How many chart instances before a setup earns a verdict
Sample size is not a magic number — but twenty cherry-picked winners still is not a study. Here is how we teach traders to decide when to keep marking.
Traders often ask for a single number: thirty, fifty, one hundred. In our intensives we refuse the shortcut. A setup that appears twice a month on the ASX200 daily chart cannot be judged on the same calendar as a five-minute FX scalp that prints candidates every session.
Start with frequency
Count how often your written rules fire in a typical month. If the answer is two, you need years of history before the sample stabilises. If the answer is twenty, a few busy months may suffice — provided you include dull weeks, not only the exciting ones.
Separate clear from ambiguous
Mark ambiguous bars in a different column. A setup with thirty crystal-clear instances and forty “maybe” bars is telling you the language is soft. Fix the wording before you chase a larger count.
Stop when new bars stop teaching
When the next ten instances look like the previous forty in both context and outcome mix, you have diminishing returns. That is a better stopping rule than a round number borrowed from a textbook.
Bring your current count to a clinic if you are unsure whether to keep marking or rewrite the rule sheet.