for systematic and repeatable processes, we use algorithms themselves to undertake the sequential processes of data mining. Each step in the process is assigned to an algorithm that is responsible for undertaking the operation , but we might need a human brain along the way to define a scope within which each algorithm can operate. Such as the scope and range of the market universe we test across, the range of variables we use to test a strategy parameter, the degree of processing to be undertaken and any threshold values assigned to processes.

A workflow process can therefore be likened to an ‘experimental process’ used by a scientist to test a model and it must be documented so that it can be peer reviewed, replicated, amended and possibly improved.

Fortunately, when using a systematic workflow method, the documentation of the experiment is quite a straightforward matter. Data files used for processing, settings applied to each algorithm, the code of execution for each algorithm and process reports and outputs are all saved and filed away for future reference.

Dependent on your coding skills, you can either develop your own workflow algorithms, or if coding skills are absent, then you could use 3^rd party data mining software for these purposes. However, it is far preferred to use your own ‘inhouse coding method’ as the workflow process then can be specifically tailored to your hypothesis. As good as some of the 3^rd party software is, there are always deficiencies in which you therefore need to develop workarounds for.

This does not preclude your ability to undertake these processes without algorithms, and this can be achieved manually, but given the data crunching that needs to be undertaken through a rigorous workflow process, using algorithms under PC power vastly accelerates the process. To undertake data mining across extensive data sets requires intensive computer power (both in terms of processing speed and memory). It is strongly recommended that you choose the systematic workflow path.

Those Dreaded Words……Back-testing and Curve Fitting

Now you will notice so far in our discussion that the term backtest is noticeably absent. This is due to the bad rap that backtests have experienced due to their frequent erroneous application that is riddled with bias.

Backtesting is NOT a process undertaken to predict future returns. Rather a backtest is simply an empirical process used to validate our hypothesis using actual data, (or more explicitly what is known).  

It is useful for testing conclusions about a systems robustness or adaptability over historical data, but that is where it ends. We do use backtests extensively as a basis to evaluate our trend following models for their robustness and adaptability using historic data, but we never use them as a basis of future expectation.

Backtesting is just a small part of a deeper process of system design. A system needs to be designed first that is configured to target a desired market condition (such as a trending condition) using logical design principles that embed causation into the derivative relationship between the system and that market condition. A backtest is then used on ‘unseen data’ to evaluate the strength of this causal relationship and whether your system does its desired job.

If markets don’t trend, then your system doesn’t perform. If markets do trend, then your system performs. There is an explicit connection between system and trending condition developed through system design principles.

Curve fitting (also known as data snooping) arises from a trader first observing data before developing their system. The system is configured to a data set that may be simply a random perturbation. You then undertake your backtest and lo and behold it performs admirably but there is no causal relationship between your system and a ‘real market feature of enduring substance’. The system therefore falls over with ‘unseen data’.

The problem with an approach such as this is that there is no logical reasoning in this process that attaches a causal relationship between the design of the system that is being evaluated under a backtest and that observed pattern of behaviour. With an absence of any causal relationship, we have no hope of receiving any conferred benefit from future unseen data.

We explore ‘curve fitting’ in a future Chapter as it is essential that we know how to identify it, so we can avoid it like the plague.

In our next Chapter we will get into the first step of the Workflow process, namely the System Design Phase. This is where we don our creative trend following hats and devise systems that obey the Golden Rules of Trend Following before we commence assessing their merit using market data.

Chapter 14: Put Your Helmets On, It’s Time To Go Mining

In the previous Chapter, we introduced readers to a data mining method we deploy at ATS, which undertakes a Five-step Workflow Process to interrogate Market Data, using long/short trend following systems that target outliers, known to reside in the left and right tails of the distribution of market returns.

The workflow process represents the procedural steps in our Data Mining enterprise that we use to test our hypothesis.

And our 5 Step workflow is described below:

Defining the Scope of Our Experiment

We can only commence the workflow process after we have defined the scope of our experimental test and have obtained the stockpile of resources (market data) from which we will be undertaking our processes. So, in this preliminary phase, before we get the wheels of our workflow process whirring, we need to collect data, and I mean lots of it  spanning vast stretches of time across geography and asset classes.

This data set, upon which we will be conducting our experiment therefore defines the scope of our testing universe from which we then want to make conclusions. Now remember that this is not representative of the actual reality of our financial markets, as we must restrict our universal scope to what can be realistically assessed under our limited resources available to our factory, however our stockpile needs to be sufficiently representative of ‘the reality out there’ so we can make pertinent conclusions about that larger reality.

Now does this mean that we need to define the universe up-front that we will be actually trading with our diversified portfolios? Not at this stage. This just describes the pool of data we draw from. Ideally that pool is diversified across geographies, asset classes and timeframes so that our workflow can derive a portfolio using a part of this spectral diversification based on how the return streams of this vast array of possibilities can consolidate together using ‘data to do all the talking’. However our hypothesis also places demands on us that we need to stretch far and wide.

Our hypothesis states that our method of extracting alpha from fat tailed markets is applicable to ANY liquid market, which is an unusual call to make for a predictive mindset that likes to specifically target a market by virtue of its own unique price action signature.  But it makes sense if we are talking about fat tailed market environments which by their very definition are unusual non predictive events.

We need all the data we can get our hands on, within our resource constraints, as we are targeting the possible causal relationship between unpredictable rare events (fat tailed environments), and the success of our trend following models. We know fat tailed environments exist, for in Primer 4 we saw them explicitly recorded in a vast array of markets spanning across asset classes and timeframes.

But we especially need very large data sets spanning a wide universe of liquid markets because our hypothesis is couched in terms of the benefits that trend followers can receive from specifically exploiting these anomalous phenomena. This puts a different tack on our whole data mining exercise from conventional logic applied to ‘predictive data mining methods’ and entirely flavours the way we attack our problem as trend followers.

Conventional practice in quantitative science suggests that we remove anomalous data from our data sets before we interrogate them, given the material impact these outliers can have over the outcomes of the experiment. But here we are now demanding that we include them, as we believe that they are a trend followers ‘bread and butter’.

So, you could now understand why we get feisty and say…”hey, you quant guys, you are deliberately diluting the potential power of our entire trend following process through excluding their impact on your data sets. No wonder you have trouble seeing or addressing these anomalies?”.

In fact, this entire Primer series so far has been preparing your, or priming you, to think differently to conventional logic and applied practice. For good reason. Because conventional wisdom fails to consider the impact that outliers have to the trading endeavour. This Primer series has been, setting your mind up, to throw away the conventional quantitative tools and metrics that are simply inappropriate in dealing with our non-linear exotic world we predate in.

It is critical that our Workflow process is honed for our specific purpose, and that we do not include any redundant or worthless process that does not assist our validation process. So, you will not find reference to the processes of Data Sample Treatment, Optimisation, Monte Carlo Testing, Walk Forward testing in  our factory processes. You also will not find the use of validation metrics such as Sharpe or Sortino in our instrument panels.

Our workflow is specifically configured to respond to our hypothesis, and ensure that any causal linkage between fat tailed market data and our trend following systems/portfolios are not diluted or eliminated by our processing method.

Having a purpose-built factory, we then organise our processes in a systematic, progressive, and logical manner to undertake a sequence of steps that progressively evaluates how each of our trend following systems respond to our defined universe of data and then, like a manufacturing plant, create a global portfolio which can hopefully validate our hypothesis.

So, we have a defined universe of market data (our resources) waiting in stockpiles outside our factory, and we have some clever engineers constructing their elegant trend following models inside the factory (which is the first step of the workflow). We will then be rigorously testing this ensemble of designs using our ‘data stockpile’ to see how they stack up and what pops out at the end of the data driven testing process.

Each step in the workflow process progressively narrows in on a potential edge that exists within the ensemble which are created within the initial design phase.

Process 1 – Design Phase

Now our engineering division comprises strange folk with weird eyes and crazy expressions and we let them do their stuff. We occasionally throw them a bone or two to keep them going on into the night, but generally we leave them to their own devices with only the strictest ‘Golden Rules’ forged from the scriptures that they must apply to every ingenious monster they make.

….and then we leave these strange denizens of the dark on their own while they conspire and create their fiendish devices.

They then come back to us with an ensemble of coded system designs which have been configured to meet these golden rules of trend following and can be classed into two different groups defined by their entry method.

• Simple Breakout Models; and
• Other Simple Trend Following Models

We need to classify these models into these grouping as the intent of each grouping is to respond to trend following conditions in different ways.

Simple Breakout Models are used to ensure we capture any significant trending condition that may become an outlier, and do not miss any opportunity. However, the issue with breakout models is that a momentum breakout is typically an explosive  and volatile affair where you need to give lots of room in the design to avoid ‘the whipsaw effect’ in these turbulent times.

The breathing room allowed for in the design unfortunately dilutes the reward to risk relationship of the model. Yes, we catch all trends using breakout, but with dilute models.

But our creative engineers have provided us with many different breakout designs using a vast array of different forms of entry breakout technique using Donchian Channels, Consolidation indicators etc.

