The one inviolable rule of trading is that you can either have a system that is high probability and low profit or a system that is low probability and high profit, but you can never have both. That doesn’t stop novice traders from searching endlessly for the setup that risks a dollar and makes ten while being profitable 70% of the time but like the elusive El Dorado it does not exist.

I bring this up because when you are considering algo trading you should know this rule cold. If your system looks too good to be true in backtest, I guarantee you it is. Forty-five degree equity curves exist only in trader’s imaginations. In real life, professional trading looks like real life, two steps forward, one step back if your are good. More often than not trading can be two steps backward for any one step forward.

Just like baseball players every system will eventually go cold. Momentum setups will get destroyed by choppy price action and mean reversion trades will be stopped out repeatedly in strong trending markets. That’s why when you are creating a system it is much more important to understand the “why” rather than the “how” behind your ideas. This is the reason that so many engineers and mathematicians fail as traders. They become enamored with the data without understanding the underlying mechanics that drive markets. Price action is only a reflection of market reality, not the actual reality itself. Patterns change as reality changes and that’s why all algos fail in the end.

Which brings me to my final point. When you are thinking seriously about risk control in system trading you have to have a portfolio approach. The best way to minimize risk is to trade several non-correlated systems at one time. The idea is that you try to make more money on your profitable system that you lose on your unprofitable one.

How do you know when a strategy has gone bad? Marcel Dion, the quant from JP Morgan offers an interesting solution. In the Active Trade interview he states, “We track a system's equity curve and make changes when it moves two standard deviations above the mean." Note that Dion will stop trading the system not only when it gets very cold, but also when it overheats. That's a good thing to keep in mind. In trading when things are going gangbusters and you fell invincible, that's not a sign of your brilliance, but a warning that the market is about to extract its punishment on your account.

Boris Schlossberg serves as director of currency research at GFT, and runs bktraderfx.com.