Skip to content
Browse › Mechanism Survival
Analytics module · as of

Mechanism survival: about half of testable ideas hold up

Survival rate = share of TESTABLE hypotheses whose latest recorded verdict is CONFIRMED_LOCAL / CONFIRMED_LOCAL_incl_2026_OOS / REPLICATED. A NULL/REJECT is honest market-efficiency evidence, not a failure. No $/edge/ROI claim.
View full size ↗
confirmednull (a finding)not testabledescriptivepending
Mechanism Survival
Chart: Mechanism Survival -- Survival rate = share of TESTABLE hypotheses whose latest recorded verdict is CONFIRMED_LOCAL / CONFIRMED_LOCAL_incl_2026_OOS / REPLICATED. A NULL/REJECT is honest market-efficiency evidence, not a failure. No $/edge/ROI claim.
scripts/platformkit/analytics_showcase/out/mechanism_survival.json

What it means

Of 287 hypotheses in the four sport ledgers, 256 were testable and 129 confirmed on their latest verdict, a 50.4% survival rate. Survival splits by sport: NBA holds up best at 0.56 while MLB is weakest at 0.3974.

Caveats & confounds

Category-level rates on tiny samples are not durable: rest_fatigue survives at just 0.3056 and weather_park at 0.0 on a single testable row.

Ask Scout about this
How does this system grade itself?Which sports have the most tested hypotheses, and how many held up?What kinds of effects survive your gate, and which mostly fail?Ask anything →