Other simple Trend Following Models are used to target an array of different forms of trend segment aside from breakouts and based on their architecture can capture a vast array of different forms of trending condition, some of which provide exceptional reward to risk relationships.

Once again, the diversity of trend following entry design is evident in the array of entry techniques provided. Our clever engineers have used a variety of different classic trend following indicators such as moving average crossovers, standard deviation channels etc.

Diversification of Entry Benefits

Diversification of entry condition is a specific objective we are seeking for a number of reasons:

• We are uncertain of the exact form a trend will take in a fat tailed environment and we need to respond to a myriad of possible forms and volatility profiles;
• Diversification of entry allows us to have many different systems attacking different aspects of a trending condition. This therefore allows us to progressively scale up in our committed position size as a trend matures; and
• Diversification of entry can offer correlation and cointegration benefits when we compile our ensemble of systems into a diversified portfolio.

Outliers are unpredictable in nature. We have so few examples of them given their anomalous nature, we just can’t neatly classify them into different types. They can be of any form or any volatility.

Chart 32 below demonstrates a diversified suite of 15 simple trend following designs that have been applied to USDCAD between 1992 to 1995. During this clear period of long trending bias of the underlying market data you can see how each of the varying designs have targeted various aspects of the trending data series. I apologise for the colour scheme of Chart 32 as it makes it hard to see the discreet system trade trajectories….but squint hard.

Having a single trending solution significantly narrows your prospects of extracting the juice out of trending environments.

Chart 32: Diversified Suite of 15 Simple Trend Following Systems used to Capture various segments of a Trending Condition

Trends can considerably vary in form and character over the course of the trending data series and, depending on the granularity of your method, having a single solution to capture a trending condition is a woeful inadequacy. When markets exhibit trending conditions, we need to take full advantage of all their trending behaviour and squeeze the trend lemon dry. Diversification of system entry whereby each system has a different way to capture an element of a trending condition is the way we achieve this outcome.

It just takes a small degree of variation in adverse price movement which can trigger our tight trailing stop. Therefore, despite that the market trend may simply continue on its merry way after this small retracement, we are left stranded and have to work out a new re-entry as without diversification of entry, we only have one solution to address them.

A diversified approach to system design whereby we deploy many different forms of simple trend following system that target different aspects of a trending condition,  therefore allows us to capitalise on a trends ability to significantly vary in form over the trending price series.

When it comes to the portfolio compilation phase of our Workflow process, where we start to compile the discreet return streams of each of our trend following systems into a united front to tackle fat tailed trending conditions, we need to mitigate adverse volatility of our portfolio and capitalise on any favourable volatility to take advantage of the compounding effect.

The correlation and cointegration benefits offered by system diversification that particularly relate to the diversification of entry condition is a major tool at our disposal that we use to generate superior geometric returns.

Standardised Design Features to Manage Risk in Fat Tailed Environments

So now we have a diverse ensemble of entry methods, but what about the features of risk management that are embedded in the architecture of each design?

We find that every entry design comprises a standardised logic encoded within them that can be applied to any market or timeframe and that risk is treated in the same way by all of them.

We find that each design adopts the Average True Range as a method to define an initial stop, a trailing stop, AND a standardised method to define risk in $ to every single design. So, it does not matter what market or timeframe we apply our systems to. They all allow for a standard dollar risk bet for each design solution. In their application in a diverse portfolio, they are all configured equally in terms of their risk contribution to the portfolio. That is great and just what we need.

We also find tucked away in the diverse range of designs are measures that specifically respond to the uncertainty of fat tailed market environment:

• No profit targets applied in the solutions allowing for potential infinite reward under fat-tailed conditions;
• Lookbacks used by entry indicators are generous ensuring that the solutions avoid the propensity to over-trade during noisy or mean reverting conditions and only allow for a trade entry into a trending conditions when they become material in nature;
• ATR based risk management methods allow for a vast array of different volatility settings to be applied that capture a vast array of different volatility profiles of a trending condition.

They are crafty and clever engineers after all so we will feed them.

Our Systems Need to be Causally Connected to the Market

You see, our engineers have realised that it is the market that dictates our fate, and our system needs to be aligned with it. So if we observe a market that is trending or a market that is not trending, then our system performance needs to express that condition. System performance is a derivative expression of the market condition.

A trading system is simply a method we apply to capture a prevailing market condition. Change the market condition and that system will flounder.

Our engineers recognised that they needed to integrate very simple causal logic into their design to be able to capture trending environments. They had a simple task to perform as it turns out. Design very simple systems that are directionally agnostic (either long or short), are only active during trending market conditions, have a tight stop to minimise the risk of adverse price moves, and have a trailing stop which progressively snugly follows the overall trend direction.

That is all there really is to the overall design logic of trend following systems.

Chart 33 below provides a brief description of the core design features around which our engineers devised their devilish devices.

Chart 33: Simple Trend Following Core Design Logic

Just four simple design principles to diversify around:

1. Entry Condition – Diverse simple entry conditions allow us to capture a variety of different material trend forms across any liquid market;
2. Initial stop – Normalised initial stop method to allow application across any liquid market, is used for standardised risk allocation and provides a risk release if we are wrong in our trade entry;
3. Trailing Stop – Normalised trailing stop that always cuts losses short at all times during the trade event from entry through to exit;
4. Exit Condition – A simple exit condition that provides a signal for that particular system design when it deems ‘the trend to end’.

As you can see these four simple design principles provide the constraints that dictate our success or failure in capturing trending conditions.

All our engineers did was to think of a system as a designed container, within which price needs to reside for the system to be profitable. If price decides to move outside those constraining parameters of your system we design, then we have a losing trade. While price remains within the envelope of these applied constraints, we remain in the trade.

If markets do not trend, then our systems do not perform. If our engineers have done their job well then our portfolios will stagnate as opposed to enter drawdowns during unfavorable conditions as they have reduced our propensity to trade during unfavorable non-trending conditions. But that is by-the-by. No trending condition equals poor or no performance from our correlated systems. In fact, if you observe your performance during unfavourable market regimes and find that you are achieving great performance results, this is a sure sign that your system is not capturing trends but rather simply capturing the spoils of the price data that has been presented to it.

This is a symptom that your system has been over-optimised, and curve fit to historical data as opposed to being designed to respond to trending market environments.

It is essential when developing a trading system that you understand the constraining variables of your system and how they ‘map to the market condition’. There is no system that can respond to all market regimes….so having an under-performing system is to be expected when conditions are not favourable to your system variables.

Avoidance of Curve Fitting through a Design First Logic

The design phase is a very important step when using computer power to undertake data mining.

Today there are many 3^rd party data mining platforms that make it so easy to generate trading strategies that ‘fit the data’. I call these ‘convenient quantitative disasters waiting to happen’. There is so much frustration by the traders that invest in these systems when they deploy their ‘convenient’ coded algorithmic expressions with the expectation they will deliver their eager expectations….and then they just fall off the cliff when they are taken into the live trading environment.

So much money spent on these quantitative ‘shiny’ behemoths and they still can’t deliver on the promises.

You see, all that the user has done, is ‘curve fit’ a solution to data noise that has no enduring potential when the ‘curve fit’ algorithm is then tasked with the fate of navigating an uncertain future.

Here is how the problem starts. Some data mining methods start with no apriori design assumption and randomly generate strategies (or use genetic algorithms) to identify solutions that meet specific performance criteria. For example, the practitioner simply enters the desired return to drawdown ratio, profit factor, minimum number of trades etc……and then lets the workflow process generate ‘any’ possible solution that meets these criteria. Often solutions are generated that make no intuitive design sense yet can pass the performance validation criteria.

The problem arising from this ‘generic’ method is that there are myriad ways a solution can be generated to meet these performance criteria by simply ‘curve fitting’ to past data without their necessarily being a ‘causal’ relationship’ between that design generated and the past data.   They simply may be curve fitting to noise where no ‘signal is present’. Primer 3 discusses this dreaded term ‘noise’ in more detail.

So how do we avoid this ‘curve fitting’ debacle? We use a workflow method that is far narrower in definition specifically targeting a particular market condition (tail events) and whose design configuration specifically responds to that condition and ‘causally’ links the solution to that condition.

The Further Tragedy of Curve Fitting and Why We Need to Keep Systems Simple

Now that we understand that it is the constraints imposed by the system that ultimately decide your fate when trading these markets, you could imagine that the more variables you include in your trading system design, the more prescriptive you become in choosing the specific market conditions  you address with your trading system.

Imagine we have the following trending series of daily closing price points of market data.

Chart 34: Hypothetical Plot of Trending Market Data

Now we want to design a trend following system that captures those trending points of data.

We could design a trend following trading system that exactly responded to that past market data with no data error to capture all the possible profit that existed in that price series. The curve would look like this.

Chart 35: A Curve Fit Response to Trending Data

Chart 35 above is clearly a curve fit response to past market data. If the future exactly plays out in accordance with the plot of past market data, we will achieve a perfect trading result with maximum profit and no drawdown.

