Independent sector errors give nominal 95% intervals that include zero depth and width slopes in seven TOI-1408 b fits to 2019–2024 TESS data

We test linear changes in the transit depth and width of the grazing planet TOI-1408 b. We use 88 transits from 18 public two-minute TESS sectors from 2019 to 2024. Each transit has a free time, depth, width, and local baseline. The width is the full width at half central depth of an empirical profile. We fit straight lines to sector means with equal weight. The primary depth slope is -14.0 ppm/year. Its nominal 95% HC3 interval is -33.8 to +5.7 ppm/year. The width slope is +0.37 minutes/year. Its interval is -0.18 to +0.93 minutes/year. All seven tested settings give HC3 intervals that include zero for both slopes. These approximate intervals assume independent sector errors with zero mean. The result depends on the error method. A seven-group cluster calculation excludes zero for depth. A separate fit to the group means includes zero. Added signals give accurate recovery of depth changes. The width response is 7.3% to 7.9% too large. A broad overlap test adds 250 ppm, three-hour boxes with independent event positions. The constructed samples give slope shifts near +/-75 ppm/year and +/-1.46 minutes/year. These boxes do not follow a planet orbit. The measured profile intervals alone therefore do not set limits on physical orbital change. The data cover the same 18 sectors as Baluev (2026). The new products are the event measurements, coverage ledger, error comparisons, and reproducible analysis. The package contains all source FITS files, hashes, code, results, and figures.

TOI-1408 b · TESS · grazing transits · transit depth · transit width · sector errors · signal recovery · public data