Local fits to 129 TESS transits leave a 0.5 ms per year period decrease unresolved for LTT 9779 b

We test whether public TESS light curves can resolve a small period decrease for LTT 9779 b. We measure 129 transit times in five native 120-second sectors from 2018 to 2026. A fixed transit template and local residual noise refits give a constant period change of -9.6 +/- 6.0 milliseconds per year (one standard deviation). The conditional Gaussian 95% interval is -21.2 to +2.1 milliseconds per year. The data do not establish a period change. We train the template and a constant period on 112 transits before 2026. The 17 separate test transits have a mean time offset of -25 +/- 27 seconds when shared prediction errors are included. A proposed decrease of 0.5 milliseconds per year gives only a -1.4 second mean signal in this test. Its expected signal to noise in the full quadratic fit is 0.084. These data cannot resolve that effect under the tested model. Changes in the flux, baseline, limb darkening, and residual block length preserve this result. Errors shared within sectors can weaken the constraint. The contribution is a homogeneous timing data set through 2026, an explicit prediction test, and a measured sensitivity. Earlier studies also found no clear period change. We preserve the five original light curves, source code, hashes, selection records, results, figures, and repeat-analysis instructions.

LTT 9779 b · TESS · transit timing · orbital period change · orbital decay · prediction test · residual bootstrap · public astronomical data