What is important to note is that the system plot above does not represent a simple trading system with few variables. If we plotted this curve algebraically it would represent a complex polynomial function with many variables such as.

y=ix⁵+ jx⁴+ kx³+ lx²+ mx+ c

This system has therefore 7 variables that need to be fit to the past data (I,j,k,l,m,c,x).

So, as we increase the number of variables used in our system design, we more prescriptively ‘map’ our system trading performance to more precisely mimic past market data. While this produces a higher correlation between the past market data and our trading performance, it also exposes us to a greater standard error if future market data does not exactly mimic the price action of this historic data set.

If we refer to Chart 36 below, we have plotted a possible future trending series against the historic trending series we used to design our ‘curve fit system’.

The curve fit nature of our trending system designed to closely match past market data, now creates significant error differences when you compare the deviation of results between the two different trending series.

Chart 36 shows the standard error between a single data point of the possible future trending series, against the red curve fit result derived from our historical trending series. The error between data points in these different trending series are significant. This error magnifies when applied across the time series if there is not a perfect match between the locations of the two discreet time series.

Chart 36: Standard Error between a single data points of two trending series

So let us see what happens when we simplify our system to more broadly, and we less prescriptively respond to the past market data seen in Chart 37.

The simplest algebraic function we can apply, that loosely represents the plot of historic market data, is a straight line. In this case a regression line that represents the line of best fit that represents the data. Clearly this does not exactly match the plot of past market data, but it does provide a simple representation of many potential trending series. This will be very useful if future conditions do not exactly represent the past.

Chart 37: Simple Representation of a Historical Trending Series

The algebraic expression that represents this curve is familiar to many of us who have studied algebra at school and is represented by the line equation of

y= mx+ c

Notice how we now only have 3 variables that describe this trending model. With simplicity comes fewer variables and greater robustness.

While we lose specificity in our design with simpler models, we significantly increase the variation of possible future outcomes which a simpler model can perform within.

So, if I now plot the future data on this simplistic model (refer to Chart 38) we find that the straight line is still a fairly good representation of the future market data and the standard error that arises from this new trending data series is significant lower than the curve fit trajectory.

Chart 38: Application of the Simple Linear Model to a Future Trending Series

So, the previous examples clearly demonstrate why it is imperative that we deploy simple models with few data variables to have the greatest chance of navigating future trends that may vary from historic trending data.

Trends come in all shapes and sizes from parabolic trends to linear trends to trends with significant volatility embedded in them or with smooth linear trajectories.

Given that the future is uncertain, future trending environments are likely to adopt a variety of possible forms. We therefore want to avoid overly prescriptive models that conform to a particular class of trend and adopt very simply trending methods that can capture a broad class of different trending condition. Having a simple core design significantly improves the robustness of our trend following models in navigating an uncertain future.

Well, that’s it for this Shift. Time to go home, clean up and get ready for your next Shift where we will reviewing the Robustness Stage of the Workflow Process.

This is where we take these engineered systems for a workout and blast them with market data to see where any risk weakness resides in these trend following contraptions.

Chapter 15: The Robustness Phase – T’is But a Scratch

In our previous Chapter we revealed how our ‘design-first logic’ has avoided any propensity to produce curve fit results, when targeting the outliers using our trend following models.

We discussed why it is essential to adopt a “design-first logic” before we apply our powerful data mining processes towards the trend following quest, to avoid ‘fitting our solutions to noise’. Our logic enforced a causal relationship to unite our trading systems performance to the market behaviour (or the signals) that we want to participate in.

We also took you on a small journey where we demonstrated how it was essential to develop simple trend following systems with few variables, to avoid our propensity to be too prescriptive in our trend following designs.

…….and then we gave our engineers a number of ‘Golden Rules’, that must be observed in every simple design solution and they went ahead and developed them……thousands of them. All variations around these Golden Rules or core principles of trend following.

And here we are now. Our clever engineers have given us literally thousands of trend following systems of both ‘both breakout configuration’ and ‘other trending condition configurations’, that span a diverse array of trending form found in outliers, and may be applicable to include in our trend following portfolios generated at the end of the workflow process.

So now we can be confident that these solutions can respond to ‘Fat Tailed Environments’, and can therefore proceed to our next processing phase in our workflow to test how well these solutions have stacked up against the ‘outliers of the past’.

Once we complete this next phase of the process, then we have a higher probability that the ‘surviving candidates’ will have a good chance of navigating the ‘outliers of the future’.

Revisiting Robustness – Invert your Thinking.

In Chapter 5 we came to the conclusion that robustness would be our chief determinant in our selection process for suitable return streams, and we adopted the MAR ratio as the preferred risk adjusted metric that we would use for our evaluation purposes. We need to keep this in the back of our minds when discussing robustness.

So now we have our vast array of system designs sitting in front of us delivered by our Engineering Division, and we want to now rigorously test them by putting them through the hoops with our large and diverse data stockpile, to evaluate any risk weakness that resides in them.

We already know that our system variations will respond well to trending conditions in these fat tailed environments, but we are unsure how these systems will perform over enduring periods of unfavourable market condition.

Notice what I am now doing now. I am flipping the mindset from thinking about the windfalls of outliers to thinking about risk associated with attempting to capture them.

Previously we came to the conclusion that our systems could catch the outlier, but now I am focussing on the risk of this eventuality being thwarted by noise and mean reversion. In other words I am now assessing the impact of our models when ‘normal market conditions’ interfere with our quest for outliers.

This is where risk lies for the Trend Follower. Our systems protect us from adverse left tail trade exposure, but we can be exposed to the costs of many whipsaws if we stray too far into the Gaussian envelope.  This is what creates slow building drawdowns which we want to avoid, if possible.

Remember that we seek to trade the tails of the distribution of returns. What I am doing now by focussing on robustness is arbitrarily defining the dividing line between the bounds of a Gaussian distribution and the Fat Tailed conditions that extend beyond that distribution into the left and right tails.

We want our systems to not only catch the tails, which we know they will do through design logic (NOT optimisation), but we more importantly now want them to be survivors so that they can always be around to participate in fat tailed conditions IF they arise. The better survivors still catch the outlier but have lower drawdowns.

So, this is what we will be assessing in this second stage of the Workflow Process, termed the Robustness Phase. The ability of our systems to survive the hostile adverse market environment of the past. In our next Primer we will be assessing the ability of our systems to survive a possible hostile environment of the future.

But before we carry on, I would like to give you another example of how inverted logic can be applied successfully to crack puzzles wide apart.

The Curious but Clever Case of Abraham Wald

Source: https://en.wikipedia.org/wiki/Abraham_Wald

How to Assess Robustness in our Workflow Process

Now down to the nitty-gritty of this processing phase. After all, we have to do some ‘brain draining’ work sometime.

This Robustness phase will be used to assess:

• The risk adjusted returns of each systems return stream on a pre-cost basis so that we can then use visual methods or the MAR ratio to evaluate weakness across the entire return path (equity curve);
• The risk contribution that each return stream makes to a possible future portfolio. Each return stream will be non-compounded and normalised to allow for direct comparison between return streams;
• The Multi-market capability of the system which is a method we deploy to increase the low trade sample size that we achieve through our systematic methods applied to a single return stream. Multi-market capability lifts the trade sample size from a small trade sample to a large trade sample and is a principle measure we use to assess robustness; and
• The suitability of the return stream for future compounding treatment. An evaluation is made of each return streams MAR ratio on a non-compounded pre-cost basis to evaluate how well they are suited for later compounding treatment.

Pre-cost Assessment

Before we complicate our exercise by not only looking at the trend following capability of each return stream, but also including the ‘broker specific’ cost inclusions that can have a major impact on our already weak edge that is present in our systems, we conduct this phase of the process on a pre-cost basis.

This means we exclude broker specific impacts such as spread, slippage, holding costs (or SWAP) during this phase of the process.

All we need for this phase of the process is the Open, High, Low, and Close data for any liquid market.

We will be backing in the broker specific costs in the later ‘recency phase’ of the workflow process to assess how the system can fare under realistic ‘live conditions’, but until then we don’t won’t to complicate this process.

Undertaking this assessment on a pre-cost basis therefore allows us to isolate performance in relation to the ability of our systems to capture trending environments and be influenced by noise and the mean reverting character that may lie in the market data.

Having achieved this, we can then use visual graphical methods to eyeball the equity curves produced to identify any ‘risk weakness in the signature. We can also use the MAR ratio as a method for this assessment. Given that we are not compounding the equity curve at this point we use the Return on Investment of the return series (ROI) as opposed to the Compound Annual Growth Rate (CAGR) and use a proxy for the MAR ratio as defined by ROI/Max Draw.

Determination of Risk Contribution

We normalise the return streams of each system using ATR based scaling techniques to allow for ‘apples for apples’ comparison amongst alternatives spanning different markets, timeframes and systems.

Each return stream has an equal $ risk bet applied per trade which ensures that volatility is normalised in equivalent dollars.

So for our thousands of ‘pre-cost’ return streams, we can now directly compare each and assess their overall contribution of risk to a possible future portfolio either by way of referring to their Return/Drawdown contribution or by visual assessment of their entire return stream which can be used under ‘visual comparative methods’.

