Basics of Portfolio Construction for Price Followers – Introduction Part 4 – Summary So Far

The purposes of this introductory thread is a bold one. It is to convince you that we are onto something special. It is to provide a vast array of different contexts to appreciate this shady art of systematic trend following that changes the way you might see things if you have a predictive mindset.

The reason we first explored the differences between price following versus price prediction is that the methods deployed in these two dominant camps that relate to trading styles are the antithesis of each other. One relates to the notion that markets are predictable in nature and that alpha can be extracted from the markets during these predictable times. That is of course right…..however it fails to take heed from a risk management perspective when markets become unpredictable or noisy.

On the other hand, as a price follower who doesn’t attempt to predict, all their emphasis is placed on risk management to preserve their capital while they wait for the unpredictable events to arise. So they manage risk at all times and across all possible market conditions but just leave themselves open for when unpredictable markets can deliver outsized rewards from time to time.

The differences between mindset effectively state which way your trading bias is inclined…..towards profit outcomes or towards risk management outcomes.. Now we know that the market caters to both these mindsets in that we can see the result in the way a system’s distribution of trade returns are biased around a normal distribution.

Predictors blossom with positive kurtosis…..while price followers blossom during negative kurtosis. Neither blossom however when markets are perfectly efficient with zero kurtosis. This is why it is very worthwhile plotting your distribution of trade returns to understand what is responsible for your overall profitability or overall loss in the long term.

Now when markets are predictable, a predictive trading system blossoms with a steadily rising equity curve. There is no extrinsic sign that risk exists in that system while predictive conditions exist. As a result, the predictive mindset starts to let go of the risk management reigns given that all appears to be going well and according to plan.

On the other hand a price following mindset that is capitalising on the fat tail at the right of the distribution of returns has no idea when that unpredictable result would unfold. The only assumption that they apply is that ‘markets are unpredictable from time to time and deliver a more extended price move than would be allowed for by a normal efficient market’. This is backed up by quantitative finance models which suggest that a normal distribution does not faithfully apply to all markets and that  better models exist that describe a real market such as models that include levy processes. So a Price Follower keeps a continuous grasp on managing risk at all times….specifically the risk that uncertainty brings.

We all understand that complex systems exhibit non-linear power laws from time to time. The shape of a market distribution oscillates between conditions of predictability when techniques such as mean reversion work as price oscillates within a defined extent around a known equilibrium point….. and times of unpredictability where price has a biased unidirectional tendency with the added impetus of power laws towards a future unknown point of equilibrium. Transitions typically overshoot the market from the power law applied to momentum before settling down to a new future equilibrium state. During these periods of market transition, predictions fail and price following techniques tend to work.

So let’s drill down a bit into what levy drift is and how this feature is more prevalent in outlier zones where participant activity is less and competition is reduced to those that apply price following methodologies.

Now in quantitative terms a market can be described as efficient  and indicative of a normal distribution, or non efficient when trade results plot outside the normal distribution of returns. A predictive market exists when future price moves to an estimated location which is anticipated by a predictor. It is based around a principle of mean reversion to some known future point. The risk embedded in this statement is ‘convergent’ in nature. The price bias in the future converges to a known point. Because price is said to oscillate around an equilibrium, this oscillation is frequent but the price moves are finite and tend to be of a shorter extent.

A Price follower on the other hand assumes that risk in the future is unknown and is divergent in nature. In other words, future price holds very little regard to what has occurred in the past. Given market uncertainty, price can move to any future level provided it is away from the convergent outcome that oscillates around the notion of ‘an intrinsic value’.  The price of the future could go anywhere and is undefined in extent. This open ended condition is what creates a unidirectional bias towards the price move and can be of any extent. As a result, given the Law of Large Numbers, the price move may be very large.

The fundamental principle that demonstrates whether you are a price follower or predictor is based on the regard you hold about whether price is always right or whether price is regarded as being undervalued or overvalued and that you can predict what ‘intrinsic value’ is actually worth. The predictive clue is based on the regard you have to price as a representative arbiter of value.

So back to levy drift. When we attempt to model how a modern efficient market behaves we can quantitatively sum up the price movements into a particular class of distribution referred to as a levy process.

