Robert Carver introduces his 2015 book, Systematic Trading: A Unique New Method for Designing Trading and Investing Systems, by stating that: “I don’t believe there is any magic system that will automatically make you huge profits, and you should be wary of anyone who says otherwise, especially if they want to sell it to you. Instead, success in systematic trading is mostly down to avoiding common mistakes such as over complicating your system, being too optimistic about likely returns, taking excessive risks, and trading too often. I will help you avoid these errors. This won’t guarantee returns, but it will make failure less likely. My framework…can be adapted to meet your needs. …Each element of the framework has been carefully designed… I’ll explain the available options, which I prefer, and why.” Based on his experience as a trader/portfolio manager and specific research, he concludes that:
From Chapter 1, “The Flawed Human Brain”: “Given there are flaws in the human brain which seriously affect our decision-making ability, what can we do about it? Easy: we should use systematic trading rules to make our decisions. These can help mitigate our own serious flaws, but they can also allow us to exploit the weaknesses that other human traders have…”
From Chapter 2, “Systematic Trading Rules”: “In my experience consistently profitable trading comes out of careful research, done by thoughtful and knowledgeable people, who seek to understand where their profits come from. The loss making systematic trading I’ve seen has often been the result of haphazard data mining, done without any consideration of the reasons why a rule might have appeared profitable in the past or might not be in the future. …many traders have highly unrealistic expectations of back-tested Sharpe ratios of 2.0, 3.0 or even higher… These values are far too optimistic and are caused by over-fitting… In reality SR consistently greater than 1.0 are rarely achieved… …a balanced combination of trading rules, with different styles that work in different environments, is better than any single alternative.”
From Chapter 3, “Fitting”: “…the more complex the method the harder it will be to realise when you are over-fitting. …reducing the pool of trading rules and variations considered is crucial unless you have many years of data. …always use rolling or expanding out of sample fitting. …Only if there is a statistically significant difference in performance between various rules across instruments should you fit them [the instruments] individually. …comparing a positively skewed rule with a negatively skewed alternative purely on SR is highly misleading, as negative skew rules will tend to have flatteringly high SR. …Be careful of focusing on outright performance, rather than returns relative to benchmarks.”
From Chapter 4, “Portfolio Allocation”: “In the early part of my career I was fatally distracted by the lovely [mean-variance optimization] equations and ended up with some terrible portfolios… How can we fix this problem? I have two techniques that I use. The first, which is quite hard work, is called bootstrapping. This involves repeating my optimisation many times over different parts of the data, taking the resulting weights, and averaging them out. …The justification for bootstrapping is simple. I believe that the past is a good guide to the future, but I don’t know which part of the past will be repeated. To hedge my bets I assume there is equal chance of seeing any particular historical period repeated. …The handcrafting method…involves constructing the portfolio in a bottom-up fashion by first forming groups of similar assets. Within and across groups you set allocations using a table of optimal weights for similar portfolios. These weights come from my own experiments with bootstrapping.”
From Chapter 5, “Framework Overview”: “Trading rules are the engine of the system. These give you a forecast for instrument prices; whether they are expected to go up or down and by how much. In a car the chassis, drive train and gearbox translate the power the engine is producing into forward movement. Similarly, you will have a position risk management framework wrapped around your trading rules. This translates forecasts into the actual positions you need to hold.”
From Chapter 6, “Instruments”: “…have some idea of the factors driving returns. If unusual forces are at play then avoid that instrument. …Volatility must not be extremely low. …hold the largest portfolio [number of holdings] you can… Assets with strong negative skew need careful handling and shouldn’t dominate your portfolio.”
From Chapter 7, “Forecasts”: “…in my framework forecasts are proportional to expected risk adjusted returns. …So expected Sharpe ratios make good forecasts… Forecasts should be capped at a maximum absolute value… You shouldn’t change your forecast once a bet is open. Instead you will be using a systematic trailing stop loss rule to close all your positions. …avoid including any two variations with more than a 95% correlation to each other… …exclude anything which trades too slowly, or too quickly.”
From Chapter 8, “Combined Forecasts”: “…what if you have multiple rules and they disagree? …In the framework you need to use a weighted average of your forecasts… How do you find the best weights to use when combining forecasts? This is an example of the problem of allocating a portfolio of assets, which we discussed in chapter four. …You need your combined forecasts to maintain the same expected absolute value…as you required for individual forecasts. To fix this the combined forecast is multiplied by a forecast diversification multiplier.”
