With fixed independent noise, annual curves leave 8.7 to 27.7 percent of 389.7 day signal information in 143 GJ 1151 velocities

We test how free annual curves change information about the reported 389.7 day velocity signal of GJ 1151. We use 97 CARMENES and 46 HARPS-N measurements from the public CDS tables. The main noise variance is the listed variance plus fixed extra scatter of 1.55 and 2.07 m/s, respectively. Each fit has instrument offsets and a common quadratic time curve. Adding a separate 365.25 day sine and cosine for each instrument leaves 8.7 to 27.7 percent of the signal coefficient information over all circular phases. This exact range uses observation dates and weights. The fitted amplitude changes from 3.06 to 3.19 m/s. Its local standard error increases from 0.39 to 1.31 m/s. At 389.7 days, the cost gain falls from 63.64 to 5.94. With annual curves, the largest gain in a targeted 300 to 500 day search is 9.08. Gaussian noise gives an equal or larger gain in 484 of 8192 draws, or 5.91 percent. The fraction ranges from 1.05 to 58.94 percent across the stated noise controls. Separate data and mean controls also retain less than 33 percent of the information at every phase. These results show a limit in separating nearby periods under the stated models. They do not show that annual instrument errors exist. The test omits HPF and does not confirm or reject the planet. We provide source data, hashes, code, figures, simulation outputs, and a fresh repeat check.

GJ 1151 · radial velocity · annual nuisance terms · signal information · CARMENES · HARPS-N · fixed Gaussian noise