Fitting the common period reduces a Ross 128 interval-change score by 81% and 97% in 305 HARPS exposures

We test whether a common circular velocity wave fits 305 selected Ross 128 observations from HARPS-RVBank version 02. Separate coefficients in 167 older and 138 later exposures improve the weighted score by 68.31 in SERVAL and 51.77 in DRS at a fixed 9.8658-day period. Fitting a common period over 9.83 to 9.93 days lowers these improvements to 12.99 and 1.54, reductions of 81.0% and 97.0%. Common-wave null draws give conditional tail fractions of 0.047 and 0.490 with 10-day sign bins. The SERVAL fraction changes from 0.0255 to 0.1075 across tested noise and baseline controls. A finer period grid and direct fits confirm numerical accuracy. The result shows that a fixed-period disagreement is not a robust physical-change test in this model. The reductions share spectra. The model omits eccentric motion and full activity covariance. No independent instrument is fitted, no new planet is found, and no orbital change is established. Original code, aggregate evidence, exact source hash and external retrieval steps are included; external database records are excluded.

astronomy · Ross 128 · GJ447 · HARPS · period uncertainty · coherence · radial velocity · bootstrap · null result