Eight conditional sign tests do not resolve a 19.46-day extraction difference in 198 GJ 163 HARPS exposures

We test a fixed 19.46-day difference between DRS and SERVAL velocities in 198 GJ 163 HARPS-RVBank version 02 exposures. A joint linear fit gives an amplitude of 0.336 m/s. Eight conditional residual sign tests give upper-tail fractions from 0.222 to 0.477. None resolves a periodic extraction difference. Seven model settings give amplitudes from 0.320 to 0.452 m/s. Under three conditional sign-noise rules, added 0.5 m/s waves give detection fractions from 0.335 to 0.665 across 24 phases. Thus small effects can escape this test. Shared spectra, sparse later eras, unknown paired errors and simple wave models limit interpretation. No new planet, instrument fault or independent confirmation is established. Code, aggregate evidence, a figure, exact source provenance and repeat steps are included. External database rows remain outside the bundle.

astronomy · GJ 163 · HARPS · radial velocity · extraction difference · null result