It is simply a theoretical model that describes price movements in random terms but highlights where randomness is not continuous but discontinuous in nature. Now the non-continuous nature of change is omnipresent in any complex system. It is the large outlier move even in random systems that dominates in the long term over the average direction of all price moves. In natural systems, change is thought to occur not as a continuum but in discontinuous phases referred to as punctuated equilibrium. A price follower is said to predate on this feature of complex systems, namely that most of the gains to be observed from the PL can be attributed to the outlier as opposed to the otherwise random outcome arising from their combination of the quantum amount of the wins and losses during less exotic market conditions.

A Levy Process is graphically illustrated by the diagram below. The step lengths of random movement are not equal in nature. Now while this represents a class of random behaviour when you also assume that randomness across the time dimension exists, it’s pattern of outcomes are remarkably close to how a real market behaves. We get concentrated clusters of price moves during predictable times and extended unidirectional bias arising from the unpredictably price extension in any direction away from these clusters of predictable outcome. This behaviour is a great way to view these markets.

Now disregard the requisite randomness as this is not the basis for discussion and look at the pattern itself as random or not, it provides a basis to understand why price following works in real markets that exhibit this behaviour.  The reason I have highlighted zones in red and green is to signify in red where predictions fail and in green where predictions blossom. In predictive market conditions where price oscillates around an existing equilibrium point, the step moves are short in nature. The price moves therefore cluster in a predictive environment. To be a successful predictor in the long term you need to have a trading strategy that has a high frequency during these clusters, but recognise when markets transition from these zones, you need to identify when this occurs….and furthermore you need to quickly be able to exploit another zone of clustering when it arises. You may be very successful during these clustered zones…..but your losses will occur when price extends beyond these clusters or in your inability to predict the nature and extent of another predictable condition in a future cluster.

So in the long term for a predictor, you need to have a successful method of when to turn on or turn off your predictive strategies and furthermore have sufficient profitable predictive models at different times in the future to pay for the losses incurred in your failure to correctly predict the outcome.

It is the discontinuous nature of predictable versus unpredictable condition that is the nemesis for a predictive mindset. It is also the central reason for why we cannot find any very long term examples of FM’s that have been successful in applying predictive systems, apart from a small handful of players with very large quant research houses such as Renaissance Technologies, Two Sigma and Winton Capital.

In the past of course we had Long Term Capital management that comprised a host of Nobel prize winners in their ranks….but we know what happened to them when prediction failed.

Now let’s focus on the conclusions of a Price Follower and the reasons for the very large number of very long term survivors who exist in this particular philosophical space.

1. Markets are simply assumed to move between noisy phases (efficient markets), predictable phases and unpredictable phases. Price Followers do not predict where price will be in the future as we concede that it (the price) is always right, however we always mitigate our adverse risk exposure for when we are wrong…..which is frequent. A predictor however must by definition be prepared to allow for adverse price movement before their predictions come into effect. They recognise that timing of their prediction can be the problem….hence they allow for intrinsic risk to build to ‘get their predictions right’.

2. Price Followers simply view price as the Gross indicator of an array of value transactions between a vast and complex array of different participants of different sizes and opinion. Wherever price is today it is right. We will therefore shadow current price at all times because it is always right with an asymmetrical trap that allows open ended profit potential but cuts losses short at all times between entry and exit. It ensures that we stay close to price as we progress into the uncertain future and never stray too far away as we do not have an opinion of whether price is ever wrong.

3. We avoid trading during those times when predictive techniques blossom by using filters to avoid noisy or predictable market condition. We therefore restrict our trading activities to those zones where the impact of the unpredictable outlier is felt more. Outliers emerge at any time and space in a complex system, but to avoid the noise and predictable oscillations we simply focus on those extreme moves where the unidirectional price extension has more impact w.r..t more normal conditions. We are therefore choosing the right place to attack the anomaly rather than all places. Furthermore we use asymmetrical traps that can lie in wait to catch the anomaly if and when it occurs at all times. We are therefore available to the anomaly 24/7. The outcome is therefore where we catch an anomaly or not as opposed to simply watching and doing nothing when an anomaly occurs.