Multi Market Capability

Now given that each return stream is only targeting unpredictable and rare outliers which can mean that we are holding trades for many years, the trade sample size per return stream is very low. We are trying to determine the impact of noise and mean reversion and their impact on overall robustness, and for this assessment, we need a high trade sample size.

We have a clever trick up our sleeve to achieve this given that we have now normalised each return stream from our prior processes.

We test our systems over our entire data set that comprises many markets as now each market is just different data whose independent qualities have just been eliminated by normalisation methods.

This lifts the trade sample size from say 15 trades over a 30 year plus market data sample to say 40x this level (eg. 600 trades per system)  when we apply our entire data set. This significantly increases the rigour of this testing phase and ensures that the noise and mean reversion affects that contribute to whipsaws have statistical significance.

You see, each market is just data to us now offering a vast array of different possible paths upon which we can now test the efficacy of our systems under different market conditions.

We apply a filtering mechanism at this point that states that we want at least XX markets of the total YY multi-markets (say 30 markets of a total 40 in our universe) to pass this test by delivering positive expectancy.

There will be some markets over this long data history that do not possess sufficient outliers to generate a sufficient edge to achieve positive expectancy (pay for all the whipsaws), so we cannot expect our systems to achieve an edge over all markets tested on. But we do expect a high pass rate, and the better the multi-market capability of each return stream, then the more we like them.

Suitability of Each return Stream by Market for Future Compounding Treatment

Finally, after all the prior steps in this process, we now need to make a decision to further reduce the scope of the hundreds of return streams that still have survived to this point. After all, we do have finite capital restrictions that does limit our ability (through minimal allowable lot sizes applied by our Brokers), to trade all these options, even though we might not like to as we hate ‘selection bias’.

Reality unfortunately starts to take a bite, and some decisions need to be made, to select those candidates that we will be taking to the next phase of the workflow process.

For the survivors so far, that have now passed the multi-market test, we therefore now need to refer to their performance at the individual market level.

We use the proxy for our MAR ratio on a pre-cost basis as our metric for each normalised return stream for each market, to identify those return streams that offer the best reward to risk relationship on a per market basis.

We eliminate any candidate that has a negative MAR (negative expectancy), and then we rank the surviving candidates by positive MAR. Of course, this does introduce weak optimisation into our method, however we counter this by including the top 100 or so in our collection per market.

We need lots of survivors for our next step in the process and their survival is related to the MAR ratio that each normalised return stream generates when compounded.

So at the end of this entire process, we may have 40 discreet markets (defined by the extent of our universe) each with 100 strategies that have passed all phases of the historic robustness test. They all possess a weak edge, they are ideally configured for compounding, and they all have demonstrated their ability to survive the turmoil that the market has thrown at them across a vast array of different market conditions.

Chart 38 below provides an example of 40 Non-Compounded Return streams using Breakout Techniques and Various Trend Following Methods that have passed the Robustness Phase. Notice the correlated step-ups during material trending events displayed across the set, and the absence of risk weakness in the drawdown profiles during intervening periods of the data history. The process is starting to sculpt optimal equity curves for compounding treatment.

Chart 38: Example of 40 Return Streams for EURUSD which have passed the Robustness Phase

We now have a set of robust candidates ready to future treatment through the workflow.

Now we are ready to move to the next phase of our workflow process, which is the Recency Phase.

This phase, like the process of natural selection, assesses how these robust candidates now fare over more recent market conditions. We inject live trading costs back into the results to simulate a live environment, and we then preferentially select those that have fitness embedded in them.

This is the adaptive phase of the process which is now required to ensure our systems can respond to dynamic market conditions.

Chapter 16: The Recency Phase – There is no Permanence, Only Change

In Chapter 15 we undertook a process in our Workflow that sought to identify robust strategies from a large list of possible trend following systems that had been logically designed by our Engineers. We found lots of them, say 4,000 solutions that met our validation criteria (about 40 markets x 100 solutions).

Chart 39 below provides an example of 40 return streams from a 50 year historical data set that have passed the prior Robustness phase comprising both breakout models and other various forms of trend following model. You will notice that there are discreet points in the time series, where positively correlated step ups occur. These are moments when our models experience ‘outlier’ moves of significance. Between these step-ups there are non-correlated periods of stagnation or slight drawdown. This is where noise and mean reversion play a role in contributing to drawdowns.

Chart 39: Example of 40 return Streams for EURUSD brought forward from the Robustness Phase for further Testing

In addition to robustness expressing a return streams ability to survive adverse market conditions of the past, the term ‘robustness’ doesn’t end here. There is more to it.

Robustness is also a statement of ‘responsiveness to change’. It relates to how well a strategy is configured to (or correlated with) its current environment. The prior test undertaken, demonstrates survival across a very long data history, where many unfavourable market conditions have contributed to adverse drawdowns and this is great to know, but not sufficient. If any of these unfavourable conditions arise again, then we can be confident that our systems can navigate them, but such a long snapshot of a past environment does not provide us with an indication of how fit these strategies are now in responding to current adversity.

Robustness is more than just a Historical Test. It is also a Statement of Fitness

So if we can imagine an environment that dynamically changes over time, we could also imagine that our robust strategies need to be periodically refreshed with newer risk mitigation methods that respond to this environmental change. In other words, that our systems can adapt to that changing state through ‘selection processes’.

You see, when we look at other complex systems such as natural systems, our perception of what it means to be a robust species is the ability of that species to respond to the changing environment around them. The environment changes and a robust species needs to cope with that change about it. Robustness is therefore tied to the ‘close’ relationship (or the correlation) between the environment and a species. If a species significantly lags in its coping state when compared to the altered state of the environment, it is exposed to far greater risk in the current environment. A species needs to be strong and only carry limited survival risk in the current environment, so it is capable of absorbing new risk relating to future uncertainty.

We need a similar coping strategy for our trend following systems in how they can be periodically refreshed to minimise current risk in their architecture. Our trend following systems need to be strongly positively correlated with current trending conditions. The current configuration must not be warehousing risk associated with the current state of the market so it can fully absorb future risk.

Trends in any complex system relate to transitions (or change events) in that complex system, and the nature of the trending condition is also subject to change over time. We therefore find that the trends of yesteryear are significantly different to current trends, yet the trends of yesterday are more similar to the trends of tomorrow.

So our robustness statement for our systems also needs to reflect this ‘correlated’ relationship. That the trend following systems of yesteryear are significantly different to current trending systems, yet the trending systems of yesterday are more similar to the trend following systems of tomorrow.

This principle of correlation between agent and system state is found in all complex adaptive systems and relates to how ‘selection processes’ work in a complex system. It relates to how ‘fit’ an agent is in that system currently, to meet the challenges of tomorrow?

Our Progress So Far

So, in our workflow process so far, we have say 40 markets x 100 solutions sitting in front of us that have passed the prior step. Their robustness so far has been assessed using a historical data set comprising an account of what is known, but now we want a process that is adaptive in nature, whereby more recent data which has an adverse affect on trend form and characterised as whipsaws (expressed by current volatility, noise and mean reversion), plays a stronger role in the selection process.

This phase of robustness testing referred to as ‘The Recency Phase’ provides an adaptive component to our Workflow  whereby our process of system selection ‘adapts’ over time in response to new trending information which is received by ‘future data’ and is injected into the workflow process.

Our prior ‘Robustness Phase’ was undertaken using historic data, so the conclusions reached about each systems robustness was a statement made about that systems overall response to the entire data set.….however we know that this statement of robustness is going to be challenged in the future as new data that has not been seen before, dilutes this ‘historic robustness statement’.

So, we therefore need to undertake this entire workflow process at annual intervals.  Each time we undertake our workflow process, a years’ worth of extra data for our universe is injected into the process which includes any new information that can assist in keeping our systems fit by releasing adverse risk. By replacing our tired systems with new vitalized versions we can keep the portfolio sharp.

We want a Conditional Outcome of both Historic Robustness and Current Fitness

So, to allow for an adaptive element in our workflow, we not only want to select for systems that are historically robust (as they have passed the prior robustness phase), but from this selection of say 40 markets offering 100 solutions, we then want to select those historically robust systems that also demonstrate current fitness. This additional requirement reduces the selection to then say 40 markets offering 50 solutions (2000 or so total solutions).

Both the prior Robustness Phase AND this Recency Phase of our workflow are contingent processes.

In our processes here at ATS, we adopt a ‘recency window’ using a 5-year look-back, but you can set this window to any desired look-back. However, we do not want it to be too ‘adaptive’ and we simply want to ensure that a degree of change can be imparted in our process.

So as trending conditions change over time from say, shorter term trends to longer term trends, or less volatile trends to more volatile trends, we will clearly lag in our response time but not by far. We don’t want to be adapting to noise, but rather the slower changes in trend form.

