The planted check at a half passes 10 of 10 runs: registered at a half, read on the planted run only¶

The ten runs planned above, repeated at PLANT=0.5 (jobs 210234 to 210243, pinned at b05a2d8), rq1-partition-openstack-*-n1850-plant0.5.json. All ten pass. Every half's lower bound is above zero (least +0.017); per-half estimates run from +0.053 to +0.142. The pass probability projected by the logged rule (normal per half, interval half-width over 1.96 as the standard error, product over halves) is at least 0.9989 for every run and 0.9999 on average. Both conditions of the rule hold, so check 5 is registered with the plant at a half: neutral.planted_convention, which reads the plant's strength and window from the run's corpus path and refuses any other, and partition_pilot.py --planted, which halts the cell when it fails.

The draft's second clause is dropped. The Stage 1 draft also required that "unplanted must not" pass. Under the repeated-partition design the unplanted dev runs are H1's own data: a pass there is H1 holding on the development window, so a halt on it would turn on the outcome, which an outcome-neutral test may not. For the record, none of the 24 unplanted OpenStack pilot runs passes (23 fail, 1 mixed). The check reads the planted run only.

Planned before any run: the same measurement on Wikimedia. Check 5 runs on every admitted organization, and its pass rate was measured on OpenStack only. Ten runs at PLANT=0.5 on Wikimedia's first ten admissible partitions (seeds 1, 2, 3, 5, 6, 7, 8, 9, 10, 11) at training seeds 1 to 10, N = 1,850, corpus v3, the jobs otherwise as its pilot's. The same rule: registered for Wikimedia if all ten pass and the projected pass probability is at least 0.95; otherwise the check is redesigned for Wikimedia before Stage 1 is submitted. The rates are logged whichever way they fall.