A Real, Reproducible, Galactic-Latitude-Dependent Sign Reversal in Blend-Covariate Contamination Tests for Threshold-Limited Proper-Motion Surveys: A Pre-Registered Fourteen-Field Analysis of CatWISE2020/unTimely

<p dir="ltr">Exomoon stability is usually assessed by comparing a single assumed planetary tidal quality factor Qp against a critical value Qerit below which a moon is disrupted within the system's age a comparison that treats Qp as known, when it is unconstrained for every exoplanet and spans four to five orders of magnitude even among the Solar System's own giant planets. Building directly on the standard constant-Q tidal-migration formalism (Goldreich & Soter 1966; Barnes & O'Brien 2002; Domingos, Winter & Yokoyama 2006), we replace this binary verdict with a survival probability Psurv (ao, tage), obtained by marginalising Qerit over an explicit prior on Qp, and pair it with a first-order log-space uncertainty budget that decomposes Qcrit's sensitivity to the planet's mass, radius, Love number, orbital separation, and the satellite's own properties separately for the inward (Roche-bound) and outward (Hill-bound) migration regimes, which we show are governed by different subsets of these parameters.

Applied to a K2-18b analogue, whose planetary spin state (and hence migration direction) is unmeasured, both the inward and outward hypotheses give a poor survival outlook (Psurv≤0.4 across the dynamically accessible orbital range at the system's ~2.4-3 Gyr age), and this qualitative conclusion is robust to substantial changes in the adopted Qp prior. A Kepler-1625b analogue is, by contrast, robustly stable (Psurv 1) under every prior we test, because its much wider orbital separation not its migration direction protects any putative moon; two further illustrative systems bracket these cases at either extreme.