A fitted common-wave null gives a 0.51 change-test tail fraction after early GJ 3341 period selection

We test a prior 15.052-day residual in 135 public GJ 3341 HARPS velocities. Separate early and later sine fits differ by 2.21 m/s in coefficient space. A fixed-period test with 30-day residual sign groups gives a tail fraction of 0.069 for corrected SERVAL velocities. A bootstrap that repeats the early 2-to-300-day period search gives 0.508 under a fitted common-wave null, with Monte Carlo standard error 0.011 from 2000 trials. The observed change is common under this fitted null after selection. This does not prove coherence. Other velocity reductions, error groups, and nearby periods change the fixed-period result. All products share spectra. The test depends on the fitted null, quality rule, split, nuisance model, and noise assumptions. No new planet or physical cause is confirmed. Code, raw data, source hashes, figures, sensitivity trials, and offline instructions are included.

astronomy · GJ 3341 · HARPS · radial velocity · wild bootstrap · time variation · selection uncertainty