Adding sibling transits with fixed shapes and linear clocks gave a median absolute time change of 7.06 minutes in 46 selected Kepler events

We tested weak transit overlaps in a public Kepler timing catalog. A sibling is another planet around the same star. The source leaves events unflagged below a sibling-to-target depth-times-duration ratio of 0.25. We found 295 unflagged expected intersections before local timing-baseline cuts. A regression that includes 153 selected intersections gives a signed overlap coefficient of 0.795 across 26 informative targets. A descriptive whole-host resampling interval is 0.632 to 0.999. We selected three targets by expected information and fit 46 events with public archive flux. Adding fixed sibling templates at linear catalog clocks changes target times by a median absolute 7.06 minutes. The maximum is 26.96 minutes, and 43 changes agree with the proxy direction. The fits allow the target time, depth, duration, and local baseline to change. A free sibling amplitude gives a median absolute change of 3.79 minutes. A run in a new directory gives the same primary numeric outputs in the same software environment. The catalog association can include dynamical and stellar effects. Tests with Gaussian noise show mean timing offsets in some cases. The paired shifts depend on the specified templates, clocks, noise model, and numerical bounds. They are not established corrections to true transit times. We preserve original data, source hashes, the full overlap ledger, code, results, figures, and offline instructions.

Kepler · transit timing · transit overlap · catalog audit · joint transit models · model sensitivity · public astronomical data · reproducible research