Backtesting properly

A backtest runs your strategy over historical data to see how it would have performed. It is indispensable — and it is where most strategies die honestly, or worse, appear to live and then die with real money. Doing it properly is the single most valuable skill in this track.

The fatal errors

• Look-ahead bias — using information that was not available at the time of the trade. Trading on a closing price you compute using that same day's close, or using restated fundamentals, silently injects the future. It is the most common and most lethal mistake.
• Survivorship bias — testing only on companies that exist today ignores all the ones that went bankrupt or were delisted. Your universe is secretly filtered to survivors, flattering every result. A proper test uses a point-in-time universe including the dead.
• Ignoring costs — every trade pays a spread, commission and slippage (the gap between expected and actual fill). Strategies that look brilliant gross are often unprofitable net, especially high-turnover ones. Always model realistic costs.
• Unrealistic fills — assuming you bought at the exact low or filled a huge order at the quoted price. Your own trading would have moved the market.

What a trustworthy backtest reports

Not just total return, but risk-adjusted return (Sharpe), the worst peak-to-trough loss (maximum drawdown), turnover, and how performance held up across different periods rather than one lucky stretch.

A backtest is a hypothesis, not a proof. Its honest purpose is to reject bad ideas cheaply — not to convince you a strategy will work. The next lessons are about not lying to yourself with it.