Twelve conditional tests do not resolve an Earth-velocity interaction in 224 GJ 674 extraction differences

We test whether the difference between two public HARPS velocity reductions has an orbital wave that changes with Earth velocity. We use 224 GJ 674 exposures in five dense seasons from HARPS-RVBank version 02. At the fixed 4.69507-day period, the interaction amplitude is 0.233 m/s per 30 km/s. Twelve conditional residual sign tests give tail fractions from 0.171 to 0.426. A fixed component fitted in other seasons raises held-season squared error in three of five seasons. Seven model and weight controls retain null tail fractions above 0.15. Conditional sensitivity trials recover 81 to 91 percent of exact 0.5 m/s interaction vectors under one noise rule. Thus the tests do not resolve this restricted extraction effect. Shared spectra, unknown error correlations and a linear response model limit the result. No new planet, physical cause or independent-instrument confirmation is established. Code, aggregate evidence, source hashes, a figure and repeat steps are included. External source data remain outside the bundle.

astronomy · GJ 674 · HARPS · Earth velocity · velocity extraction · instrumental artifacts · null result