The same can be applied to a stock market. I am a big fan of looking into historical data, and I was using WealthLab for quite a while.

One of the funniest things is when you find "strategy" that performs best over one year by making from 50 to 100 deals. But don't get fooled, it's just a random parameters, and when applied to the next year or years, you won't get these results, of course.

So you're getting reliable results only when you can reproduce your success (no matter what it is) consistently.