Sample Size in Backtesting

Why a small number of historical trades can make a backtest look cleaner than the evidence really is.

Small samples can be fragile, even when the chart and table look tidy.

Small samples are noisy

If a backtest only produces a few trades, one unusual winner or loser can dominate the result. That makes the summary less reliable as evidence.

A sample-size warning is not a technical footnote. It is a signal that the result needs extra context.

A larger sample can help, but only if the rule is still coherent and the data remains relevant to the question being tested.

Hundreds of low-quality trades can still describe a poor idea. Sample size works best alongside clear rule review.

ZadVest treats sample size as part of evidence quality. A single backtest does not become proof just because the result table is tidy.

Sample size helps distinguish a research artifact from a conclusion.