Prop Win-Rate Samples: Eight Wins Do Not Establish a Pass Probability
A sample win fraction needs its denominator and model assumptions. Eight wins from ten observations and eighty from one hundred both give 80%, but they contain different amounts of evidence. Neither statistic directly supplies the chance of passing an evaluation.
Explore this site’s existing research resource
Define the outcome and sampling process
Record what counts as a win, the observation cutoff, cost treatment and every included outcome. Open, missing or untriggered cases require their own classifications. A selected set of favourable screenshots is not a complete sample of independent trials.
The Wilson interval is one method discussed in NIST’s proportion guidance. Applying a binomial interval assumes a suitable binary-outcome model; actual trading outcomes can be dependent and selected. The interval does not repair a biased dataset or turn a research sample into account evidence.
Compare two fictional samples
Using z=1.96 for an illustrative 95% Wilson interval,8 wins from 10 gives approximately 49.02% to 94.33%. With 80 wins from 100, the interval is approximately 71.12% to 86.66%. Both sample fractions are 80%, while the second interval is narrower under the stated model.
These are intervals from hypothetical counts, not measured trading probabilities. The method’s confidence interpretation concerns repeated sampling under its assumptions. It does not state that a particular account has a95% chance of meeting a profit target or loss rule.
| Sample | Win fraction | Approximate interval |
|---|---|---|
| 8 wins /10 | 80% | 49.02%–94.33% |
| 80 wins /100 | 80% | 71.12%–86.66% |
| Actual sample validity | Not supplied | Dependence/selection need review |
| Evaluation pass probability | Not calculated | Different account-level question |
Review the full programme separately
The existing small-sample resource remains the companion on this site. An evaluation depends on payoff sizes, costs, account path, activity and the current product conditions. Even a well-estimated trade win fraction would not contain all of those facts.
The optional sponsored HashHedge link offers current programme information. No customer win rate or pass probability is asserted, and no trading outcome is forecast. The example makes uncertainty visible rather than using a small sample as a registration promise.
Sources and rule dates
- NIST: confidence intervals for proportions
Checked 2026-10-23. Selected published description only. Worked numbers and event records are fictional; no actual provider account, search result, live price, eligibility or payment is established. - HashHedge: challenge types
Checked 2026-10-23. Selected published description only. Worked numbers and event records are fictional; no actual provider account, search result, live price, eligibility or payment is established. - HashHedge: terms and conditions
Checked 2026-10-23. Selected published description only. Worked numbers and event records are fictional; no actual provider account, search result, live price, eligibility or payment is established.