When you’re creating trading strategies, most traders eventually consider trying to improve the system’s performance by adding an overall market rule.
For example, you have a system that trades stocks gapping up long and stocks gapping down short.
Instead of taking all the trades in the system, you consider only taking the long trades when the overall market is positive or only taking the short trades when the overall market is negative as determined by some metric.
(In Trade-Ideas, there are several of these overall market filters available. For example, S&P Change in 30 Minutes, S&P Change Today, NASDAQ Change Today, etc. See those links for the full details about how they work.)
While this might seem like a perfectly reasonable approach, I think it’s a mistake to use these filters that measure the overall market in your trading strategy.
Here’s why.
When you design a trading system, you identify an event as the impetus for your trading signal.
This event needs to be significant, strong, and unusual – otherwise, it’s just random noise.
For most trading systems, the overall market behavior at the time of your trade should be irrelevant.
When you consider applying a rule based on the overall market, on some level you are admitting that your system’s trading signal isn’t as strong as it should be and that it “needs help” from a largely unrelated factor to function sufficiently.
When you think about applying an overall market filter, it’s usually a sign that you need to find more significance in the trading signal itself.
But even beyond that, there’s another important reason I avoid overall market filters.
Let’s say you can apply an overall market filter that significantly improves the performance of your trading system.
By definition, this improvement will come from increasing the profit per trade and reducing the number of trades in your system.
In other words, you’ll remove some relatively poor trades from your system when the overall market isn’t favorable.
Unlike a normal trading rule, you’ll end up with large windows of time where your system isn’t making trades at all.
For example, let’s say your rule is to skip short trades in your system when the SPY is above its 50-day moving average.
One look at a daily chart of the SPY and you’ll see there will be weeks or even longer when your system makes no trades, just because of this one rule you applied.
This is a terrible tradeoff to me.
I want my trading systems to trade frequently and regularly, without large arbitrary windows of time when they’re effectively disabled.
Instead of considering a market filter, use your effort to identify more power in your trading signal and your system will be more robust.
