Wilson lower bound · model result
90 positive out of 100 reviews. Observed positive rate: 90.00%.
Results are rounded to three decimal places.
What this score means
This is a cautious score for comparing positive review counts. It accounts for how little evidence a small sample provides; it is not the observed percentage of satisfied reviewers.
9 positive reviews out of 10 and 90 out of 100 both give 90% positive reviews. Their Wilson lower bounds are about 59.6% and 82.6%: the larger sample leaves less uncertainty under this model. A higher bound does not prove better product quality.
How it works
(p + z²/(2n) − z × √(p(1−p)/n + z²/(4n²))) ÷ (1 + z²/n)
Uses z = 1.959963984540054, the lower endpoint of a two-sided 95% Wilson confidence interval. This is a ranking statistic, not a star average or a guarantee about the true positive rate. It assumes independent binary observations.
NIST: confidence intervals for proportions
Results update as you type. Calculations run in your browser; submitting with JavaScript disabled sends the numbers to the server to calculate the result.