We view this process of ‘sharpening our portfolios’ as a method that ensures we do not attempt to ‘time’ the market. By keeping the portfolio razor sharp through an annual ‘re-balancing process’ we avoid having to make decisions such as ‘is the portfolio in too much drawdown?’  and ‘when should I turn the systems off with under-performance?’. We integrate a process into our Workflow that ensures that these decisions are made for us, without having to interfere under discretion. It is the discretionary over-ride that frequently is the reason for long term poor performance.

Think of it this way. The time a predictive mindset wants to ‘change’ is usually at the worst time, when drawdowns have already been introduced that compromise geometric returns. Avoid that impact by never letting your portfolios get severely blunt in the first place.

Bacteria and viruses and perhaps the most adept creatures at responding to change. They don’t predict change (attempt to ‘time’ change).  They already have mutated strains sitting in their collections, that are ready for it….at all times. If they waited to change, the slow response time would probably be too late to perpetuate their existence.

This Recency Phase is likened to processes of Natural Selection whereby we reduce our selection of robust systems to those that are more fit for immediate purpose. They are more likely to carry less warehoused risk and possess system characteristics that are suited for the near future.

Each year we undertake the workflow process, we therefore replace all our portfolio systems with a new generation of fitter candidates.  But we allow systems with existing active trades to play out before they are deleted from our portfolios.

In our historical robustness test we historically tested across a 30 plus year look-back and in our recency test we use a 5-year look-back using a component of this historical data set. So we are reusing a portion of this data again. Now this poses a problem of ‘reuse of data’ which we need to understand as it can introduce a bias in our selection process.

A Word on the Dangers of the Reuse of Data and Why we Insist on Using Large Data Sets for Our Decision Making Processes

We need to take a slight detour now, as we need to explain this possible dangerous bias that arises from the reuse of data for testing and more specifically any process of decision making between alternatives when randomness has a say in the matter. The danger arises from a subtle bias associated with selecting the best candidates from amongst alternatives where randomness has a say in the matter.

Now unlike alternative data mining methods, we do not break our data set into In Sample or Out of Sample data to design and then validate our systems. We need all the data we can get our hands on to test the robustness of our systems given the low trade sample size per system. We have avoided the use of In Sample data by designing our systems from scratch without the need for data to test their efficacy.

This is great as it avoids our propensity to over-fit our solutions through any form of optimisation process  but using an entire data set to test for robustness for example can lead us to problems where we then re-use the data more than once. We actually do need to reuse 5 years worth of our historical data for our Recency Phase.

The issue of bias that unfolds from the reuse of data relates to a ‘selection bias that creeps into our decision making process’ if there is a component of randomness in the results. A random distribution of returns can offer a range of possible paths from positive equity curves to negative equity curves and anything in between. If there is randomness in our 5 year look-back that influences the recency phase (which there is likely to be), then by selecting the equity curves with a positive result and eliminating other random curves that are unfavorable, we are actually biasing the result with our selection process.

You see, our decision has now introduced a Bayesian Bias into the selection process….just like the Monty Hall problem. Warning…if you want to get your head messed up then go ahead and click on that Monty Hall Problem.

A Visual Way of Understanding the Problem of Selection Bias

Let’s understand this better visually. We all now understand that an equity curve which is a derivative expression of underlying market data, comprises elements of non-randomness (signal) and elements of randomness (noise). In a ‘mostly efficient’ market , the randomness in the equity curve is considerable and should never by overlooked. For Trend Followers, this is particularly pertinent as a great deal of our equity curves are a feature of the noise that resides in our long term equity curves. There are only a few moments where we have ‘outlier anomalies’ residing somewhere in an otherwise noisy equity curve. In Primer 3 we showed an example of an equity curve of a trend following program that was derived from a trade sample size of 500 trades (refer to Chart 40 below). The Trend Following Program (in red) was found to have a weak edge, but only when examining that equity curve over a far greater trade sample size (refer to Chart 41 Below where we included this weak edge example of a Trend Following system in a sea of otherwise random equity curves).

Chart 40: Trend Following Program embedded in a sea of Random Equity Curves (Trade sample size 500)

Chart 41: Same Trend Following Program embedded in a sea of Random Equity Curves (Trade sample size 5000)

So let us assume that we only had a small trade sample size  of 500 trades  (like we found in Chart 40) and had to make a decision about which equity curve we would choose for our Trend Following Portfolio from amongst these alternatives.  Let us say that we could only pick 5 from this sea of alternatives. We decided to use profit over the time series as out prime method of choice determinant. In Chart 42 we therefore select the top 5 only from this list of available alternatives.

Chart 42: Selection of Top 5 Equity Curves from a Trade Sample of 500

If you refer to Chart 42 having selected the top 5 return streams, you will note that they do not include the return stream of the Trend Following System with an edge that was displayed in red in Charts 40 and 41. You have only selected random return streams.

This is why you find that when you take these selected candidates to the live trading environment in your Diversified Portfolio, they immediately fall of the cliff when they are implemented. They only ever were random equity curves and such curves have no enduring projection power into an uncertain future.

However if we had made this decision with a Trade Sample size of 5000 trades (Chart 43) we would have avoided that problem. We would have then selected the system with the weak edge (red) albeit, we would also have selected a few random candidates as well.

Chart 43: Selection of Top 5 Equity Curves from a Trade Sample of 5000

So hopefully you now agree that using large trade samples sizes is always preferable when having to make decisions between alternatives.  In fact, as you can now imagine, we can actually evaluate the power of our edge by comparing our Trend Following equity curves against random equity curves over large trade samples.

Back to our Recency Test

Okay so now that we have made this slight detour, we return to our description of the Recency Phase. Have we adopted a slight Bayesian bias in our process by reusing 5 years of data for our recency test?

Well yes, we have slightly, but this Bayesian selection bias is reduced from our use of a fairly long look-back (5 years), where randomness has less say and we feel it is worthwhile to take this slight statistical risk. Furthermore, when considering our selection process including the Recency Phase you will note that we are using the entire long-term series of equity curves in our decision making process. However, this form of bias is inevitable in any form of selection process, and ultimately given our finite capital limitations, we must make selection decisions from alternatives.

Fortunately the statistical bias we have recognised in our process is significantly lower than the more pronounced selection bias arising within alternate data mining processes such as the dreaded ‘curve fit word’.

It is exceptionally difficult to avoid statistical biases in data mining, but we do our best.

Being aware of the problem is important as it allows us to identify it, consider ways to reduce it and note any residual possible bias in our assumptions. Many traders who adopt powerful 3^rd party data mining software are blissfully unaware of this issue.

Inclusion of Brokers Trading Costs

In our prior historical Robustness phase, we undertook our testing on a pre-cost basis, but now with our 5-year Recency phase, we introduce brokers trading costs back into our assessment such as commissions, interest holding costs (eg. SWAP), and slippage assumptions. This ensures that realism starts to enter our process.

We conservatively apply these costs to deliberately understate results so that we are unlikely to get any surprises from our brokers claws when we enter the live trading environment.

By including trading costs into our assessment, we now need to lift the pre-cost positive expectancy threshold for strategies to pass this phase. Given that our solutions only offer a weak edge, this post cost inclusion is essential, as it ensures that any weak edge that now resides in our systems is sufficient to accommodate these additional conservative costs.

Validation Method Used by this Process

So having tested all our robust systems during this recency phase we are now at the pointy end of this exercise where we need to select suitable candidates to move to the next phase of the workflow process.

We once again base our selection of suitable candidates of the Recency test using MAR as our preferred metric. The recency test is non-compounded and normalised like our robustness phase, but this time we also include brokers trading costs into the 5-year return streams that are generated.

Chart 44: Example of 20 non Compounded Return Streams for EURUSD that have passed the Robustness Test AND Recency Test

Those that pass this test with positive MAR then move onto the next phase of the process.

If you closely look at Chart 44, you will notice that not only do these return streams meet the long term robustness test, but components of each return stream in the last 5 years offer very good recent performance with minimal drawdown. Yes, they have been selected for their recency attributes and hence ‘curve fit’ for recency, but we are not expecting them to continue on with this recency trajectory in the future. We don’t know the future, but at least each of these return streams are now fully ‘fit’ for function to navigate an uncertain future.

So let’s say for the purposes of example, that we are now left with 40 markets each with 50 systems (2000 total return streams) that are robust and fit for purpose (aka responsive to change). We are now ready to progress to the compilation phase of the Workflow where we use correlation methods to blend return streams  together, first by individual market as Sub-Portfolios and then across markets and a Portfolio.

Well, my head hurts now. This was a tough Chapter and I had many attempts at putting it to the pen for the sake of coherence, but you will be pleased to know that it gets back to good old simple trend following from here on in, and the fun stuff of sub-portfolio compilation commences. This is where we see the fruits of all this philosophy and applied practice unfold.

Chapter 17: Compiling a Sub Portfolio: A First Glimpse of our Creation

In Chapter 16, we left our readers at the point of our Workflow Process where all aspects of robustness had been completed, with validation that each return stream meets ‘historical robustness‘ and  ‘responsiveness to change‘ objectives.

We were left with say 40 markets offering 50 systems each or 2000 total return streams from which we could then weave our portfolio magic in this Primer.

