Symmetric infrared fits place the March 2026 WASP-12 b transit 91 to 118 seconds before a 2020–2023 TESS constant-period forecast

We test the 10 March 2026 WASP-12 b JWST transit against clocks fit only to 117 TESS times from 2020 to 2023. We use public NIRSpec PRISM data from program 5924. The original broad-band fit fails quality checks. We then use the fixed 3.8–5.0 micrometre band as an exploratory choice. A symmetric transit with a quadratic time baseline gives BJD TDB 2461109.87605008. Residual-block refits give a conditional error of 12.9 s. Eleven tested long-wavelength fits span 26.5 s. All put the center 91.0 to 117.5 s before the constant-period forecast. This range measures model sensitivity, not confidence. The main center is 114.8 s early, with an 18.5 s comparison error under independent TESS timing errors. A fixed published shrinking-period clock reduces the discrepancy to 41.2 s. Assumed shared sector offsets of 30 s increase the comparison error to 42.7 s. The first delivered integration midpoint precedes the exposure header by 4.575 s. A separate segment check confirms the delivered time table but does not resolve that discrepancy. The adopted clock floor does not cover all possible product timestamp errors. We publish reduced data, source code, all fit checks, and a raw-data reproduction. The result supplies a new event test of specified clocks. It does not establish a new decay rate or a physical cause.

WASP-12 b · JWST · TESS · transit timing · orbital decay · NIRSpec PRISM · model sensitivity · public data · reproducible research