ALGOINDICATEUR
Login Create an account
๐Ÿค– Algo AI Trader ๐Ÿง  AI Assistant ๐Ÿ“ก Algo Signals VIP ๐Ÿ’น Trading

Platform

๐Ÿ“ˆ Markets โญ Watchlist ๐ŸŒ Market movers

Tools

๐Ÿงฎ Trader calculators ๐Ÿ”„ Converter ๐Ÿ“… Economic calendar ๐Ÿ“– Trading glossary ๐ŸŽ“ Trader academy ๐Ÿ““ Trade journal ๐Ÿงฐ Advanced calculators ๐Ÿ“Š Chart patterns โœจ App features

Resources

๐Ÿ“ฐ News โ“ FAQs โ„น๏ธ About ๐Ÿ“จ Telegram signals
Create an account Login Become VIP
Trader academy

Testing a strategy without fooling yourself

Backtest, sample, fees and chance: how to avoid false hopes.

Lesson 8 / 96 ยท 8 min read ยท Beginner

A backtest shows what a strategy would have done in the past. The trap: with enough attempts, chance ends up looking like a good strategy.

The four traps

  • Over-fitting: tuning until the history looks good.
  • Multiple tests: out of 500 strategies, a few succeed by pure luck.
  • Forgetting fees and the spread.
  • Too few trades: twenty trades prove nothing.
ResultThe best by chance
Among 500 randomly drawn strategies, a few show an excellent result by pure luck.

How to protect yourself

  • Split the history: learn on the older part, check on the recent part the search has never seen.
  • Keep a last โ€œblindโ€ slice, which you look at only once.
  • Compare with random entries: if chance does as well, it is not an edge.
  • Test on other markets, then on a demo account.
LearningValidation?BlindOlderRecent
You learn on the older part, check on the recent part, and keep a last โ€œblindโ€ slice looked at only once.

This is exactly what the strategy search in Algo Signals does: fees, blind test and chance test.

๐Ÿ“ Lesson test

Log in to take this lessonโ€™s test. Once you have passed all the tests, you will receive a certificate in your name.

Log in

These lessons are for understanding and are not investment advice. No method guarantees winning.