Each of these possible candidates were uncompounded, yet we understood from their risk to reward architecture, that they would fly to the heavens under path dependence when we compounded their story.

Now before we can reveal our artwork (our diversified systematic trend following portfolio), there are two further processes we need to undertake to finalise the Workflow.

1. We need to compile our return streams into Sub-portfolios (or discreet markets); and
2. We then need to compile our Sub-portfolios into a single consolidated Global Portfolio.

We will be looking at Sub Portfolio creation in this Primer, and in our next Primer, which describes the final phase of the Workflow, we will be compiling our Masterpiece (the Global Portfolio).

Sub Portfolio Creation

In this step we need to compile our return streams for each market into a Sub Portfolio, using principles of correlation and co-integration discussed in Primer 11. This therefore achieves the best risk to reward configuration for the entire sub portfolio for at least a year.

Remember that we undertake this entire workflow process annually to ensure our portfolios can stay razor sharp.

What we mean by the best risk to reward configuration in terms of Trend Following, is the best risk-adjusted performance that can be achieved through the compilation we ultimately select. This concept is discussed further in Primer 9.

Having the best risk-adjusted compilation allows us to therefore reach to infinity and beyond, when we apply compounding treatment to the portfolio, which we will be demonstrating in our Walk-through video shortly.

Now given that we have, say 50 return streams, for each portfolio, we need to decide how many of these we want to compile into a single Sub Portfolio. This ultimately is a decision that must be faced by a Trend Follower with finite capital limitations when faced with Broker restrictions.

As a general principle (discussed in Part 13) there is no theoretical maximum level that should cap our aspirations as “Diversification is Never Enough”. However the reality of finite capital under Broker minimum lot size restrictions, define the maximum threshold of system diversification that we can use.

Now as discussed heavily throughout this series, we avoid using single statistical measures such as the correlation coefficient to make a statement about the suitability of an entire return stream for inclusion into a Sub-Portfolio. We require an assessment of the strengths and weakness across the entire path of each return stream and then ‘like puzzle pieces’ how these entire paths can unite cohesively into a Sub Portfolio to produce ‘the biggest bang for the finite buck’.

If we can complete the entire risk adjusted puzzle for each sub portfolio (or market), then we have a grand solution that we can apply to a global portfolio when we start compiling sub portfolios together.

We therefore use an iteration process of compilation that, swaps in and out, every possible combination of return streams that meet our requisite number of total solutions we can use. Then using a ranking criteria of MAR, we sort these possible permutations into a listing from best to worst.

Now this iteration process really chews up PC resources. For example, let’s assume we can only have a maximum of 10 trend following solutions per market.  But we have an available 50 solutions to consider for each market. This means that with 50 available solutions where we can only use the best 10 risk adjusted solutions, requires a total of 10,272,278,170 iterations to be performed over 50 years of data. And this is for each market (or sub-portfolio).

Did I say it was an exercise in computer power?  I wasn’t kidding. Forgive me if you want to use manual methods to undertake this rigorously. It can pay to be systematic.

So when you undertake these processes, this is when you let the PC do all the hard work, and then you curl up next to a warm fireplace with your pipe  and open the Johnny Walker Black and get guzzling.

Because at the end of this process a small miracle is being generated. Your first sub-portfolio. Your first glimpse of what you have been working for…over so many Primers.

Now, giddy with delight, you come back to the PC and scan the ranked selection. You take the sub portfolio collection of 10 or so strategies that offer the best MAR, and then all you need to do is to apply a trade risk % of equity to this Sub Portfolio, integrate the code of each system into a single Sub Portfolio algorithm and…… “Bob’s your Uncle”.

You are now ready to have another chug on your Scotch, kick back, turn your Walk Forward Tester on, and see how this man-made miracle can surf the trend.

So here we are. We have a Walk-through Prepared for 10 trend following systems combined into a Sub Portfolio for EURUSD between 1980 to the 16^th May 2021 using an Initial Balance of $50,000 and a trade risk % of 1% per trade. We could of course have 20 trend following systems using 0.5% risk per trade to get an even better non-correlated result….but this gives you a picture of the power of the Workflow Process.

For added inspiration while you watch this walk-through play out as you sip your Scotch, we have accompanied it with our favourite music, to sing along with. We call these “chill out” scores the sweet songs of Positive Skew.

Take note of the walk-through as you revel in the joy it brings from the fruits of your hard work  in reading this Primer, as we will be returning after this demonstration to debrief and discuss how this Walk-Through can demonstrate that there is “Method to All this Madness” we have put you through.

Welcome back and sorry to interrupt the chillout.

We now have some debriefing to do to demonstrate how our philosophy, and the processing steps that we have taken in our workflow has now been embedded in this outcome.

First up. Did you notice how the separate systems were attacking different segments of some big long or short trends? Have a look at this segment for example of a bullish trend (Debrief 1).

Debrief 1: Diversified suite of Trend Following Systems attacking Various Aspects of Trend Form

Did you also notice how many systems are active during major trending conditions, but few systems are active during noisy or mean reverting market conditions, before or after major trends end? We squeeze all the juice out of trends and try to avoid noisy or mean reverting environments. Notice how we ‘smash’ the outliers with all our solutions blazing. Also note how we perform well when outliers are around, and under-perform when they aren’t. This supports our ‘Goldilocks’ claim that these little beauties are “not over-fit”, “nor under-fit”….”but just right fit”.

Debrief 2: Squeezing the Juice out of Trends and Remaining Relatively Inactive during Non-Trending Periods

Did you notice how there are correlation offsets and co-integration benefits being applied through this diversified suite of trend following systems? You will see when trends end, that there is a degree of hedging going on, to not leave too much profit on the table.

Debrief 3: Correlation and Co-integration Benefits from a Diversified Suite of Trend Following Systems

Did you notice the very long holds of some systems when we catch an outlier, sometimes lasting many years but also far quicker and nimble trend following models that take shorter term opportunities when they can?

Debrief 4: Diversification across Time-frame Benefits

Did you notice the fairly low winning percentage of 44%, but the very high average win when compared to the average loss? Smells like some positive skew is cooking.

Debrief 5: Some Helpful Performance Measures

Did you see how Unrealised Equity (green) always sits above the Closed Trade Balance (blue) at all times, reflecting that the Sub Portfolio never warehoused risk? Also do you see the positive skew in this equity curves signature characterised by fast step-ups and slow building drawdowns?

Debrief 6: No Signs of Warehousing Risk and Positive Skew all the Way

Did you notice the overall robust nature of the equity curve, and the more recent performance associated with the fittest series of the Collection? Can you see historic robustness and Response to Current Market conditions in this trajectory?

Do you also see how compounding has taken the bull by the horns delivering an exponential growth profile of the curve over time?

Debrief 7: Evidence of Robustness in the Equity Curve that favours Geometric Returns

…..and finally, do you see the pain arbitrage in the monthly performance results where you need to stay insanely patient over long periods of time in continuous drawdown…..but then occasionally and unpredictably you just have a great and I mean really great month or maybe even two?

Debrief 8: The Pain Embedded in the Monthly Returns

I think through this debrief session, we may have just convinced you, that our philosophy is not just smoke and mirrors. We can translate our philosophy into an outcome.

Well that is it for the Sub-Portfolio Phase. Remember that we now undertake this process across all our 40 markets or so. Gulp……….Watch the electricity costs soar with your PC setup.

At the end of the Sub Portfolio Phase we therefore have some risk adjusted ‘juggernauts’ that we then take to the final phase of the Workflow Process, to smooth the drawdowns even further, and lift the global portfolio equity curve to the heavens and above. You want more risk adjusted performance….well it is coming.

Chapter 18: The Court Verdict: A Lesson In Hubris

Finally, we are at the end of our Workflow Process. This is it. This is where we compile our global portfolio using our available sub-portfolios and do one final test. The ultimate test of long-term historic portfolio performance. This is when we can then validate our hypothesis.

In the previous Primer we started to get excited and provided an inspiring Walkthrough of ‘the first glimpse of our creation – The Sub Portfolio’, but wait, there is more wonder to be had. “Wait till you see the Second Glimpse of our Masterpiece – The Portfolio?”.

We are nervous of course as we approach our day in Court, where we can prove our hypothesis to one and all, and fortunately in the short time we have left before this auspicious moment, we can speed through this last phase of the Workflow Process to compile our Portfolio as it is punchy and brief.

We have our collection of Sub-Portfolios to work with (comprising say 40 separate markets), and without the need for repetition, we have simply followed the exact same process as that adopted in our prior process where we created Sub Portfolios. We have simply replaced the term ‘return stream’ with the term ‘sub-portfolio’ in this final phase.

