Why Mechanical Trading Systems Fail
Training yourself to be a fully systematic trader is hard. The virtue of a well-designed trading system is that real-time results come in as the backtest leads you to expect. The challenge of a mechanical trading system is . . . you.
You have to trade exactly as the system dictates in order to duplicate test results. For some, following every single system dictate is impossible. Here are some of the common pitfalls of mechanical trading systems:
Fooling around with new ideas: Mechanical trading systems often fail because the trader can’t resist fiddling with system components — either indicators or rules. Most traders’ systems are never really finished. They evolve while the trader experiments with new ideas. The problem with new techniques is that people are impatient and try to fit the new idea into an existing system without fully backtesting it — or sometimes without backtesting it at all.
Backtesting until you’re blue in the face: To find the perfect parameter for your indicator on your securities, you can spend countless hours backtesting. No sooner do you discover the ideal parameters than market volatility shifts, and the parameter is no longer optimum.
A lot of indicator testing is just spinning your wheels. Indicator signal accuracy isn’t 100-percent reliable to begin with, and fiddling with indicators never cures the accuracy problem. Before you spend a zillion hours adding or perfecting indicators, remember that your goal isn’t to have the perfect indicator; your goal is to make money.
Not knowing your time frame: Technical analysis contains rules that are valid in the context of their own time frame but work a whole lot less well in a different time frame.
Practicing self sabotage: Although a mechanical system imparts confidence in the eventual profit-and-loss profile over some period of time, it has the drawback of occasionally being wrong on any single trade. Sometimes you can see the wrong trade coming, which makes you want to override the signal. To override technical signals is called discretion. Discretion is an innocent-sounding word, but in fact, it’s dynamite. To exercise discretion means to abandon your hard-earned, high-probability systematic trading signals in favor of personal judgment.
Because you can’t backtest judgment, the only way to evaluate discretionary overrides is to keep a diary and write down every override you want to do. Every so often, go back and do an honest accounting of your judgment. A trading diary has many benefits:
You get ideas about additions to your system to overcome a shortcoming. The diary becomes a wish list. Then while perusing the technical literature, you can see a gem when you come across it — it’s the solution to an issue on your wish list.
You may discover that your eye was detecting patterns that math-based indicators don’t catch. If you had a feeling that you should stop out a position but your indicators didn’t agree, and in retrospect you can see a pattern that was correct, you may have a hidden talent for patterns.
You discover personal characteristics you didn’t know about yourself (and may or may not like). A common finding is that you saw a continuing trend because you wanted to see it and willfully ignored reversal warnings from other indicators that were present on the chart.