4. So let’s now go back to that diagram of Levy drift. We design our trading systems to avoid the clusters but lie waiting in those zones where the extended anomaly prevails. We don’t know if the direction of the anomaly is going to be long or short…so we need to take a 50% punt in our system design that it will be either way. We bias the outcome in our favour however by trading in the direction of the primary price bias (the long term trend). The times we get it wrong we are immediately out of the trade with full risk release through our system design.  We never warehouse risk. Furthermore our diversification in our models ensures that we bet small in each trade event so even if our system design fails to exit the trade with an initial stop or trailing stop condition, the impact to our PL from being wrong is never large.

5. So by being in the right place (the outlier zone) and the right time….we prevail when the anomaly is of sufficient extent to pay for all the mistakes we have made along the way in refusing to be predictive. We just need to ensure that our systems are open ended in design and can capture the bulk of the unidirectional price movement that an anomaly can deliver from time to time. Diversification again is the major way that we can increase our trade frequency  to catch these seldom and unpredictable outlier events. A single system alone is not likely in the long term to survive the impacts of all the whipsaws in trying to catch these anomalies. We must be diversified.

6. Financial markets are all connected but we don’t typically see it that way. An anomaly may exert it’s impact across markets and timeframes by virtue of their loosely connected nature. For example, during bull markets phases, the individual markets tend to call their own tune. We have some markets performing strongly and other markets performing weakly. Anomalies however have the ability due to their impact on participants to cascade across timeframes and the markets themselves. For example the 2008  GFC was not restricted to the mortgage markets. This anomaly cascaded across all market sectors that were thought to be previously uncorrelated with each other. ….but don’t think this is restricted to anomalies arising from fear. Anomalies can arise from the collective behaviour associated with greed as well.

Consider the impact of QE programs since 2010 on the financial markets. The release of gargantuan fiat money has to reside somewhere. It was easier to see where the can of worms was hiding in 2008 (namely the mortgage sector), but the QE release now could be hiding anywhere and everywhence. In fact it is probably everywhere and may be a reason for why the markets appear to be all so correlated today. Who knows….it may be an asset bubble of cataclysmic proportions across all asset classes? The bottom line however is that a Price Follower does not care. In following price the outcome will be known in hind-site later on. We prefer the potential for profit while managing risk at all times as opposed to educated guesses.

7. For a price follower who focuses on catching anomalies while restricting adverse losses, if they are diversified across timeframes, markets and systems, then they can catch the windfall when markets become correlated in nature from time to time. That is why the Price Follower blossomed with say 80% returns during the 2008 GFC event while the predictive community suffered horrible drawdowns and many met risk of ruin head on.

8. If we view things from a competitive viewpoint….the story goes like this. During predictive market times, predictive techniques blossom. Competition using predictive discretionary and automated systems blossom such as HFT etc. The very finite alpha that exists from the predictable market condition needs to be shared amongst an explosion of predictive participants. They do well….but only as well as finite competition allows. The most successful predictors therefore do well but to a defined extent. They are significantly capped in their profit expectations by virtue of the predictive community they need to share the finite spoils with.

With the increase in predictive methods, alpha for predictive technologies naturally depletes but the coordinated efforts of the mass of speculative predictors extends price movement in a singular direction away from it’s natural equilibrium which predictors hitherto have been pin pointing. The markets therefore become overheated through competitive actions of its predictive participants. Price followers during this phase manage to pay for some of their losses during these asset bubble phases….but as every predictor and their dog are making money…..there is a lop sided bias exerted in the participant mix which significantly departs from any notion of underlying fundamental value basis. Complex systems are elastic however to a certain point but become increasingly fragile until a critical threshold is reached….where a small seed that disrupts the elastic emergent structure of the current market  such as a single failure in it’s 1 to many relationships (eg, an institution, an economy or a bank) can exert a critical failure impact that then unwinds the intrinsic risk buildup of that emergent structure. It’s impacts can cascade across timeframes and markets. The swathe of predictors are now caught on the wrong side of the market bias. It is during the inevitable transition that is a common feature to all dynamic complex systems ….where predictions fail.

9. Now unlike the Predictive community who are vast in number during predictable times, the Price Following community is far less in number and a very small part of the speculative participant mix. Why so few in number…..because it is just so hard to stick to. Everyone out there wants to predict while conditions are predictive. The Price Follower just sits back and accepts the inevitable drawdown that occurs over this phase.

