Fixed-shape eclipse fits to TESS sectors 96 and 103–106 do not resolve a depth change in WASP-18 b

We test whether the measured TESS secondary-eclipse depth of WASP-18 b changed between a 2025 visit and four 2026 visits. We use 99 eclipses from public 120-second SPOC light curves. A fixed eclipse shape and a separate quadratic time background describe each event. No event is removed because of its flux or fitted depth. The inverse-variance mean depth changes by -35.6 parts per million (ppm), with 2026 minus 2025 as the sign rule. A bootstrap that resamples complete events within each sector gives a nominal 95% interval of [-95.3, +18.7] ppm. Blocks of two or three events also give intervals that include zero. Across 18 fits with different backgrounds, flux products, times, or shapes, point estimates range from -44.5 to +6.2 ppm. All approximate 95% intervals from event errors include zero. A constant synthetic phase curve gives a false year contrast of 0.024 ppm for the primary fit. It also gives a 24 ppm bias in absolute depth. These data do not resolve a change under the tested models. They do not prove constant planet emission. A shared error within the single 2025 sector can escape the event bootstrap. The package includes the ten source light curves, source hashes, the analysis plan, code, numerical results, and figures.

WASP-18 b · TESS · secondary eclipse · planet emission · time series · photometric variability · reproducible research