Fitting orbit shape removes over 90% of local amplitude information for some 365.25-day orbits at eccentricity 0.8 in all 152 selected HARPS series under independent noise

We tested how an unknown eccentric orbit shape changes local information about velocity semi-amplitude K in 152 selected public HARPS schedules. The main model has a fixed period of 365.25 days, eccentricity 0.8, independent nightly errors, era offsets and a linear trend. We compare information with orbit shape fixed and with mean anomaly, argument of periastron and eccentricity fitted. Every series has tested angles that retain less than 10% of fixed-shape information. For the median series, this occurs in 39.4% of a uniform 64 by 64 angle grid. In 148 series, some such cells also have fixed-shape information at least half of a uniform reference. Complete uniform phase coverage retains 43–55% at the same eccentricity. The median of within-series median retention is 0.297 for K and 0.633 for the mass-related scale K sqrt(1-e^2). Controls vary period, eccentricity, weights and the fitted period. Independent source and numerical checks passed. We provide the 152-series information atlas, exact time-basis reduction, source fields, code, figures and instructions. This is a local information test under fixed noise assumptions. It does not search the measured velocities, measure planet masses or establish a new planet. Grid fractions are not detection probabilities.

HARPS · radial velocity · eccentric orbit · Fisher information · observing schedule · orbit uncertainty · reproducible atlas