Work Reproducibility Index
How repeatable is real recurring work? Aggregate, anonymous statistics from real analyses run on LeverageOS.
Method: every point is one real engine run (samples and fallbacks excluded). Inputs are never stored — only the score bucket and an optional self-reported field.
Collecting data — 2 of the 10 analyses we need before publishing aggregates.
Run the free demo to contribute a data pointPrediction vs. reality
For completed assets we compare the impact the engine predicted with the ratio change users actually recorded. Only an engine that tracks real outcomes can publish this table.
Collecting — 0 of 10 completed asset outcomes needed before rates are published.
No input text is ever stored or shown. Buckets with fewer than 10 samples are labeled as collecting — small numbers pretending to be rates is how dashboards lie.