Skip to content

Comment on How Not To Run An A/B Test

Comments

Interesting that frequentist A/B software packages let you essentially break the test without telling you. Are there bayesian A/B testers that give you a likelihood ratio instead?

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.