Backtesting

What backtesting actually is

Backtesting means taking a set of trading rules and checking them against price that has already happened. You apply the rules to history, write down every trade they would have produced, and add up the result.

The purpose is not to predict anything. It is to find out whether the idea has ever worked, and roughly how it behaves, before you pay to find out live.

Doing it by hand or with code

If your rules can be written down exactly, software can test years of data in seconds. If part of your decision is judgement, and for most discretionary traders it is, then the honest way is to replay the chart bar by bar and make each call yourself.

Manual replay is slower but it has one advantage that matters. You cannot see the future, so you cannot accidentally cheat.

The three things that ruin a backtest

Too few trades, so the result is luck. Tuning the rules until they fit the data, so the result is a description of the past rather than an edge. And leaving out costs, so a losing system looks like a winning one.

Each of those has a guide below, because each is common enough to be worth its own page.