coiled / benchmarks

BSD 3-Clause "New" or "Revised" License
28 stars 17 forks source link

A/B tests - calculate confidence interval against all combinations #1436

Closed fjetter closed 6 months ago

fjetter commented 6 months ago

When running multiple configurations it is sometimes useful to caculate the differences between the various configurations and not just the baseline.

I think this is useful and cheap enough for us to do every time.

phofl commented 6 months ago

thanks