What do you do if the trading system that you developed has a really impressive equity curve the past year, but performs terribly on a ~4 years timeframe?
Do you just scrap it and start working on a new strategy? What would be your approach?
It depends on how you have developed it.
If you coded it while only looked at how it were performing on the last year, it’s certainly over-fit on the last year data. If you want to be sure, the only “easy” way is to test the strategy in the next weeks or months and see how it goes, with these new market datas.
Thanks for your response Nicolas.
I came up with an idea, and then I added the indicators which the idea was based on onto the chart. After giving it a quick look, It seemed to perform pretty good on the recent data, so I coded it. And this is when I realized that the strategy was splendid the past year, but before that it wasn’t pretty at all.
Did you change the indicators default settings? What is the indicator?
I did not change any default values. The two indicators were MACD (12,26,9) and Stochastic(14,3,5).
That’s already a good point, but I need more informations about how many trades, markets, spread, etc… Or the code itself of course 🙂
I will definetly give you an update on this. I Was going to upload the code right now, but for some reasons my most recent strategies were gone in PRT when I opened it today. Will write the code again and post it here whenever i have time.
Ok that’s nice. From time to time, I usually make a CTRL+S to save the platform and all codes in progress, that way i’m sure everything are safe.