Look-ahead bias, curve fitting, survivorship bias, ignored costs and the other errors that inflate backtest results, and how to avoid each one.
Most strategies that fail live looked great in a backtest. These are the usual reasons.
Using information that would not have been available at the time, such as the day's closing price to decide a morning entry. Fix: step through charts bar by bar and only use data to the left of the current bar.
Tweaking parameters until past results look perfect. The more rules and parameters, the more likely the strategy fits noise. Fix: keep rules simple and test on data you did not use to design them.
Twenty trades can show almost anything. Fix: aim for at least 100 trades across different market conditions.
Commissions, spreads and slippage can turn a small edge negative, especially for short-term strategies. Fix: subtract realistic costs from every trade.
Testing stocks only on companies that exist today ignores the ones that went bankrupt or were delisted. Fix: use data that includes delisted symbols when testing stock strategies.
Skipping signals you “wouldn't have taken”. Fix: take every signal the rules produce, or write the filter into the rules.
A strategy with great returns and a 50% drawdown is one most people will abandon. Fix: check maximum drawdown and the longest losing streak, and ask whether you could sit through them. The drawdown calculator shows what it takes to recover.
Forward test with small size before trusting the numbers. See how to forward test a strategy and our full guide to manual backtesting.

Sync your trades, set daily loss and drawdown limits and see exactly where your edge comes from. Free for one account.
Educational content only, not financial advice. We may earn a commission from partner links. See our affiliate disclosure.