BAT data from 2020 to 2026 do not confirm a 14.981-day cycle in IGR J16393−4643

We tested a proposed 14.981-day X-ray cycle in IGR J16393−4643 with public Swift/BAT data. Earlier work could not confirm a physical cycle. We used 19,638 later pointings on 1,919 observed days, after the earlier study's cutoff. We fitted the orbit and changes in the mean rate with the candidate cycle. At the fixed candidate period, simulations with random signs in 15–120-day calendar blocks give null tail fractions of 0.50–0.58. These results do not confirm the cycle. The tests recover the fitted early waveform in only 49–66% of trials across block lengths and phase offsets. Thus, the result does not show that the cycle stopped. A conservative conditional 95% RMS bound is 3.66–3.78 × 10⁻⁴ count cm⁻² s⁻¹ for a fixed signal with three harmonics. A search from 14.90 to 15.10 days gives a weak peak at 15.040 days. Its tail fractions after correction for this search are 0.0225–0.0430. The selected peak needs independent tests. All fractions depend on the stated block noise model. We also find 5,259 pointing dates listed one day late in both ASCII and FITS files. The primary analysis uses mission seconds. The package preserves the data snapshot, source records, code, simulation results, five figures, and offline reproduction instructions.

IGR J16393-4643 · Swift BAT · X-ray binary · superorbital modulation · time series · injection recovery · data quality · reproducible research