In the Beginning there was Trend Following – A Primer – Part 18
Trend Following Primer Series – Compiling a Sub Portfolio: A First Glimpse of our Creation – Part 18
Primer Series Contents
- An Introduction- Part 1
- Care Less about Trend Form and More about the Bias within it- Part 2
- Divergence, Convergence and Noise – Part 3
- Revealing Non-Randomness through the Market Distribution of Returns – Part 4
- Characteristics of Complex Adaptive Markets – Part 5
- The Search for Sustainable Trading Models – Part 6
- The Need for an Enduring Edge – Part 7
- Compounding, Path Dependence and Positive Skew – Part 8
- A Risk Adjusted Approach to Maximise Geometric Returns – Part 9
- Diversification is Never Enough…for Trend Followers – Part 10
- Correlation Between Return Streams – Where all the Wiggling Matters – Part 11
- The Pain Arbitrage of Trend Following – Part 12
- Building a Diversified, Systematic, Trend Following Model – Part 13
- A Systematic Workflow Process Applied to Data Mining – Part 14
- Put Your Helmets On, It’s Time to Go Mining – Part 15
- The Robustness Phase – T’is But a Scratch – Part 16
- There is no Permanence, Only Change – Part 17
- Compiling a Sub Portfolio: A First Glimpse of our Creation – Part 18
- The Court Verdict: A Lesson In Hubris – Part 19
- Conclusion: All Things Come to an End, Even Trends – Part 20
Compiling a Sub Portfolio: A First Glimpse of our Creation
In our last Primer, 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.
- We need to compile our return streams into Sub-portfolios (or discreet markets); and
- 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 16th 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.
Enjoy….and by the way Cheers
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.
Stay tuned for our next installment in this Primer Series.
Trade well and prosper
The ATS mob