What a backtest cannot tell you

A backtest asks a narrow question: if these rules had been followed on these past prices, what would the arithmetic have been. It is useful. It is not a tour of the future.

It cannot feel the spread you will actually be quoted, unless you put a realistic spread in. It cannot feel the hesitation that skips the uncomfortable trades and takes the pretty ones. It cannot feel a news spike that fills you worse than the bar's low. If those are missing, the curve is a ceiling, not a forecast.

Watch for the rules that were added after the author saw the chart. A filter that exists to hide one ugly month is not a filter. It is a memory. Honest testing decides the rules first, then looks.

Use a backtest to reject ideas that do not survive their own costs. Do not use it as a promise that the next month will rhyme. Forward practice, with the spread written down, is the next piece of evidence. It is slower, which is why people skip it.