A 10% shared dip rule marked no peaks in 96 test stars in TESS Sector 70, Camera 1, CCD 3, with QUALITY = 0

We tested whether shared dip times in separate training stars select large dip contrasts in corrected TESS light curves. We fixed the rules before reading target flux. We selected 192 stars in Sector 70, Camera 1, CCD 3, with magnitudes from 9 to 12 and separations of at least 600 arcsec. We used 96 stars for training and 96 for testing. We kept QUALITY = 0 cadences and formed two-hour contrasts in ten-minute bins. Scores use each star's median absolute deviation. The primary rule required at least 80 valid training stars and at least 10% with scores of 3 or more. It made an empty mask: the largest eligible training fraction was 7.5%. There were 44 test peaks with scores of 5 or more. A planned 5% rule selected six peaks and covered 792 of 249,297 usable test-star centers, or 0.318%. The six peaks lie near a boundary of usable score coverage. A stricter coverage check also selects six peaks, with five stars common to both sets. We report all 27 planned threshold combinations, quality and coverage checks, SAP comparisons, and 36,864 separate added-signal trials at known centers. An additional check applies all nine masks to those same trials. Shared event checks are prior methods. The new contribution is this reproducible measurement and its complete data and code package. The ratios are descriptive. The results do not classify individual events, establish a new planet, or measure blind transit-search completeness.

TESS · Sector 70 · light curves · transit vetting · shared event times · threshold sensitivity · negative result · reproducible research