Sample size is the difference between evidence and a story. A 70% win rate over 20 trades is meaningless — a single hot streak produces it by chance. The same win rate over 500 trades is a genuine finding.
The rough rule
Aim for at least 100 trades before you take any metric seriously, and 300+ before you trust it enough to size up. The rarer your setup, the more history you need to accumulate those trades.
Why small samples fool everyone
With few trades, one oversized winner can create a fantastic profit factor and expectancy that vanish on the next batch. Randomness looks exactly like skill until the sample is large enough to separate them.
How to get enough trades
Test over many years, and across several markets and timeframes if your logic is general. If a setup only produces 30 trades in two decades, its rules are probably too strict to trade or to trust.
See the sample size instantly
Bulk Quant runs your rules over 20+ years and shows the trade count alongside every metric, warning you when there are too few trades to trust. Run it free.