Weighted 2021–2026 BAT fits support an 11.894-day component in IGR J16479−4514 under a block noise model

We test two known superorbital periods with public Swift/BAT data after the cutoff of Islam et al. (2023). The new interval is 2021-10-29 through 2026-10-20. We fix the periods and fit offsets, yearly orbital terms, and two superorbital harmonics. Earlier data set the weight rule. Restricted residual signs in calendar blocks give a conditional noise calibration. In IGR J16479−4514, 1184 daily rates support a mean 11.894-day component. Its fitted RMS is 0.0006403 count cm^-2 s^-1. The two-source adjusted tail fractions are 0.0004 for 64-day blocks and 0.0031 for full sign enumeration with 128-day blocks. Equal weights give an adjusted fraction of 0.0888 and do not meet the study threshold. Finer pointing data and an eclipse cut retain support but use the same observations. In IGR J16418−4532, 1183 daily rates do not resolve a fixed 14.702-day component. Recovery of its earlier fitted waveform is only 17.9% in conditional injections. Thus, non-resolution does not show that this cycle stopped. The result extends known-period measurements under a fixed mean model and independent, approximately symmetric noise blocks. It does not identify the physical cause. The package preserves raw data, provenance, code, random draws, figures, and offline reproduction checks.

Swift BAT · superorbital modulation · IGR J16479-4514 · IGR J16418-4532 · X-ray binaries · stellar winds · public astronomical data · conditional residual calibration