Under one reference profile, fixed pixel sums give a larger HAT-P-70 b half-depth width in TESS Sector 98 than in Sector 32

We test a reported HAT-P-70 b transit width increase with new sums of public TESS target pixels. The main rule uses a fixed circular aperture of radius 2.5 pixels and a local residual background estimate. The background pixels are selected from images outside transit. The extraction tests use common samples and weights within each sector. Each event has its own fitted center, depth scale, and linear baseline. A common asymmetric reference gives half-depth widths of 182.40 minutes in Sector 32 and 187.38 minutes in Sector 98. These visits cover 2020 and 2025–2026. The increase is 4.97 minutes, or 2.73%. We resample nine and twelve whole transits and refit the reference in 400 trials. The conditional 95% interval for the increase is 2.99 to 6.73 minutes. All tested aperture and background choices give positive differences. Omitting the local subtraction gives 4.40 minutes. Adding the pipeline background model back before local subtraction gives an exact duplicate. The stored model has no spatial variation in each retained stamp. It is not an independent check. Noiseless input differences of 0 and plus or minus 5 minutes are recovered to within 0.0028 minute. The result retains the earlier light-curve signal under fixed pixel sums. It shares TESS pixel calibration and does not measure errors common to a sector. A half-depth width is not a contact duration. The study does not establish orbital precession. The package preserves exact selected pixels, source hashes, comparison light curves, code, all resample draws, results, and figures. A fresh run gives identical bytes for 13 result files.

HAT-P-70 b · TESS · target pixel files · transit photometry · half-depth width · aperture photometry · background subtraction · reproducible research