Gaia DR3 radial-velocity scatter flags on-target binaries among TESS planet candidates: 41 of 46 flagged 2020 candidates became false positives, and 133 open PCs show stellar-mass scatter

Question: Does the Gaia DR3 radial-velocity (RV) variability test find on-target stellar binaries among TESS Objects of Interest (TOIs), before ground spectroscopy? Method: We matched 7,502 of 8,175 ExoFOP TOIs (2026-11-09) to Gaia DR3. For 3,334 TOIs with G_RVS <= 12, at least 10 RV epochs and template temperature 3900-8000 K, we applied the Katz et al. (2023) criterion (rv_chisq_pvalue <= 0.01, rv_renormalised_gof > 4). We derived the epoch RV scatter from radial_velocity_error, subtracted a magnitude noise model, and converted the scatter to a minimum companion mass at the TOI period. Results: The flag marks 5 of 447 hosts of planets with K < 0.3 km/s (1.1%), 46 of 50 spectroscopic-binary false positives (92%; ROC AUC 0.97), and 24 of 54 on-target eclipsing binaries. For 734 testable TOIs that had no TFOP verdict on 2020-11-20 (before Gaia DR3), 41 of 46 flagged TOIs with a 2026 verdict became false positives (89%), against 108 of 307 unflagged TOIs (35%; p = 3e-12). The 5 flagged TOIs that became confirmed are all brown dwarfs. The scatter gives the published companion mass within 26% for WASP-18 b, TOI-4603 b, TOI-5882 b and KELT-1 b. Now 386 open PC/APC TOIs have the flag; 133 TFOPWG PCs without a binary note have a Gaia NSS period match or stellar-mass scatter. WASP-163 and TOI-7154 show RV scatter that their known planets cannot cause. Limitations: The test applies to 41% of TOIs, sees only the target star, gives false flags for broad-lined and young stars, and a flagged planet in a wide binary is not a false positive. FP reasons come from comment keywords. TFOP observers could see Gaia data, so the forecast is not fully blind.

TESS · TOI · false positives · eclipsing binaries · Gaia DR3 · radial velocity · brown dwarfs · planet candidate vetting · forecast test