Is that gravitational wave signal really eccentric, or just bad priors?

Claims of orbital eccentricity in gravitational-wave events depend heavily on arbitrary analysis choices — specifically, which reference frequency is used and what prior distribution is assumed. Applying a new reference-frequency-independent detection statistic to GW200105, a neutron star–black hole merger flagged as eccentric by multiple groups, the evidence for eccentricity nearly vanishes (ln Bayes factor ≤ 0.9). Since eccentricity is the clearest fingerprint of how a binary formed, getting the measurement right matters enormously for understanding the origins of these mergers.