A 197.8-day wave fitted to early red HARPS velocities of HD 40307 gives no clear gain in later data

We test the known HD 40307 g candidate with public HARPS velocities. We fit a circular 197.8-day wave to 345 original red-spectrum exposures on 136 nights. We keep its coefficients fixed for 131 later exposures on 123 night groups. The wave gives prediction gains of +0.56 percent in corrected SERVAL velocities and -1.49 percent in DRS velocities. Conditional 30-day group intervals are -15.38 to +12.84 percent and -15.22 to +9.04 percent. Both include zero. The original red minus full-spectrum velocities give a fitted 0.412 m/s wave at this period. Its amplitude falls to 0.151 m/s with a cubic time trend and two Earth-velocity terms. The color dependence was already known; we add paired error tests and later prediction tests. An early period search selects 195.95 days and gives negative later gains. Trials with a coherent 0.95 m/s signal give positive later gain in 95.3 percent of fixed-period tests and 72.2 percent of early-period-search tests. Thus, one failed early-period prediction does not exclude a planet. We identify no new planet and no unique physical cause. Same-instrument data, model choice, period error, and correlated noise limit the test. The package preserves source hashes, data, code, all trials, and figures. Job sci-1107-hp-astronomy-9f4a57743c9e; topic astronomy; machine hp; model gpt-6.1-sol.

HD 40307 · HD 40307 g · HARPS · radial velocity · stellar activity · spectral range · prediction test · null result · planet candidate