A shared delay fit leaves 12.14 seconds of RMS scatter in six Beta Pictoris signals in 10-day TESS Sector 98 windows

We test whether six Beta Pictoris pulsation signals share one nonlinear delay in TESS Sector 98, from November 2025 to January 2026. We fit 37 fixed published frequencies in six 10-day windows. Each signal has its own phase offset and frequency correction. The covariance-weighted shared fit leaves 12.14 seconds of root mean square residual scatter. An exact shared fit needs at least one phase-time estimate to change by 20.94 seconds. A saved optimization certificate and a six-value contrast verify this bound on the measured values. With 0.25-day residual blocks, the main chi square is 46.54 for 20 degrees of freedom. Its nominal tail area is 0.000679. Only one of 4000 centered block-sign draws has at least this chi square. The 12-day fit gives chi square 22.99 for 15 degrees of freedom and tail area 0.0844. Thus, the 12-day test does not reject a shared delay at 5%. Removing F8 also weakens the result. Window choice and the conditional error model limit the conclusion. Earlier work already found unstable timing. We add later phase values, their full covariance, original data, code, figures, and reproduction checks. We establish no new planet or physical cause.

Beta Pictoris · TESS Sector 98 · stellar pulsations · phase modulation · time delay · shared delay · pulsation timing · public astronomical data · covariance · reproducible research