From Chapter 9, “Volatility Targeting”: “Deciding your overall trading risk is the most important decision you will have to make when designing your trading system. …I use a single figure to measure appetite for risk–an expected standard deviation, which I call the volatility target. …I find it’s easier to look at an annualised cash volatility target, which will be the annualised expected daily standard deviation… Nearly all professional gamblers, many professional money managers and some amateurs in both fields know that this optimal point [position size] should be calculated using something called the Kelly criterion. …using the full Kelly criteria is far too aggressive, because of the risk of getting a poor run of luck… It’s far better to Half-Kelly and set your risk at half the optimal.”
From Chapter 10, “Position Sizing”: “Using recent price volatility is generally a good way to get the right risk adjusted position size. …you need to know the volatility of the instrument value in the currency of your account…”
From Chapter 11, “Portfolios”: “It’s best to simultaneously run a portfolio of as many trading subsystems and instruments as possible and allocate your capital between them. …asset allocating investors can use handcrafting or bootstrapping [Chapter 4] to find weights. Correlations can be estimated from a back-test of trading subsystems, or using rules of thumb. …Semi-automatic traders should use equal instrument weights… If the current position is within 10% of the target then you don’t need to trade…”
From Chapter 12, “Speed and Size”: “…it’s much better to trade systems that aren’t vulnerable to…high levels of costs. …You should use a measure of costs that accounts for how risky different instruments are…how much of your annualised raw Sharpe ratio (SR) you’ll lose in costs for each round trip. …you’ll need to count the number of round trips done annually [turnover]… …considerable evidence is required to justify trading a faster, and apparently superior, variation rather than a slower, cheaper and supposedly inferior alternative… …I recommend that you set a speed limit; a maximum expected turnover you will allow your systems to have. …I recommend that the costs of an instrument’s trading system should be at most one-third of a conservative estimate of the pre-cost Sharpe ratio…”
From Chapter 13, “Semi-automatic Trader”: “…semi-automatic traders…make their own discretionary forecasts about price movements, but then use my systematic framework to manage their capital and position risk. …With stop losses, risk targeting and position sizing taken care of you can focus on getting the buy or sell decision correct. This will give you the best of both worlds-your human ability to interpret and process information, combined with a system giving the correct amount of risk. It will not be easy sticking to the framework. The system may force you to trade, or prevent you from trading, when you would rather do otherwise.”
From Chapter 14, “Asset Allocating Investor”: “…an asset allocating investor [believes] the best returns can be obtained by investing in a diversified portfolio of assets without trying to predict relative risk adjusted returns…you think all assets will have the same Sharpe ratio. …I assume you’ll be using exchange traded funds (ETFs)… I am assuming that you’re not comfortable, or unable, to use leverage or to do short selling. …In practice rebalancing could be done more frequently [than weekly] for large funds or in times of market stress, others may choose to rebalance quarterly…, and amateur investors might be happy with annual rebalancing…”
From Chapter 15, “Staunch Systems Trader”: “…the staunch systems trader [uses] both the framework and multiple systematic trading rules for predicting asset prices. …The specific example…will cover futures trading… I will assume that you have $250,000 of initial trading capital. The example will focus on building a system suitable for part-time traders which trades daily and for which data can be obtained for free. …choosing which futures to trade is mostly a balance between getting reasonable diversification and running into the issues with large minimum instrument block sizes… You should avoid instruments…with low volatility.”
From the Epilogue, “What Makes a Good Systematic Trader?”: “…humble…simple trading rules…sceptical…pessimistic…thoughtful…thriftiness…nervous…diligent…lucky…”
In summary, investors will likely find Systematic Trading a rational and practical approach to building diversified, risk-managed investment/trading portfolios.
The book offers quantified examples throughout.
Cautions regarding conclusions include:
- As with much research on asset pricing (and noted in the book), the author assumes that return distributions are tame enough to support a “normal” interpretation of distribution statistics. To the extent that actual distributions are wild, these interpretations break down.
- The book does not address actual performance achieved by applying the framework presented. The closest approach to this disclosure is in Chapter 9: “I run a highly diversified futures trading system with around 45 instruments, eight trading rules drawn from four different styles, and 30 trading rule variations. In a 35 year back-test, conservatively fitted with out of sample bootstrapping, it has a Sharpe ratio (SR) of around 1.0 after costs… …for an out of sample bootstrap, as I’ve used in my own system, a ratio of 0.75 should be applied to find a more realistic Sharpe ratio. …the highest volatility target I’d advocate for it is 37%…” But, how does this trading system actually perform?
