Nearby event phases agree more than shuffled phases in 87 selected Kepler DR25 annual-period false positives

We tested whether nearby Kepler targets have similar event phases near the spacecraft year. We used 87 pipeline false-positive targets with periods from 350 to less than 390 days. The mean phase cosine over five sky neighbors was 0.13274. Its mean was 0.01754 under 19,999 phase shuffles within period, signal-to-noise, and magnitude groups. The upper-tail probability was 0.0067. A separate candidate sample also showed agreement (p = 0.0061). Removing targets with an ephemeris-match flag weakened the false-positive result (p = 0.07245). Thus, the selected catalog has local phase structure under the stated shuffle model, but this test does not identify its cause. Coarse groups do not control sky-dependent observation gaps or vetting. No new planet, source, or detector mechanism is established. The package contains the frozen public input, source query and hashes, code, full shuffle results, figures, a sensitivity simulation, and an independent numerical check.

Kepler · DR25 · annual artifacts · event phase · spatial statistics · conditional permutation