Fixed-phase daily fits give a lower hard-to-soft cycle amplitude ratio for LMC X-4 in 2024–2026 than in 2017–2024

We compare the strength of the long X-ray cycle of LMC X-4 in public MAXI 4–10 keV and Swift/BAT 15–50 keV daily data. Both bands use the same days, exclusions, and weights. A fixed phase model and four Fourier harmonics give a first-harmonic amplitude in each band and interval. The ratio of the hard amplitude change to the soft amplitude change is 0.765. Its conditional 95% interval is 0.656 to 0.886 from 4,000 paired three-cycle block bootstrap draws. The soft amplitude changes by a factor of 1.233 [1.069, 1.425]. The hard amplitude changes by a factor of 0.943 [0.836, 1.050]. The reference has 1,961 matched days from June 2017 to May 2024. The later sample has 389 matched days from June 2024 to September 2026. Wider outburst exclusions, equal weights, a linear phase, and equal interval lengths retain the lower ratio. The second later half is sparse and has 73 non-estimable resamples. A separate Crab mean-rate control gives 1.074 [1.047, 1.093], in the opposite direction. A post hoc check before the first listed nearby outburst gives 0.774 [0.629, 0.918]. These results measure the selected daily products. Calibration errors, source confusion, and uneven exposure can remain. The evidence does not select a physical cause or establish a sudden transition. The package preserves source bytes, provenance, code, selections, bootstrap draws, figures, and instructions for a repeat without a download.

LMC X-4 · MAXI · Swift BAT · superorbital modulation · Fourier amplitude · paired block bootstrap · public data · source confusion