Evidence for period decrease in 2009–2025 PTF J0533+0209 survey data weakens when PTF profiles or noise groups change

We test public survey evidence for period decrease in the known double white dwarf binary PTF J0533+0209. The sample has 5,247 PTF R and ZTF g/r measurements from 2009–2010 and 2018–2025. We check exposure times and use midpoint BJD TDB. The model has a quadratic phase, separate band profiles, and free night or object means. With one profile for both PTF seasons, the fitted rate is (−3.39 ± 0.84) × 10⁻¹² seconds per second. This error is one standard error from seven-day residual groups. A constant-period fit has 19.25 more weighted squared loss. With free PTF season profiles, the rate is (+1.20 ± 3.19) × 10⁻¹² seconds per second. ZTF alone also leaves the rate unresolved. We use 499 residual draws for each of three time-group widths. The plus-one null tail fractions are 0.008, 0.142, and 0.316 for one-, seven-, and 30-day groups. Thus the fitted standard error does not give robust detection strength. In a retrospective test, clocks and profiles fitted through 2021 predict 327 ZTF rows from 2022–2025. The changing-period fit gives 4.04% more loss than the constant-period fit. This test does not propagate earlier-fit uncertainty. All selected PTF rows are flagged as not photometrically calibrated. Free means cannot replace independent calibration. The evidence from this survey test depends on profile and noise assumptions. The PTF data overlap the discovery observations, so this is not a fully independent confirmation. We preserve exact public responses, provenance, code, controls, random draws, figures, and offline reproduction instructions.

PTF J0533+0209 · double white dwarf · orbital period change · PTF · ZTF DR24 · survey photometry · profile stability · barycentric time · reproducible research