Systematic trading, systematic risk control

From the Issue of Managed Futures Today
Systematic trading is simply a method of creating rules that can execute a trade idea and manage risk with some degree of consistency and predictability over time.

Although the majority of professional futures fund managers trade systematically — at least to a significant degree — there is still a great deal of confusion about what systematic trading really is. Often associated with complex computerized trading approaches, systematic trading is just a way of methodically defining trade goals and risk controls, from the portfolio level to the individual trade level.

While many investors might think money managers wake up each morning and think of novel ways to make their clients money based on the latest financial news, no manager worth his salt proceeds without a detailed game plan. Systematic traders simply formalize the process of taking profits out of the market to a greater degree.


Achieving investment goals
Ultimately, all money managers have the same basic goal — capturing available profits while, hopefully, minimizing risk. Markets, time frames, and trading styles can vary.

Consider two money managers, both attempting to capture medium- to longer-term trends — say, one to six months. The first manager might trade on a discretionary basis, assessing markets individually to see if their supply and demand situations or other fundamental factors indicate a likely trend. The second manager might use a collection of systematic rules to capture the trends that have historically proven to develop with a certain degree of regularity over time. Same goal, different approaches. But the second money manager, by virtue of using a trading system to objectively capture trends when they develop, might actually have the more understandable and consistent trading program.

Let’s take the case of a hypothetical commodity trading advisor (CTA) trading a simple trend-following breakout system on a 20-market futures portfolio. For the sake of argument, let’s assume the basis of the system goes long when a market pushes to the highest price it has made in 62 days (approximately three months) and goes short when a market falls below the lowest price of the past 62 days. The logic behind this common “breakout” approach is that if a market moves in a certain direction by a certain amount, it implies the strength to continue in that direction — that is, form a trend.

That’s one part of the puzzle — entering the market. But any money manager also has to know when to exit (either with a profit or a loss, if the market doesn’t move as expected), how big of a position to put on, and how one trade fits in with all the rest in a multi-market portfolio. Systematic trading, in fact, lends itself to risk control precisely because it allows money managers to define profit targets, loss points, trade size, and system shutdown points objectively and in advance of entering each trade.

What makes it possible to proactively determine such levels is extensive “back-testing,” which is the process of applying trading system rules to historical price data to simulate how the system will perform. By testing the system multiple times and over different periods of time, the money manager can determine the reliability of the trade rules — if they are capturing repetitive patterns in a market — and the odds they will succeed in the future. To be able to test rules accurately, those rules have to be objective and consistent; they cannot vary from one trade to the next.

This objectivity allows systematic traders to test rules governing virtually every aspect of their trading programs, from how much they should risk on each trade to how much the portfolio can lose overall before trading should be halted.

In the hypothetical breakout system, let’s say we add a rule to exit long trades when the market falls below the lowest price of the past 11 days, and exit short trades when the market rises above the highest price of the past 11 days. The logic here is that if the market has been trending up or down, if it moves in the opposite direction by a certain amount — in this case, a new 11-day high or low — it is a sign the trend is probably losing its momentum. Like the original entry rule, the money manager would have determined this system “parameter” through extensive testing. (In reality, these rules are merely illustrative. Basic tests on a handful of markets showed these rules were profitable in some markets and unprofitable in others.)

Figure 1 shows the buys and sells resulting from these trade rules in the crude oil futures market (ticker symbol: CL). The system did what it is supposed to do: capture trends when they occur. There were five trades during this period, but the only two that really matter are the long trade that captured most of the uptrend from February to July 2008 and the short trade in September 2008 that caught the majority of the huge downtrend that continued into spring 2009. The other trades were small winners or losers, as are most trades generated by this type of system. All markets wander aimlessly more than they trend. The rules ensure that if and when future trends unfold, the system will capture those as well; there will be no need for guessing.

Figure 1: Objective Rules

Layers of risk management

A typical systematic trading approach might include some or all of the following tools for managing risk.

1. Portfolio level:

Portfolio, system, and time frame diversification

• Trading different markets.
• Trading different systems.
• Trading on different time frames.
• Halting trading if and when account equity experiences an x-percent drawdown.

2. System level:

• Allocating a fixed percentage of equity for each position to equalize per-trade risk.
• Capping the number of positions/risk for different sectors and/or all markets.
• Risk balancing among portfolio components (ensuring each trade represents the same risk to the portfolio.

3. Trade level:

           • Predefined stop-loss and
             profit-target levels.

More risk control
In the real world, trading systems typically include more-vital rules that determine the size of each trade — that is, how many contracts to trade — and other factors that control risk. For example, one systematic way many CTAs manage risk is to size trades consistently across all markets and market conditions. Futures contracts vary enormously in size and volatility; holding a one-contract position in 10-year T-note futures (TY) represents a much different level of risk than a one-contract position in coffee futures (KC).

Because the manager wants to make sure every trade has the same potential impact on the portfolio’s performance, he will determine how many contracts of market A it takes to equal the same risk as trading market B, and so on. Also, to stabilize portfolio risk over time, CTAs will often further adjust trade size depending on how volatile a market is. In choppy, wide-swinging markets, the trade size will decrease to limit portfolio exposure, and the number of contracts will increase when market conditions are quiet. As was the case with the basic entry and exit rules, the rules determining these changes are tested and defined in advance. It would, actually, be detrimental to performance to make such important decisions on an ad hoc basis.


Capping losses
Another systematic risk management tool is determining the maximum amount of account equity that can be risked on an individual trade, in a particular market sector, and in the portfolio as a whole. CTAs typically limit per-trade risk to a small percentage of total equity (for example, 1 to 2%), and often set comparable limits at the sector and portfolio level. Once these levels are reached, no additional trades can be opened.

Similarly, many futures fund managers institute fixed “shutdown” points at which they will halt trading. For example, if a trading program drops 20% or more from an equity high, the manager may temporarily halt trading to assess whether the system has stopped performing or market conditions have rendered the system ineffective. Again, historical testing will provide logical guidelines for what constitutes a reasonable drawdown vs. a breakdown in system performance.

“Layers of risk management” lists some representative rules a systematic trader might incorporate to control risk at different levels of a trading program.


Diversification on many levels
Although some trading programs focus on specific markets, trading styles, or systems, many diversified CTAs consciously managed risk by diversifying across markets, time frames, and strategies. Systematic trading ultimately simplifies this process. Also, managed futures investors can diversify by allocating funds among different CTAs with different areas of specialization. (There are also futures funds of funds that provide this service.)


Theory vs. practice
Although the process of researching and testing a trading system — especially on a portfolio basis — can be quite complex, that doesn’t mean the underlying logic triggering trades is. Systematic traders simply employ technology to be able to mechanically and objectively capture the profit opportunities their research finds in the markets. Those opportunities can be as straightforward as trading long-term trends.

In reality, there are relatively few purely “systematic” or “discretionary” traders, although there are more of the former than the latter. Discretionary traders base their decisions on rules and execute certain tasks systematically, while most systematic traders leave at least some room for discretion in their trading programs.