In plain English
What Backtesting means
A backtest converts a strategy into rules that could have been followed at the time, then runs those rules through past prices and other available data. It can show how often the rules traded, their hypothetical returns, and their drawdowns. A credible test specifies items such as the instrument, timeframe, entry and exit rules, position size, trading costs, and whether decisions used only information available at that historical moment. It cannot establish that the strategy will work in future conditions.
Why it matters
Backtesting helps reveal whether a strategy’s logic is internally consistent and whether a claimed result is sensitive to costs or market conditions. It is also easy to overstate: selecting favorable dates, repeatedly changing rules after seeing results, or assuming unrealistic fills can make a weak strategy appear strong. Hypothetical performance should therefore be kept separate from live, independently verifiable results.
Example
A simplified EUR/USD strategy buys at the close whenever a 20-day moving average rises above a 50-day moving average and exits when it falls below. A backtest over 10 years might calculate 48 hypothetical trades. If it assumes every trade fills exactly at the closing price and charges no spread or slippage, its reported return is likely more favorable than a result using realistic execution costs.
Quick answers
Common questions
Is backtesting the same as paper trading?+
No. Backtesting applies rules to completed historical data. Paper trading applies them in a simulated account as markets unfold. Paper trading can expose practical issues such as delays, missed alerts, and order handling that a historical simulation may not capture.
Should a backtest include spreads and commissions?+
Yes, where they apply. Costs, spreads, financing, slippage, and realistic assumptions about order fills can materially change a strategy’s hypothetical result. The appropriate inputs depend on the instrument, broker arrangement, account type, and trading frequency.
Sources