Lorentzian fits to 101 selected GX 301−2 flares give 0.45-day scatter about a quadratic clock in 2012–2026 BAT data

We test whether brightness can explain changes in the times of GX 301−2 X-ray flares. We use public Swift/BAT data in the 15–50 keV band. We fit each flare with a constant plus a Lorentzian curve. The main sample has 101 flares with sufficient coverage from 2012 to 2026. A quadratic clock leaves 0.454-day root mean square scatter. A fitted brightness term gives +0.127 day per doubling of peak height. Its 95% interval is [-0.157, +0.385] day when we resample blocks of three accepted cycles. Brightness removes 1.6% of the fitted timing variance. Alternate pointing samples give time differences with standard deviation 0.491 day. Sampling and fit errors therefore limit the physical meaning of the scatter. We provide a catalog, alternate fits, conditional simulations, figures, and frozen source data. The result describes this selected sample and these flare models. It does not measure a change in the true binary orbit. Earlier work already found flare-time scatter. This contribution gives current cycle data and tests a brightness correction to the flare clock.

GX 301-2 · Swift BAT · X-ray binary · flare timing · orbital evolution · reproducible catalog