To briefly summarize;

1. Commence with say 40 uncompounded sub-portfolios of equal risk$ weighting;
2. Use as many of these Sub-portfolios as you can to compile your Global Portfolio within your finite capital restrictions. Ideally, we would use all we have available to us as market diversification particularly across asset classes is a big deal in delivering diversification benefit. Don’t worry about equal representation across asset classes as we let the data speak for itself through this iteration process. With say 50 years of data across say 30 to 40 markets, there is more than enough information embedded in each sub portfolios correlated relationship with the balance of the portfolio without the need for discretionary judgement;
3. Iterate to build the optimal risk-weighted Portfolio adopting the same process as we did for the constituent Sub Portfolios;
4. Return to your PC’s and select the best non compounded portfolio composite using the MAR metric;
5. Compound the result using your desired trade risk % which applies compounding treatment and leverage to your chosen solution. Adjust up or down this trade risk % to achieve a drawdown that is half your tolerance……..and then
6. Run the final portfolio test

….and there we have it. Simple, logical and straight forward. Wait till they get a load of this?

We are now ready to present our findings to the court. Now all we have to do is find an even better musical accompaniment to our final Masterpiece, to truly do justice to this achievement of Modern Man.

The Proceeding

The courtroom was austere and Rumpole (the Chief Prosecutor) was deep in a cryptic crossword and simply didn’t pay any attention to me.

“I am going to teach him a lesson, and not just him, all of them” I thought to myself, as I scanned around the courtroom .

I confidently unpacked my LightPro and attached it to my laptop to present my findings to the court and straightened my tie. I was ready to prove the Hypothesis of My Workflow.

The Judge then entered the court and bid us all to take a seat, as he announced the intent of the proceedings.

“We are here today to validate a hypothesis that seeks to empirically demonstrate the power of a quantitative approach to Trend Following (specifically targeting the tails of the distribution of market returns) that can deliver sustainable long term returns”

The Lightpro illuminated the bold Hypothesis for all to read.

The Judge continued…….

“The author of this experimental test will now do the court a courtesy and present his argument.”

I stood up, bowed slightly to his reverence, and introduced the final portfolio.

“Thank you, your honour. Here I have a brief presentation of the fruits of my workflow process where I demonstrate my hypothesis using a backtest of a small but diverse systematic portfolio spanning 22 separate markets with 10 diversified trend following systems applied to each. This test is conducted over a 50 year backtest starting with an initial deposit of $50,000 using a 1% Trade risk of Equity to allow for compounding”.

I pressed the play button to my presentation which was going to blow the Court’s socks off, but was it the right music to truly represent this Masterpiece?

Rumpole however was still deeply engaged in his crossword puzzle.

The Barrister….and the Twist

The silence in the Courtroom was deafening.

“That music was the perfect accompaniment after all” I said to my colleague. “Sufficiently grand to inspire”

Rumpole stopped his crossword and the Courtroom was held  in suspense as he deliberated.

“Thank you for your presentation he grumbled. Most impressive. I even liked the elevating music or should I say elevator music.”

The Courtroom sniggered. Was this going to be another hammer blow by the adept Barrister?

“I almost even believed you had the Holy Grail in your hands for a brief moment…….but I see a slight snag in your argument”.

Guffaws are heard in the Courtroom and my heart starts to spasm

“You see you missed one small detail in your process logic. You presented to us an exhibit of an Equity curve arising from 22 separate sup portfolios over a 50 year history that attempts to validate your claim that this curve is a proof of robustness in your hypothesis.”

Exhibit A: Proof of Robustness

Rumpole continued….

“You go on to state that you undertake this entire process annually to refresh your systems and keep your portfolio razor sharp, but I see no evidence of this annual replacement in your equity curve that you have supplied with this exhibit.

You have simply presented us with the backtest results of your current selection of systems in your portfolio which is only valid for one year into the future. Your demonstration is not yet complete. In fact, it is woefully deficient”.

A Shattered Ego and the Need for a ‘Humble Mind’ Reset

Rumpole was right of course. I stood defeated, as I recognised the deficiency in my presentation. How could the moment change so suddenly from sublime awe to one of shattered delusion.

The answer lay in that brain of mine in which I was so sure played no role in this elegant systematic process.

Somehow that quantitative process I had become so attached to was found wanting…..and Rumpole knew it, as he had come across thousands of supremely confident quantitative traders before. He just needed to find a simple floor in the systematic logic that would shatter the argument and demonstrate this hubris for all to see.

He was, like the Market, going to teach me a lesson that all quantitative traders need to heed.

There was an essential step missing in the work flow. The final process of the flow just beyond Step 6. A piece of the process was missing.

1. Commence with say 40 uncompounded sub-portfolios of equal risk$ weighting;
2. Use as many of these Sub-portfolios as you can to compile your Global Portfolio within your finite capital restrictions. Ideally, we would use all we have available to us as market diversification particularly across asset classes is a big deal in delivering diversification benefit. Don’t worry about equal representation across asset classes as we let the data speak for itself through this iteration process. With say 50 years of data across say 30 to 40 markets, there is more than enough information embedded in each sub portfolios correlated relationship with the balance of the portfolio without the need for discretionary judgement;
3. Iterate to build the optimal risk-weighted Portfolio adopting the same process as we did for the constituent Sub Portfolios;
4. Return to your PC’s and select the best non compounded portfolio composite using the MAR metric;
5. Compound the result using your desired trade risk % which applies compounding treatment and leverage to your chosen solution. Adjust up or down this trade risk % to achieve a drawdown that is half your tolerance……..and then
6. Run the final portfolio test.

Here was the fatal missing step

7. Undertake this process annually to retire and replace systems developed through this workflow process. Allow active trades to naturally close but do not take new trades from the tired sequence. New trades are only allowed through the new ‘refreshed’ sequence of data mined systems.

The ego overrode the moment and, in my haste, to present my Masterpiece to the court, I forgot the smallest matter. This one tiny step in the process that makes all the difference. I forgot to annually revisit the strategy and undertake this process to project the performance results for the next year.

As systematic and non-discretionary as I thought my processes were, my brain had the last laugh and set me up for a lesson in hubris.

Fortunately, the mistake as it turns out, is actually an under-estimation, as opposed to an over-estimation of the power of the described Workflow in this Primer series.

By rotating systems each year to replace ‘blunt systems’ with new ones, the process lifts, as opposed to lowers, the equity curve. As we progressively compound the results, the impact of this annual ‘sharpening of the portfolio’ lifts the curve over annual intervals and we no longer see the obvious ‘recency’ acceleration in the curve as depicted in Exhibit A above. Rather we see a far stabler equity curve over its lifetime.

Now why aren’t I presenting you this effect to demonstrate what I mean? This would make the proof all so much more powerful. The performance metrics would rival the benchmarks of the Professional Funds. It could even makes us famous?

Because the lesson this Primer is teaching is a necessary lesson of humility. The Primer was building the momentum of excitement and ego, which was challenging the efficacy of the process.

Rumpole saw it for what it was. In fact he had seen it again and again from a legion of quantitative traders who are attached to their models.

Not a rigorous proof of concept, but rather a statement of hubris. He was literally more interested in a cryptic crossword to what I was planning to say after all. He knew his craft. I unfortunately did not know mine. I failed to recognise the role a brain actually plays in a systematic process.

The Real Reason for the Missing Step

The real reason for the essential missing step in the process is that to undertake the process of annual rotation of systems at both the sub portfolio and portfolio level over a 50-year backtests would have my PC’s working for many years.

I simply do not have the time or the inclination to direct my PC resources to this endeavour to further validate the hypothesis of this Workflow process. I am suitably convinced with the processes I can demonstrate now with a single generation of systems from a 50 year rigorous workout, without having to go that extra ‘exhaustive step’ to prove it.

There comes a time where you have to make a decision to step into a live environment and take your ‘possibly slightly’ deficient models with you into the fray. There is always modelling risk lurking there somewhere. Learn to live with it, otherwise you could become a victim of ‘paralysis by analysis’.

As it currently stands it takes about a month to undertake this Workflow process each year, and this only gives us a years’ worth of life out of the output results. However, I hope that I might have convinced you in this series so far that the Workflow process is worth the effort.

Another reason for not demonstrating a proof of concept for the past 50 years is that I don’t have a data set of 100 years to allow me to calculate the first annual rotation of the 50 year test series. I only have 50 years of data in my universe which at best would allow a rigorous result to be generated over a 25 year horizon.

Furthermore, a backtest on historical data only provides limited information that can assist in navigating an uncertain future. It is a bit like asking the question at 57 years old, will I live till I am 100 using my 57-year history to make that assessment. History only has limited value in the information it provides when embedded in an uncertain future.

From the point of this ‘Now’, there is only one historical record behind you but there are an infinite array of possible future paths ahead of you.  A backtest simply doesn’t pick this nugget of wisdom up.

The Brain Versus the Machine – A Last Word on Quantitative Workflow Processes

The moral of the narrative in this fictional Court of this Primer has been to strongly suggest that any quantitative developer needs to remove any pre-conceived bias that they may have in their models. Having this bias will only lead to failure with your systematic methods.  Despite the efforts adopted by quants to removing the interfering bias of the brain from the process, it is always lurking within the assumptions of the models such as the decisions made regarding which process driven steps we choose for our Workflow.

You simply cannot cover enough bases in a simulated systematic environment to validate any proof statement. Back-tests as part of this simulated environment are simply useful tools to help you assess what the past has possibly delivered to your systems. In fact there is still considerable material variation between a simulated environment versus a live trading environment. There is the psychology of the reality, the slippage of illiquidity and the errors of simply being human that lead to this material difference.

