The 2021–2026 BAT fit gain for IGR J16479−4514 changes sign between tested periods of 11.888 and 11.894 days

We tested whether early long-cycle profiles improve later Swift/BAT fits for two X-ray binaries. The early interval is MJD 53416–59515. The later interval is MJD 59516–61333, after the last day in Islam et al. (2023). The primary model has a fixed published period, two long-cycle harmonics, four orbit harmonics, and weights set by the early data. Only the later constant, trend, and orbit terms are refit. For IGR J16479−4514 at 11.894 days, the early profile lowers later weighted squared error by 1.283 percent. A nominal 97.5 percent interval from 60-day residual blocks is 0.518 to 2.035 percent. At 11.888 days, one published period error lower, the gain is −0.664 percent. Its interval includes zero. At 11.900 days the interval also includes zero. Equal weights also remove a clear positive lower limit. Other checks retain a positive gain at 11.894 days, including single-pointing data and removal of each later block. For IGR J16418−4532 at 14.702 days, the gain is −0.555 percent, with an interval of −1.433 to 0.471 percent. Recovery tests show low power for its early amplitude. The measured gain depends on the assumed period and weights. It does not establish a stable physical clock or cycle absence. The package includes four original BAT tables, source hashes, code, all checked choices, figures, and repeat instructions. The intervals assume independent, approximately symmetric block errors and hold periods and weights fixed.

Swift BAT · IGR J16479-4514 · IGR J16418-4532 · superorbital modulation · fixed profile transfer · period uncertainty · wild block bootstrap · reproducible data