i could write a book on the fallacies of back testing they are so numerous. in a nutshell back testing platforms will give you results than are not realistic if you use them wrong. most all people misuse back testing without knowing it, the number one problem is using limit orders. when using limit orders you will most often see great systems results that can not be replicated in real trading. back testing is great i use it all the time but i do so with the knowledge of all the pitfalls. m
Hello MarkBrown, I agree, many times I thought I hit the lottery in back test, only to find out I was still poor in forward testing.
I would tell my younger self to look at trades as a sequence to scale into vs one-shot decisions.. Eg instead of buy 600 at $23, buy 200 @$23, 200 @$23.4, 200@$23.8. type of thinking... so yeah position management is key And wide range clean charts. And hookers n blow for the up weeks (jk)
Price is not random, but time is. That invalidates all back testing (curve fitting) on time based indicators.
How is time random? You mention indicators. Something I don’t really use. You can back-test / back-check historical prices manually. No indicators needed. If you have a statistical model (which I have), you can use that model to make your tests or studies. Curve-fitting or optimizing classical indicators seems like a waste of time. Not my idea of back-testing / checking.
If I use back testing it's to prove that whatever bright idea I just had doesn't work. If I can't prove that it doesn't work then I'll forward test the idea. Have a theory; backtest it over a period of time. Now back test it over a different period of time. Now back test it on a different market or different asset.
Understanding how FOMC, CPI, interest rates, employment data, and GDP expectations and actual numbers move the markets is something I wish I paid more attention to. I can struggle all week, and make it back and more when I'm on the right side of economic data.