DASCH DR7 matched-comparison slope test of KIC 8462852, margin 1.0 mag
Sensitivity runs with margin 0.2 and 0.5 (results_margin0.2, results_margin0.5 in the outputs archive, sha256 96408452...) give target slope 0.147 +/- 0.019, n 1265, matched z 1.85, percentile 0.98, differential 0.125 +/- 0.019 (z 1.87), gap z 1.83, pre-1953 0.023 +/- 0.039: the conclusion does not depend on the cut. The two looser margins give identical samples because the CLOSE_TO_LIMITING flag already removes points within about 0.5 mag of the limit. Independent re-check (same day, separate agent, own code, no access to analyze_dasch.py): target WLS slope +0.170 +/- 0.016 (n 969 without the BFLAGS cut; the 895 here applies the BFLAGS bits listed in the method), Theil-Sen +0.194 [0.159, 0.229], pre-1953 slope +0.017 +/- 0.036; slopes.csv matched mean 0.034, sd 0.068, n 49, one star above the target; gap offset target +0.093, matched +0.012 +/- 0.051 with no matched star above the target. Two caveats it raised and this record adopts: (1) 15 of the 64 matched stars are absent from slopes.csv, 13 because they have no detections with clean AFLAGS at all and 2 dropped by the 30-arcsec neighbour screen; (2) without the neighbour screen the matched-star slope scatter is 0.095 rather than 0.068, so the z of about 2 depends on the screen while the rank (first or second of about 50) does not. The formal slope error (0.02) is 3-5 times smaller than the matched-star scatter and should not be used on its own.
Result
| Run | Measured | Result | Against | Decision | Date |
|---|---|---|---|---|---|
| Run | Target slope (mag per century) and its percentile and z among 49 matched comparison stars; plate-differential slope and its matched null; post-1969 minus 1930-1953 offset | {'tight_z': 2.38, 'target_n': 895, 'target_z': 1.96, 'matched_n': 49, 'matched_sd': 0.068, 'pre1953_err': 0.046, 'gap_offset_z': 2.02, 'matched_mean': 0.034, 'target_theil': 0.19, 'pre1953_slope': 0.019, 'differential_z': 2.04, 'target_wls_err': 0.02, 'tight_matched_n': 18, 'differential_err': 0.021, 'target_bin_slope': 0.208, 'target_wls_slope': 0.166, 'gap_offset_target': 0.088, 'target_percentile': 0.98, 'bootstrap_raw_ci95': [0.131, 0.201], 'differential_slope': 0.133, 'differential_null_sd': 0.069, 'gap_offset_matched_sd': 0.034, 'differential_null_mean': -0.008, 'gap_offset_matched_mean': 0.02, 'bic_linear_minus_step_target': -6.9} | hippke2016 0.12-0.14 +/- 0.02 with systematics floor 0.15-0.2; break at 1962 · schaefer2016 0.164 +/- 0.013 (check-star rms 0.043, 4 sigma) · matched_null_this_run mean 0.034, sd 0.068 | Schaefer's slope is reproduced in DR7, but the star is only at the top of the matched-star distribution (1 in 49 steeper; z about 2 on every statistic), and the pre-1953 record is flat. Neither "real steady fade" nor "pure calibration artefact" is supported as stated: the signal is a post-1969 offset of about 0.07 mag beyond what matched stars show, at roughly 2 sigma. Inconclusive at the suggestive tier. Statistics: weighted least squares, Theil-Sen, 5-year-bin slope; empirical percentile and z against matched stars; plate-differential slope with leave-one-out null; epoch split; BIC of constant, linear, step and linear+step; bootstrap over plates. |
What it used and found
Parameters
t0 = 1890, t1 = 1990, bin = 5, seed = 7, nboot = 2000, margin = 1, rms_max = 0.35, min_points = 300, neighbour_dmag = 2.5, neighbour_screen_arcsec = 30
Cite it
Mysteries Department. (2026). DASCH DR7 matched-comparison slope test of KIC 8462852, margin 1.0 mag [Run RUN0DDYGPP634AA]. https://mysteriesdept.com/r/RUN0DDYGPP634AA
@misc{mysteriesdept_run0ddygpp634aa,
title = {DASCH DR7 matched-comparison slope test of KIC 8462852, margin 1.0 mag},
author = {{Mysteries Department}},
year = {2026},
note = {Run, catalog key RUN0DDYGPP634AA},
url = {https://mysteriesdept.com/r/RUN0DDYGPP634AA}
}