Sample Size
Also called: sample size, number of observations, n, how many trades.
Formula
The plausible range around a measured rate narrows roughly in proportion to the square root of the sample size. Quadrupling the observations halves the width.
How it is read
Sample size is what makes a percentage readable. A hit rate on a dozen occurrences carries a range of roughly plus or minus thirty percentage points. That is wider than any edge worth having. So the honest statement about it is what it cannot rule out. It is not what the number shows. The range narrows with the square root. So going from ten to a hundred observations is a large improvement. Going from a hundred to a hundred and ten is not.
What it does not tell you
Counting observations is not the same as having independent ones. Overlapping windows inflate the count without adding information. Several positions in the same move do the same. Many rules tested on one history do the same. A large sample also does not rescue a biased measurement. It only makes a wrong number more precise. No sample size can fix one other thing. A result taken from the data that suggested the rule is not a test of that rule.
Related concepts
Sample Size: What Twelve Occurrences Actually Prove
A hit rate measured on a dozen occurrences carries an uncertainty wider than any edge worth having.
Open the lesson