Literature period and reference time pairs reduce mean timing offsets below one minute in the last TESS seasons of HAT-P-16 b, KELT-18 b, and WASP-20 b

Do three extreme catalog transit timing offsets require a change in orbital period? We selected HAT-P-16 b, KELT-18 b, and WASP-20 b from an earlier SCI survey before fitting the data. We measured 68 complete transits in all 12 public TESS SPOC two-minute sectors returned by a new MAST query. In the last seasons, the mean measured-minus-predicted times from the saved NASA Exoplanet Archive composite table are -13.899, -75.320, and -12.746 minutes. The latest complete named-paper pairs in the saved archive give -0.387, -0.328, and +0.961 minutes. These pairs are from Yalçınkaya et al. (2024) and Kokori et al. (2023). The first and third composite pairs mix newer periods with reference times inferred using old fixed periods. We reconstruct those reference times within 0.081 and 0.606 minutes. KELT-18 b uses an old discovery period that was already corrected in published work. Earlier TESS sectors predict the last season within the estimated errors for each system. Quadratic-minus-linear BIC values are +2.41, +2.66, and +2.49 under independent Gaussian timing errors. These data do not require curvature but give weak limits on period change. In an added HAT-P-16 b test, sector 84 occurs 2.308 minutes after the complete quadratic prediction printed by Sonbas et al. (2025). This is about 4.00 combined estimated standard errors even with the largest parameter covariance allowed by the printed marginal errors. The same decay rate remains compatible when reference time and period are fitted again (profile delta chi-square 0.090). This is a conditional prediction discrepancy, not an exclusion of that decay rate. Errors shared within a sector can change the interpretation. We preserve the input arrays, provenance, code, 68 times, model controls, full covariance, figures, and offline reproduction instructions. The result applies to these three selected systems and does not measure the rate of errors in all archive ephemerides.

TESS · transit timing · ephemeris · NASA Exoplanet Archive · HAT-P-16 b · KELT-18 b · WASP-20 b · orbital decay · reproducible research