Earlier GX 1+4 flux fits give lower later mean squared error than constants for 60-day windows and higher error for 90-day windows
This study tests whether an earlier flux fit predicts later rotation changes in GX 1+4. We use frozen public Fermi/GBM pulse frequencies and Swift/BAT daily 15-50 keV count rates. We fit a line and a constant on accepted windows before 2022. We hold both fits fixed for later predictions. The 60-day rule accepts 28 earlier and six later windows. The later mean squared error ratio, flux model divided by constant, is 0.92395. The 90-day rule accepts 24 earlier and six later windows. Its ratio is 1.02266. The sign of the measured gain therefore changes with window length. The 30-day rule accepts no windows. The six later windows cover four years in 2022-2025. Resampling ranges depend on block length and are not calibrated confidence limits. Fits with separate year averages still have negative flux slopes. These data do not establish that the physical flux-torque relation has ended. The package preserves exact inputs, source URLs and hashes, all coverage cuts, source code, predictions, controls, and figures.
GX 1+4 · Fermi GBM · Swift BAT · pulsar timing · accretion torque · chronological prediction · public data
References
5- doi:10.1086/310666
This paper found a negative torque-flux relation in GX 1+4. We do not claim that relation as new.
- doi:10.1051/0004-6361/201117893
This paper found a negative short-term relation in early GBM and BAT data. It motivates the later prediction test.
- doi:10.1093/mnras/stx1045
This paper found changes in relation state and long-time torque noise. These results limit the physical interpretation.
- usesdoi:10.3847/1538-4357/ab855c
This paper describes the GBM pulsar monitoring method. We use the public GBM frequency history.
- usesdoi:10.1088/0067-0049/209/1/14
This paper describes the BAT transient monitor. We use its public daily count rates.
Cited by
4 publicationsPublic catalog entries on Sciscape.
- Fits to GBM counts with a cubic 768-second background rule give a 12.76-microhertz GX 1+4 frequency loss between two eight-day intervals in 2026
This prior study already reported smaller recent annual frequency losses. It defines part of the prior evidence.
- usesDirect Doppler shifts cannot explain smaller GX 1+4 spin loss in 2026 for total mass ≤5 solar masses, period ≥1100 days, eccentricity ≤0.2, and the tested errors
This study already noted smaller recent annual frequency losses and left direct orbital motion as a possible contributor. We add an explicit bound for that alternative on a new GBM snapshot.
- With random walk noise, 30-day unweighted BAT means lower later frequency RMSE for 4U 1907+09 at a 2019 split and raise it at a 2018 split
Related later-data flux test in GX 1+4. This study uses a different source, fits frequencies with random walk noise, and reuses no code or data from that study.
- Three 2017–2018 intervals reverse the 2023 ranking of two fitted torque models for Swift J0243.6+6124
Related earlier-to-later rate-spin tests show setting sensitivity for GX 1+4. No code or data are reused.