Article Surfing Archive

When should you override a trading strategy?

Let's use the following example for trading the e-mini S&P:

Your system establishes a long position at 1190.00 and the profit target order was placed at 1192.25 ($112.50 profit per contract). Prices moved up to 1192.00 and reversed. One hour later the system tries to reverse at 1191.00. Again prices moved up to 1190.75 and reversed. Two times the system missed the profit target by one tick.

Should you change the strategy?

or

Should you manually override the strategy when something like this happens?

Doing any of this is like opening Pandorra's Box: Let's say you start lowering your profit goal by one tick. Of course you will be instantly rewarded, because the number of winners would increase. Next week you might experience the following situation: Your stop is hit and you are taken out of the trade, but then the market turns and takes off and you are missing a nice winner. What now? You start moving your stop a little bit further away and again you are instantly rewarded: The number of losers decrease.

One week later you experience a similar situation and you continue "finetuning" your system by slightly moving down your profit goal and minimally increasing your stop loss. And very soon the winning system that you once had turns into a losing one, because your losses are much bigger than your profits.

I have seen it many times: A trader backtests his system over 700 or maybe even 1,000 trades and then "finetunes" it after the first 5 trades. This doesn't make sense: If you have a sound logic why your system should work then it won't need "finetuning" after 5 trades.

You should evaluate your system periodically, but instead of curve-fitting your system's parameters you should ask yourself: "Does the logic of the system still apply?"

If you have a trend-following system and markets are trading sideways, then "optimizing" the system parameters won't help: It's the wrong market condition and the market is just not right for your system.

Exercise discipline: If your system worked well over 1,000 trades, and you have a sound logic and didn't curve-fit the system, then you should not override the system or "finetune" it after only a few trades.