A decreasing period improves 2024–2025 ZTF predictions for J1539 with fixed flux profiles

We test whether a changing clock improves later survey predictions for the known 6.91-minute white dwarf binary J1539. We fix a public ZTF response with 1,652 rows and keep 1,452 g and r measurements with zero flags, magnitude errors below 0.20, and 30-second exposures. We fit separate flux profiles for each filter, separate object constants, and a shared constant or quadratic phase clock. A broad search tests both signs of period change. Fits to 1,343 points before 2024 then predict 109 points in 2024–2025 with all coefficients fixed. The quadratic clock reduces the later weighted loss from 631.41 to 159.97, a gain of 471.44. The gain is positive in each later year. The full-sample period derivative is −2.390 × 10⁻¹¹ seconds per second. This is a new fixed-data prediction test of known orbital decay. It does not improve the best published precision. The split is retrospective; its numerical reference comes from a 2025 paper. The result depends on stable filter profiles and the selected detections. The package includes the raw response, time and row audits, code, fitted clocks, individual predictions, residual block checks, and figures.

J1539 · white dwarf binary · orbital period change · ZTF · public data · prediction test · reproducible analysis