Earlier ephemeris pairs reduced 3-sigma offsets from 29 to 12 for 129 TESS hot Jupiters with zero assumed covariance

We tested whether earlier published ephemeris pairs explain large transit timing offsets in a recent SCI study. We fixed its 153 quality-selected sector times. We selected period and reference-time pairs from the NASA Planetary Systems table by their stated prediction error, without using measured residuals. Each pair had to state BJD-TDB and have paper and archive release dates before the tested sector. Of 153 planets, 129 had an eligible pair. Within these same 129, the number above 3 sigma fell from 29 with composite values to 12 with earlier pairs. The number above 5 sigma fell from 13 to four. These counts assume zero period-reference-time covariance and use stated formal errors. The median absolute offset fell from 6.63 to 0.78 min, while the median prediction error fell from 4.32 to 0.74 min. Eight original outliers remain, 21 fall below 3 sigma, and four others cross above it. All 12 selected-pair outliers remain above 3 with the largest prediction error allowed by covariance. The three largest original normalized offsets, HAT-P-16 b, KELT-18 b, and WASP-20 b, become smaller than 1.5 sigma. Separate fits to their same TESS light curves agree with the saved times within 0.13 min. We preserve the data, code, source checks, selection audit, and figures. This tests direct composite-table use, not the archive transit prediction service. The errors are not calibrated for all noise or model effects. No new planet or changing orbit is established.

transit timing · ephemeris · TESS · hot Jupiters · NASA Exoplanet Archive · reproducible research · catalogue validation