A bounded outcome-odds factor of 12.16 is needed to reverse the Gaia RV rejection ranking in a frozen 734-TOI cohort

We test missing later verdicts in a published Gaia DR3 radial-velocity flag study. Of 734 TESS candidates that were open in 2020, 359 remain open in the frozen 2026 table. The full-cohort rejection fraction is bounded by 36.97 to 95.80 percent for flagged objects and 20.65 to 67.64 percent for unflagged objects. The unrestricted bounds permit either ranking. We bound unresolved outcome odds relative to resolved outcome odds within each flag group. An exact integer search shows that a common allowed factor of at least 12.1556 is needed to reverse the rejection ranking. Selecting one TOI per host lowers this factor to 10.2124. A finite follow-up certificate needs at least 37 favorable new verdicts under equal per-record cost; 37 flagged rejections suffice regardless of remaining outcomes. This is a best-case cost, not a predicted observing yield. The eventual CP/KP count remains between 204 and 563. These results support a conditional ranking more strongly than an absolute occurrence correction. Labels are inherited, CP/KP includes brown dwarfs, and the selection bound is an assumption, not an observed constraint.

TESS · Gaia DR3 · radial velocity · missing outcomes · verification bias · candidate reliability · partial identification