mars
MARS-639
estadisticas: dead outlier-detection rule, never fires (since MARS-193)
Backlog low
cmcoder02-mars-cc
computePositionDetail's >2-sigma outlier test on the 4 group-A tooth positions is unreachable by construction: for n=4, max possible population-sigma deviation from the mean is (n-1)/sqrt(n) = 1.5, always below the 2-sigma threshold. isOutlier has been false for every row since MARS-193 -- the 'Atipico' badge has never rendered in prod. Pinned with a KNOWN-DEFECT test, not fixed, because the right behavior for a 4-position report is a SCOPE decision (redesign the test for small-n, or drop the badge/feature), not a typo fix. Needs Elazar's call on which.
Activity
-
coder-mars-ca / coder
-
coder-mars-cc / coder
-
coder02-mars-cc
-
Coder lane finding, no code shipped: outlier rule is k-sigma at fixed n=4, max attainable 1.5 sigma vs a 2 sigma bar -- unreachable by construction, so the badge has never fired since MARS-193. Already pinned as a known-defect test in multiradicular-stats.test.ts under MARS-637; no implementation work remains. Remaining half is a product call with two incompatible answers (redesign the rule for small-n, or drop the badge) -- both change what estadisticas asserts to teachers. Routed to pm-mars-cc as blocked 2026-09-01 for an Elazar decision. Left in backlog, unassigned, nothing touched.
bug
2026-08-14 by wi-cli-venus
6w ago