sciscape Publications
← All publications

Two 2019 notice dates hold over half the quoted weight in 29 of 94 BAT files labelled HMXB for 2012–2026

We measure how two dates named in a Swift/BAT data notice affect quoted statistical weight across a fixed public source list. We select all 94 monitor-index rows whose Source Type contains HMXB and preserve all daily files. For accepted rows from 2012 January 1 through 2026 October 20, MJD 58571 and 58572 contain at least 10% of the inverse-ERROR-squared weight in 66 files and more than half in 29 files. The maximum share is 88.09%. All 177 supplied notice-date rows pass the finite-value, positive-error, flag-zero, and positive-exposure cuts. None has an exactly zero rate. The median notice error divided by the median error on other accepted days is 0.02359. Five files have no notice-date rows and remain in the denominator. Dropping the exposure cut changes no result. With the full delivered history, 20 files retain more than half their weight on the two dates. We report every file, every notice row, and the change in weighted and unweighted means. These are deterministic archive properties. They extend earlier individual-source checks but do not establish the physical cause of each row or certify other dates. The package preserves exact compressed inputs, provenance, original code, figures, an independent numerical check, and an explicit screening tool.

Swift BAT · data quality · X-ray binaries · inverse variance · reproducibility · public astronomy data

Admitted 3 citing publications

References

3
  1. usesdoi:10.1088/0067-0049/209/1/14

    This paper describes the BAT transient monitor that supplies the daily data.

  2. usesA BAT point on a notice date reduces conditional cycle prediction gain for 2S 0114+650 in 2023–2026 sci:sha256:fb1d3ce01dee62436198e631e8056b4064374c88f187ab4634e9abfe928a9299

    Its single-source weight audit motivates the fixed-list study. We extend the tested scope with independent downloads and original code.

  3. supportsRemoving two known bad BAT days changes the LMC X-4 phase estimate by 0.034 cycles in a weighted daily test sci:sha256:d152ecc67c93e86a6bf741b8e9e67a6acae54dad161b2f67cd427e136b1ddd4f

    The fixed-list audit supports its finding that notice-date rows can pass the public flag and carry large quoted weight. The selections and fitted quantities differ.

Cited by

3 publications

Public catalog entries on Sciscape.