In my last message, I talked about the worst thing that can happen with curve-fitting.
Now I ask the question: Is your strategy curve-fitted?
Yes – to some degree.
Because there’s a little bit of curve-fitting in every strategy.
Curve-fitting is not a binary metric – you can’t use a curve-fitting ruler and measure to see if your strategy meets some standardized definition.
Think of curve-fitting as having a whole range of degrees rather than some black or white, true/false measurement.
Consider all the rules that make up your strategy. For example, relative volume needs to be greater than X for a signal to meet a strategy’s criteria.
Even the most obvious, predictive rule that’s highly correlated with profit has a little bit of curve-fitting.
That’s because strategies, by definition, are inherently trying to predict the future – which, as it turns out, is hard. Things are always changing, and there’s no guarantee a pattern you see will continue forever.
If you accept there’s a little curve-fitting in every trading rule you apply, it shouldn’t cripple you with fear. The fact there’s a non-zero chance of injury or death when you cross the street doesn’t keep you locked up in your house forever so you can avoid the miniscule risk.
Tomorrow, I’ll discuss things you can do when designing your strategy to minimize curve-fitting so you end up with a robust strategy with staying power.
-Dave
P.S. If you’re just getting started trading and you’re not getting the traction you know you could be, take a look at the Get Systematic Roadmap. Most beginning traders make the same few mistakes over and over without ever realizing it. This custom blueprint gives you a solid foundation so you can create and improve your own trading strategies with real edge.
