c00069f208
The report told readers to 'check significance' but gave no way to. Add a 95% Wilson interval on each variant's conversion rate (accurate for the small n and near-zero rates this report has) and a two-proportion z-test between the two leaders, Bonferroni-corrected for the three pairwise comparisons, with an explicit small-sample guard. Printed to the console and posted as a Discord field so 'is there a winner yet' has a numeric answer. |
||
|---|---|---|
| .. | ||
| Commands | ||