A constant drift reduces the summed spin prediction error for selected 2022–2026 EXO 2030+375 data at brightness power 6/7

We test whether a fit to the 2021 giant outburst of EXO 2030+375 predicts later spin changes. We use public Fermi/GBM frequencies and Swift/BAT rates. We fit integrated brightness at power 6/7 on MJD 59400–59600. We then test 154 pairs in 14 groups from February 2022 to October 2026. The model with zero constant drift predicts a summed change of 63.79 microhertz. The measured change is 52.00 microhertz. A constant drift, fitted on the same training data, changes the prediction to 52.31 microhertz. The group root mean square error falls from 1.95 to 0.73 microhertz. The error for individual pairs stays near 0.33 microhertz. These results apply to 426 covered days at detected pulse epochs. The first test group supplies 73% of the primary prediction. Planned changes to power, smoothing, and gap cuts give measured-to-predicted ratios of 0.777–0.970 without drift and 0.953–1.047 with drift. The summed error thus depends on the empirical model. It does not establish a new physical torque state. We preserve the source snapshots, a duplicate-row audit, all pair selections, code, figures, and added error checks.

EXO 2030+375 · accreting pulsar · Fermi GBM · Swift BAT · spin frequency · prediction · reproducible analysis