Have no doubt. Always opt for a live track record rather than a simulated one, when evaluating system performance. The difference is chalk and cheese.

However, on an even deeper level, whether simulated or live, no one and I mean no one can predict the outcomes of the next day, let alone the most profitable future path into an uncertain future.

So here we are after the humiliation of facing our own Hubris. We humbly accept that while we can come very far in our modelling processes using the privilege of hindsight  (‘look-back’ methods), we will never know how far or how close we are to the actual reality.

But is a systematic workflow process that uses the ‘data to make all the major decisions’ better than a more discretionary path to portfolio development?

It certainly appears to offer sound logic in the process that avoids most biases lurking in a human brain, so surely it must be a better way to model this complex system we call the Market?

Well unfortunately we find that all models have their weaknesses. There is this thing called reality out there, but there are only ever models that we can use to describe it.

I have spent the last six Primers leading you through a quantitative workflow process that seeks to prove our trend following hypothesis, but as close as we may feel that we are approaching an answer about reality using some heavy computer power, quantitative methods can be as guilty as discretionary methods in their assumptions about any reality associated with a complex system.

You see there are no fully non-discretionary quantitative methods. They are all ultimately steered by the fallibility of a brain. The biases that arise from this fallible organ inevitably contaminates our models to some degree.

The heuristic limitations of the brain that create illusory bias in our discretionary model making (further described in Primer 2 of this series), has now led to statistical biases of data mining and curve fitting in our quantitative methods that stifle our ability to correctly match any interpretation with the reality.

People often accuse me of being a ‘brain bully’ where they see my intent as ridiculing that grey soft squishy stuff that lies in our skull but ‘au contraire’ my friends. This amazing organ that has been crafted over deep time within a natural environment is beautifully aligned with that environment. It has been perfectly ‘curve fit’ for purpose through eons of applied selection processes. Give me a brain anytime as opposed to an artificial intelligence (AI) that seeks to replace it’s amazing capabilities. ‘AI’ has got such a long way to go before it can challenge the master at interpreting the universe that surrounds us.

Now, provided that the human mind remains as the dominant contributor to the machinations of these financial markets, then there is a lot of merit in discretionary approaches, provided that any bias that resides in that human brain of ours is understood and its impact in our financial markets.

We believe there is always a necessary role played by a brain in the development of a workflow process. So much so, that we leave the hard stuff of process design to the brain itself. The brain decides what the steps should be and why and the brain decides what limitations to place around the experimental method and the assumptions we make for our model making.

The quant stuff which is embedded in this workflow process is simply harnessing what a computer, processing binary bits, can do extremely well to take lots of the heavy lifting away from the brain in those areas it simply cannot compete.

So, we view our workflow processes as the combination of what the brain does best and what our information processing systems can do best. We feel that you simply cannot get better than that. It is going to take a long time before anyone can exactly copy what we do. Any small edge that resides in our workflow process is protected for a time.

However, this final word on ‘brain versus machine’ has a caveat. We touched on some of the incredibly important biases in our statistical treatment using our machines that our brain may not see, and we can therefore be blind to. The biases of over-fitting and the bias of selection processes used in the workflow process itself. Over-fitting is a curse to the machine that seeks to investigate the reality of complex financial markets as is selection bias arising from decisions made in selection between competing return streams.

The question is, do these biases that we know of which creep into our statistical treatment, outweigh the benefits of moving towards this more systematic path of portfolio development using quantitative methods? If everything was simply left to the brain, as limited as it is, are we any better off with the workflow method? I mean Einstein was able to crack relativity without the use of a machine, and the machines today still cannot work out what the brain can magnificently simplify.

Well, we believe that there still is a small edge in our workflow that allows us to better distinguish ‘the Signal within the Noise’, provided we let the brain steer the process and that we take all measures to identify any biases that may arise in our statistical treatment and avoid them. In tandem, we believe the brain plus the machines can simply make better models that interpret the reality out there.

But here are a few guiding pointers to help to protect you from adverse bias that may arise when using statistical methods to find that damned elusive ‘signal’ using quantitative methods such as that delivered through this workflow.

• Adopt Design First Logic using the brain as opposed to ‘hocus-pocus’ optimisation methods. Configure your system through Design Principles to deliberately respond to a market condition you are targeting. Do not let an automated system generator decide for you. It will over-fit to the data which ‘mostly is noise’.
• Avoid where possible choosing your systems to use based on the most profitable equity curve. If there is randomness in a portfolio equity curve, which we know there is in a ‘mostly efficient market’, you are only exacerbating ‘fictional’ performance as you are removing the adverse random elements that make up the equity curve in your selection process.
• Most Monte Carlo methods applied to trade results themselves (as opposed to market data) tell you nothing as they disrupt the serial correlation in the equity curve (related to the market signal) and not the noise element of the equity curve (related to the randomness in the market data). They turn the entire equity curve into noise. So don’t use them as methods to test for robustness.
• Walk Forward should only be used for Predictive techniques, that are seeking to capture a repeatable market condition. They are not applicable for trend following or momentum methods. A nice straight equity curve for a single system can only be maintained if market conditions remain favourable over the entire extent of that equity curve. That is only relevant to predictable methods for the period of time that a predictable market condition persists, and that can be very short. Every equity curve must display periods when they are performing and periods when they are under-performing as we know that no system can address all market conditions. Be suspect of straight equity curve for a single return stream.
• The market condition determines your fate. A good system simply allows you to extract the elusive signal from otherwise noisy market data. Bad systems……well they are just bad in every shape and form.
• Straight equity curves over the long term are the result of a portfolio comprising different successful systems attacking many different conditions. Trend follower’s equity curves are superior to ‘predictive equity curves’ as they avoid the cost of failed predictions.
• Always data mine using the most data you can get your hands on. This is the only way you can reduce the impact of randomness in your overall performance results. Data mining over a few years of data is asking for trouble.
• Avoid getting too attached to your Models. Your brain and the way it likes to assign causality to a reality is the ultimate weak spot in your quantitative models, despite the fact that you feel you have objectively eliminated this weak spot through your systematic processes. Always treat any outcome derived from your quantitative processes with scepticism. There are always better models out there…….so keep modelling.

I have now demonstrated how we here at ATS use Data Mining to apply our Trend trading philosophy. Some of the systematic workflow processes we apply may be new to many trend traders who have a different way of interpreting their craft, but when you dig down into the weeds of this process, you can probably see that the core logic of trend following safely resides within it.

We are coming to the end of of this Primer…..and I am quite literally running out of things to say that might assist a trader who wants to pursue the Trend Following path.

Just the ‘Conclusion’ to go…….

Conclusion: All Things Come to an End, Even Trends

All things come to an end, even this Primer, but what a ride it has been.

We have explored the very many facets of a trend following philosophy from a narrator, who is dedicated to this cause.

This is just one interpretation of the “Trend Following Way”, as there are just so many more others. We can pick up pieces of this philosophy from a host of legendary traders, philosophers and scientists who have discovered that there is never any permanence in this vast complex system of ours, only change.

Having this Primer Series in electronic form is of great value to me. It allows me to continue working on this series via updates as new ideas and new opinions evolve over time. Such flexibility is not offered by a publication with a due date.

I hope you enjoyed the read up to this point in time. It only represents one opinion, but at least now that I have this opinion in writing, it provides a journal and a reminder for me of how I eventually came to this very strong viewpoint about Trend Following. I can comfortably say that it is the most superior method of trading these markets I have ever encountered over my now long trading career.

I am getting on now and unfortunately I wasn’t able to jump on board this gravy train in my youth, where I could have taken full advantage of the miracles of compounding, however I feel that this series might give a leg up to those more youthful types, where they can apply some of these principles to make their trading careers more fulfilling and sustainable.

This journey has explored the mind, the natural world, the financial systems and a class of trader who are embedded in this epic riddle, namely the diversified systematic trend following community. I feel very close to this group and hope that I have done a service to them with this series.

Trend following is far more than just a way to trade the financial markets. It is a recipe to lead a fulfilling life. It gives us a way to deal with the very thing we most fear. Change.

Most of our lives, we struggle with this concept. We want to hold on to the familiar with our dear lives and control every moment within it.

But there is an alternative way to deal with change and that is to embrace it. Go with the flow and recognise that we are just a small piece of this majesty, we call this complex system….our universe.

Will trends die like many tend to think. That is a bit like asking, will the planet stop changing, will society stop changing and will the financial markets stop trending?  There will of course be long periods of time where the surf is mute, but if we are patient and just sit long enough on that beach somewhere, the swells will inevitably roll in. You can count on it. This isn’t a trading statement, this is a universal statement. We will all be long gone before the Trends end.

Follow the trend my dear friends, until of course it all must end.

Many thanks for stopping by and reading this journey. It has been a blast but look at the time…..I will see you to the door with one of my favorite ‘thinking and melancholic tunes’ once again from Paul Schwartz with thanks going out to those that inspired this writing.