What is overfitting when optimising an EA?

Chris Taylor Asked 1 day ago 1.3K views

I optimised my EA and the back test looks perfect, but it loses on a demo account. Why?

๐Ÿ‘1๐Ÿ’ก1๐Ÿ˜ฎ1๐Ÿ™‚React
8
Report

Log in to report content.

For education only. This is not financial advice. Trading carries a high risk of loss.

Log in to reply.

1 Answer

Best answer

Overfitting (curve fitting) happens when you tune the settings until the EA looks perfect on past data. It then fits the noise of that period instead of a real edge, and it fails on new data.

How to reduce it:

  • Use fewer parameters.
  • Test on out-of-sample data and use walk-forward testing.
  • Prefer settings in a stable zone, where nearby values also give good results, over one sharp peak.
  • Test on different pairs and periods.
  • Include realistic spread and slippage.
Marcus LindqvistAlgo Trading and MQL5 Answered 1 day ago
14
๐Ÿ‘1๐Ÿ’ก1๐Ÿ˜ฎ1๐Ÿ™‚React
💬 Reply

Log in to reply.

Your Answer

Log in or sign up to answer this question.