Early ZTF fits recover part of the later variation near TOI-1546.01 but leave model-dependent timing offsets

We tested a published eclipsing-binary lead near TOI-1546.01 with a time split of 1,504 frozen ZTF observations. Separate-band Fourier fits to 940 early measurements predict 564 later measurements with 18.0 to 23.5 percent less squared error than constant brightness. Fits use a local period interval near the known 1.13-day signal. Later timing offsets range from 20 to 42 minutes as waveform complexity changes. Resampling both time segments gives broad, model-dependent timing intervals. The result supports a repeatable component in the ground data. It does not establish that this star causes the TESS transit, strict phase stability, or orbital period change. Residual scatter is much larger than quoted errors. Crowding, calibration, and waveform errors remain possible causes. We include the frozen input, provenance, code, predictions, controls, bootstrap samples, and reproduction instructions.

TOI-1546.01 · ZTF · eclipsing binary · prediction · timing uncertainty · contamination