OpenStack's simulation on the twelve-month window: two-cell bounds +0.0248 and +0.0234¶
partition-sensitivity-openstack.json, SPORC job 21794711 (CPU only, pinned at 89753ef), in
place of TIGRIS job 210563, which was cancelled when the TIGRIS queue slipped. It replaces the
artifact of 2026-09-30, which ran on the ten-month 1,635-change window at 300 trials per step.
Its placebo input is read from the cluster data root; that file, redacted, is byte-identical to the
committed rq1-placebo-openstack-v3.json. 1,000 studies per power step and 4,000 null studies,
K = 24, N = 1,850, and the twelve-month projection (2,017.6 changes, simulated at 2,017, rounded
down), with sigma_run calibrated on the pilot's 206 changes:
| spread point | spread | sigma_run | FP 97.5% | FP 95% | detects 97.5% / 95%, 2 cells | null bounded (97.5%, 2 cells) | reproducibility failure |
|---|---|---|---|---|---|---|---|
| pilot 90% lower | 0.0118 | 0.0085 | 0.0125 | 0.02325 | +0.0220 / +0.0202 | 0.991 | 0.000 |
| pilot estimate | 0.0140 | 0.0111 | 0.0125 | 0.0235 | +0.0232 / +0.0210 | 0.991 | 0.003 |
| sizing bound (90% upper) | 0.0177 | 0.0152 | 0.012 | 0.0255 | +0.0248 / +0.0234 | 0.991 | 0.022 |
| pilot 99% upper | 0.0215 | 0.0195 | 0.01075 | 0.02575 | +0.0274 / +0.0256 | 0.991 | 0.077 |
The false-positive rate sits at or within Monte Carlo error of nominal at every point (0.01075 to 0.0125 against 0.0125, 0.02325 to 0.02575 against 0.025, Monte Carlo errors 0.0018 and 0.0025 over 4,000 null studies). Reproducibility holds at the spread K was sized on (0.022 against beta = 0.05) and fails past it (0.077 at the 99% bound), as sizing at a 90% bound intends.
Read under the registered rule at two cells, the count once Wikimedia's admission lands:
OpenStack's bounds are +0.0248 at 97.5% and +0.0234 at 95%, each cell at power 0.9747. The
one- and three-cell bounds are in the artifact (by_cells). Against the replaced artifact
(+0.0239 and +0.0223 at a per-cell power of 0.928) the bounds rise by about 0.001; the two runs
differ in per-cell power, window size, trials and resamples, and the rise is within the replaced
run's bisection resolution, so no one cause is read from it. registered-decisions.md and the
Stage 1 sample-size paragraph take OpenStack's figures from this artifact.