One of the frequently voiced misgivings about systematic trading strategies usually goes like this: “yes, but your model can’t know if tomorrow XYZ happens and market prices go haywire…”

The immediate aftermath of U.S. elections saw fairly volatile market reactions particularly in equities and treasuries. Both had previously turned slightly bearish but on November 4, the day after the elections, prices rallied sharply higher. Such events tend to be adverse to trend following strategies and often raise doubts about the merits of systematic strategies since they can’t react to the changes in a timely manner. Many traders feel they could do better by anticipating the turbulence and maneuvering around the ongoing market events and see systematic strategies’ inability to anticipate such events as a weakness of the approach. A rather remarkable event in September of 2019 offered a good perspective on this issue.

Case in point: oil price and the surprise military attacks in Saudi Arabia

In July 2019, a prospective client of ours was planning to launch an energy focused hedge fund. They had asked me to put together a portfolio of long-, medium- and short-term I-System trend following strategies for all the major energy futures. We began tracking this portfolio on July 8.

On Saturday, 14 September a surprise military attack on Aramco’s Abqaiq production facilities in Saudi Arabia took place. On Monday, when the markets opened oil price staged the largest recorded one-day jump. A few months later I summarized the experience, which underscored two important issues related to systematic trend following:

• The impact of the unforeseen major event and,
• The problem of underperforming strategies

Both issues appear in the following chart:

Over the observed period, I-System strategies generated a loss of $6.87/bbl (at 22 Nov. market close), Over 84% of which occurred on the September 16. N.B. – in the above chart, only 16 strategy curves are visible as 4 of them fully overlap. Two elements in the chart stand out: (A) the price spike on 16 September and (B) the presence of an “outlier” strategy.

(A) The impact of an unforeseen major event

At market open on Monday, September 16, we had 60% short exposure in Brent (16 out of 20 strategies) resulting in an average loss of $5.80/bbl for the day. To be sure, this was a very unpleasant experience. However, unforeseen events that can and do affect market prices to varying extent happen almost every day – this is the essence of uncertainty that’s integral to the price discovery process.

But the idea that we could always correctly anticipate these events and react to them consistently to our advantage is unrealistic. I believe that disciplined adherence to a set of well-formulated, time-tested rules is more likely to yield positive results over the long term. In this sense, the inertia of systematic trend following is not a weakness. To the contrary, it may the crucial anchor preventing haphazard, emotional overtrading. We saw this clearly as we revisited the performance of our oil strategies through March 20, 2020:

Unaffected by September losses I-System strategies continued generating trading signals with an unchanging focus on achieving their objective: capturing profits from major price trends. Several months later, such a trend emerged and the strategies did and continued to do what they were meant to do (see here for the most recent update). Although their September loss was a highly unpleasant experience, longer-term it proved to be a relatively minor setback.

 (B) The problem of underperforming strategies

The problem with quantitative trading strategies is that we only know how they have performed in the past. We have no way of knowing how any strategy might do in the future: today’s top performer could fall short while an underperforming strategy might excel. For this reason, I am always reluctant to favor past winners and eliminate underperformers. As the above chart shows, our underperforming strategy marked “B” caught up nicely and ended up being our top performer from December through March.

Isn’t there an easier way?

We’d all like to find some gimmick giving us instant trading gratification and predictable profits without much risk of losses. This desire is met with a flood of offers promising very large returns on investment, risk-free trading, 80% accurate forecasts, etc. Such claims will usually prove false. We know that most speculators lose money. In some jurisdictions, retail brokers are obliged to disclose the percentage of their customers who lose money. Here are a few such disclosures:

• ETX Capital: 75.6% lose money
• IG Group: 74% lose money
• Ava Trade: 71% lose money
• Plus 500: 76% lose money

Quant contests: 79% lose money

One company’s experience provides a valuable empirical case: in December 2006, world’s most popular trading platform provider MetaQuotes began staging an annual Automated Trading Competition. Through 2012, the $100,000 prize attracted a total of 2,726 participants – all highly motivated to win the prize and sophisticated enough to formulate and test quantitative strategies they hoped would win the competition. However, only 567 (21%) of them finished their competitions in the black while all others – 79% of them, lost money.^[1] This data speaks to the unrealistic claims of many providers of quantitative systems: they all look great on paper but usually prove less useful in reality.

How is trend following different?

Trends are the product of human psychology and the collective action of market participants. Our psychology is the one constant in all markets that’s not likely to change any time soon, which is why we can observe price trends in any market throughout history.

Here’s how the trending dynamic emerges: in making investments, our rational goal is to obtain the best possible return with the least risk necessary. If we buy some asset, we want to receive a stream of rents or dividends and have the opportunity to sell that security for a higher price than what we paid. Since those outcomes depend on other market participants, we are obliged to reflect on what they might do. Thus, if the price of some asset is rising, we may conclude that it is attractive to other investors, and that the rising demand would push prices even higher.

Informed by the actions of others, we might accept inflated asset prices and proceed with the investment anyway. In one form or another, this dynamic is at work in every market where people speculate and transact – it is simply a matter of human nature, which is why we often see the prices of things like Oil, Gold, Tesla, Amazon, Apple or Bitcoin advance far higher and for far longer than anyone might have predicted in advance. In investment speculation, the simple axiom is, “trend is your friend.”