D1268 retains a longer recovery but D992 changes direction in 24 local fits of KIC 3542116

We test three known dips in the DR25 Kepler light curves of KIC 3542116. We fit eight local models to each dip. The models vary the flux product, quality mask, and baseline order. D1268 has a recovery-to-ingress width ratio of 1.53 to 2.09 in all eight fits. D992 changes direction, with ratios of 0.83 to 1.27. D1176 has ratios of 1.18 to 1.22. A conditional test adds symmetric events to real off-event noise on the event time grids. For D1268, zero of 31 quality-zero controls and one of 31 all-finite controls have a score at least as high as observed. The other events overlap the control scores. These small, overlapping control sets do not give a survey false-alarm probability. This study supports a stable local shape direction for D1268 under the tested choices. It does not show a new object or supply independent instrument confirmation. The package gives code, derived samples, fit tables, noise tests, figures, exact input hashes, and a retrieval script.

Kepler · exocomets · KIC 3542116 · transit asymmetry · quality flags · injection tests