Pointing terms reduced position noise in linear pixel fits from 14 TESS sectors of TOI-2095

We test pixel source positions for TOI-2095.03 and the known planets b and c in 14 public TESS sectors. We keep catalog times and transit shapes fixed. A linear time baseline gives excess position noise in nearby time controls. Adding the recorded column and row position terms reduces this noise. For the candidate, RMS values of two whitened spatial coefficients fall from 14.40 and 8.44 to 1.21 and 1.35 in 36 control groups. Known planets b and c show the same reduction. The candidate offset changes from (11.88, -0.12) to (1.31, 0.31) arcseconds in east and north relative to an empirical reference image. Its formal errors are (1.74, 1.75) arcseconds. Control-based error scales are (2.17, 2.36) arcseconds. Alternative baseline and pixel-radius fits retain small offsets. Synthetic image shifts show bias in the first-order spatial model. Residual image noise also exceeds the formal model. These tests support a reduction of false position signals, without a precise sky position, a nearby-source exclusion radius, or planet validation. The package contains 45,270 mean pixel images from 225,218 accepted exposures, source hashes, all event decisions, fitted images, controls, figures, code, and repeat instructions. We report source FITS checksum mismatches and preserve the exact archived input hashes.

TESS · TOI-2095 · TOI-2095.03 · target pixel files · difference images · pointing correction · source position · time controls · reproducible research