The weaker of the Price Followers with a wavering philosophical mindset will succumb as they watch the predictive systems take the market spoils while they incur stagnation or worse still drawdowns. As a price follower, it is just so psychologically tough to live with a large number of small losses and building drawdowns and the temptation to leave the Price Following camp is significant. So the numbers of Price Followers particularly during predictive conditions declines…..but then when the spoils of prediction are over (as the alpha of prediction is transient)…the game suddenly changes where the market spoils become available to a small population of patient speculative participant with a strong philosophical mindset….namely the ardent Price Follower.   These transitions are generally quick but significant to the PL of the Price Follower. Quick enough to ensure that the predictors cannot change their philosophy to price following to steal part of this windfall for the patient Price Follower.

Now the reason for the size of the impact on the PL is that we get a significant disruption to the participant mix of speculators.  in those rare moments when prediction fails, we start to see a different form of behaviour emerging from the traditional predictive camp. They drop their predictive notions and start exhibiting coordinated behaviour of panic during adverse events and greed during favourable events. This coordinated behaviour that emerges from the  predictive camp ‘dropping their predictive tendencies’ causes these unidirectional anomalies. You can see this occur when markets are said to capitulate.

10. Sitting on the other side of this ‘now’ coordinated behaviour is the Price follower, who has simply been following price all along. They lie on the other side of the transaction. Being far fewer in number than their predictive cousins, there is a transfer of wealth from the predictive pie to the price followers pie….and given the scarcity of price follower, their share of that pie is just sooooo big. This is where the positive favourable outlier exerts so much dominance to the Price followers PL.

Now with a predictive mindset you would prefer this repeatable pattern of oscillation. Let’s face it we would all love this degree of certainty. This arises from the coordinated impact of a predictive trader……

….but sometimes the market can do this at times where prediction breaks down. A Price Follower makes their alpha from the unpredictable price move where predictions fail exhibited by this example. We all hate this unpredictability below…but that is why the edge is so persistent and enduring for those traders who design their system to accommodate this uncertain pattern of market dynamic.

Complex markets are called complex for a reason. There is no central mean of predictability. There are many means and many transitions between means given the complexity that exists in the non linear relationships that exist within a complex system. Markets move between states of predictability and unpredictability and noise is omnipresent but particularly noticeable during normal market conditions….but the outlier that has so much to say in a traders long term success or failure reside in the fat tailed movement associated with uncertainty.

So let’s call it a day with this summary. I hope it provides a very justifiable reason for why we prefer to adopt Price Following as opposed to Predictive methods in our system design.

Now that we have discussed system design for a price follower at length you can see that it is the broad principle of design, namely cutting losses short and letting profits run’ that is important as opposed to specific fine tuning that a predictor might deploy to catch a future price move with precision. By addressing these broad principles with a simple design outcome that addresses all forms of uncertainty you can see how you can deploy many different types of asymmetrical trap to achieve a Price Following outcome.

In fact….as we will observe shortly when discussing specific methods of diversification adopted by Price Followers in the forthcoming editions of this basic introduction to Price Following, you will see how very much diversification matters. It is within the methods of diversification that the edge of a Price Follower really resides.

Trade well and Prosper

The ATS mob

Want to know more? Then join us at ATS Trading Solutions. We offer unique services for the retail trader wishing to learn more about the techniques applied by the systematic diversified trend followers to generate wealth building sustainable returns. Membership includes access to all Programs and Trading Tools available in our shop, systematic development and portfolio construction education and practice and participation in our Members Forum. 

For a short time only we are offering free access to the exclusive areas of our website reserved for paying members with 100 free vouchers on offer for the first month, so you can have a good look around and get a feel for what we offer. Just go to the membership page of our website and choose the Premium Membership Plan and proceed to checkout. On the Membership Checkout page just click on the discount code and apply the code VIP100 to receive your first month free. So long as you cancel your membership in your first month then you won’t be charged a penny. If you don’t cancel in the first month it will automatically go to $9.99 AUD per month following your first free month.

Looking forward to seeing you over there. So come and join us and participate in our forum with other like minded systematic diversified trend following folk.

You must be logged in to post a comment.