Yesterday’s message was a thought exercise on the best ways to curve fit, hypothetically assuming that was your goal.
Now we return to the real world, where you’re trying to minimize it.
The best way to steer clear of curve fitting is usually quite easy: add lots of trades to your backtest.
The more trades you have, the less chance you’ll curve fit.
It’s not impossible, but with many trades and a healthy dose of common sense, you can worry a lot less.
How do you add more trades to your backtest?
If your backtest has a small number of symbols, add more. (This is one of the main reasons I prefer equities to futures and day trading to swing trading – there are simply way more profitable trades to be had!)
If your backtest covers a certain window of time, explore increasing that window to a longer period that covers more market types.
If you’ve filtered your signal so it doesn’t occur frequently, loosen the filters so it happens more often.
Remember, you can’t eliminate the risk of curve-fitting, but you have a lot of control over how you set up your backtest to minimize the chance of it.
I’ll focus soon on the exact steps you can use to set up a backtest to give you the best chance of finding a profitable edge from it.
(If you’re like most traders, you’re probably backtesting wrong!)
-Dave
P.S. Speaking of backtesting, there are still a few days left to beta test a new product I’m working on that helps you optimize your backtests (without curve fitting). If you’re interested, hit reply and let me know which backtesting software you use.
Here’s what one trader said about it:
“As someone relatively new to trading without a strong quant background, I initially struggled to build and refine a profitable strategy. That changed when I started using Dave’s optimizer. It has completely transformed my trading experience. It has made trading more accessible, significantly less time-consuming, and much more enjoyable—all while improving my results.” – Brandon C.
