Estimate a Wilson confidence interval for an observed pipeline success rate rate. Enter the number of successful outcomes, the total evaluated sample, and a confidence level to see the point estimate, lower and upper bounds, and margin of error. This is useful when a single percentage would hide how much uncertainty remains.
Here, p̂ = successes ÷ total, z is selected from the confidence level, and n is the evaluated sample.
The point estimate describes this sample; the interval shows the sampling uncertainty around it. Compare the lower bound—not only the point estimate—with any operating requirement.
This statistical interval does not account for selection bias, label errors, dependence between records, or future drift.
With 920 successful outcomes out of 1,000 at 95% confidence, the observed rate is 92.00% and the Wilson interval is approximately 90.16% to 93.53%.
It gives a plausible range around the observed rate, reflecting uncertainty from the sample size rather than guaranteeing where the true rate lies.
More observations reduce sampling uncertainty, so the standard error and the resulting margin of error become smaller.
Yes. The calculator uses the Wilson score interval, which remains useful near 0% and 100% where a simple normal interval can perform poorly.
Use the level required by your reporting standard. Higher confidence produces a wider interval when all other inputs stay the same.
No. It quantifies random sampling uncertainty only; biased labels, drift, and nonrepresentative samples require separate investigation.
| Variable | Meaning |
|---|---|
| p̂ | Observed success proportion |
| n | Evaluated sample size |
| z | Confidence-level critical value |