skillrank_skill ▸ PostHog/diagnosing-experiment-resultsMIT · open registry

skill detail

← registry

Diagnosing Experiment Results

PostHog/diagnosing-experiment-results

Diagnoses bias, anomalies, and strange-looking results on a specific PostHog experiment. Covers empty / 0-exposure experiments, sample ratio mismatch, identity fragmentation, multi-variant exposure, uneven-split exclusion bias, significance traps (peeking, A/A, Bayesian vs...

communityprovisionalotherdiagnosesbiasanomalies

skillrank score

████░░░░░░░░░░░░23

SkillRank score blends community stars, real usage, and our eval lift; provisional until a skill is evaluated -- so popularity alone can't reach the top tier.

source

★ 58

PostHog/skills

skills/omnibus/diagnosing-experiment-results

open on GitHub ▸

eval status

eval pending

Success delta, token delta, and trial count are not available yet. No eval number is shown until this skill has measured results.

install

$ skillrank install PostHog/diagnosing-experiment-results
$ curl -fsSL skillrank.dev | sh