Who this is for — Anyone who wants to know when a metric is informative and when it is just statistical noise dressed up as a result.
Sample size is the number of trades (or signals) used to estimate method performance. The smaller the sample, the more win rate, payoff, and expectancy swing by chance.
In plain terms — With few trades you can tell yourself any story: you need enough data to separate skill from luck.
Bronze prerequisite — Before this lesson: trading-journal, trade-result, trading-mistake. See bronze-path.
Why it matters in validation
An adequate sample reduces premature conclusions:
- limits the effect of a few extreme trades;
- stabilises the estimate of expectancy;
- shows drawdown depth and duration better;
- helps compare periods and different setups.
There is no one-size-fits-all sample size, but "definitive" judgements on 10–20 trades are almost always fragile.
Example — After 18 trades a method shows 72% win rate. At 140 trades the same method drops to 49%, with better payoff but a very different reality: the first figure was incomplete.
Practical use on the Silver path
- Define a minimum observation threshold before changing rules.
- Separate results by market regime and instrument.
- Compare homogeneous samples in terms of risk and costs.
- Always integrate with backtest and out-of-sample tests.
Card
- What it is: amount of data on which you judge method quality.
- When to use it: in metric reviews, setup comparisons, and change decisions.
- Typical mistake: changing strategy after a few lucky or unlucky trades.
Silver path — Module: Validation. Part of silver-path.