A fixed activity covariance gives a 1.40 ± 0.44 m/s amplitude for the 50.81-day GJ 674 template in 248 HARPS nights

We test a reported 50.81-day velocity template in 248 public GJ 674 HARPS nights. We keep the candidate period, eccentricity, argument and phase fixed. We fit a signed amplitude, two instrument offsets and local changes in the five parameters of known planet b. With a fixed SHO activity covariance from Liebing et al. (2024), the amplitude is 1.3977 ± 0.4354 m/s. The error is one conditional standard error. A diagonal control with the same marginal variances gives 1.6133 ± 0.2898 m/s. Time correlation therefore increases the error by a factor of 1.50. A nonlinear refit of planet b gives 1.4038 m/s and a profile interval of 0.9685–1.8392 m/s at delta chi squared equal to one. Across 64 corners of published marginal activity intervals, the amplitude ranges from 1.2550 to 1.4902 m/s and the local score ranges from 2.8276 to 3.7630. The corners measure sensitivity; they are not posterior samples. For one fixed template and a two-sided 1% test, the central model needs 1.6793 m/s for 90% rejection probability. The covariance is transferred from another reduction and is not fitted here. The same data selected the template, so the local score is not a search significance. The result does not establish a second planet. The package contains input data, source hashes, source code, numerical tables, figures and offline reproduction instructions. An independent rebuild from the full CDS table reproduces all 248 nightly times, velocities and errors exactly.

GJ 674 · HARPS · radial velocity · stellar activity · Gaussian process · SHO covariance · fixed template · planet candidate · reproducible analysis