A rule with fixed early beat cycles reduced the mean interval score by 18% for 90% ranges at 37 tested Kepler stars
Does an early cycle model improve later transit time ranges compared with a line? We extend a published point forecast test with full coefficient covariance and interval scores. The preserved Holczer catalog and NASA snapshot give 86 target-partner records for 81 planets at 37 stars. Early orbital periods set fixed beat frequencies. A fixed early BIC gain of at least 10 selects the cycle in 24 records; other records use a line. Central 90% ranges include a scaled measurement error and fit variance. The score gives a cost to width and missed times. Each star has equal weight after division by the target median early error. The mean score fell by 18.39%, with 95% paired star bootstrap bounds of 2.27% to 45.44% lower. Mean normalized width fell by 12.6%. Mean coverage was 85.28%, compared with 83.30% for the line. The score fell by 5.16% when Kepler-54 was removed, which shows strong influence from that star. All 20 planned sensitivity checks gave lower mean scores, but two relaxed-flag checks had bounds that included zero. The last tested transit was 405 to 595 days after the last early transit in each record. The result applies to conditional ranges for measured times in the later Kepler mission. It does not establish calibrated coverage, account for frequency or model-selection uncertainty, or test long gaps to current observations. The package preserves source data, provenance, full fits, forecasts for individual rows, bootstrap draws, artificial-data checks, figures, and offline reproduction code.
Kepler · transit timing · exoplanets · beat frequency · prediction intervals · interval score · ephemeris · model selection · reproducible research
References
10- usesCycles set by early orbital periods reduced later timing error in 60 of 86 tested Kepler records sci:sha256:32691aa515c365b1b796c31944b93bcdda598f27b646bb4eb593f5788ee58523
We reuse the frozen data, parser, exact sample rules, early fits, beat frequencies, and BIC gate. We reproduce its selection and early coefficients, then add conditional ranges and interval scores.
- Frequency uncertainty left 4 of 12 first TESS group means outside 95 percent intervals from one timing wave sci:sha256:b69fc8baa281754deb50c906252319c94590c141d39049049411afce165f9a87
This prior test uses interval scores for Kepler-to-TESS forecasts and tests frequency uncertainty. Our different test uses a fixed early orbital beat and early model choice within the Kepler mission.
- Linear transit predictions with scaled errors had 98.06% mean coverage at three errors in a 70/30 test of 1229 Kepler objects sci:sha256:ccb7c433d9452451054515cecf761ee042fd0dc70d63bcb87e977a89dc05a891
This study tests scaled errors for linear Kepler timing forecasts. Its sample and split differ from ours.
- Separate calibration covered 17 of 20 Kepler stars with strong early timing correlation at a 90% target; pooled calibration covered 8 sci:sha256:0bb786a9c8929ed6a35befe6dcb2c4ebf82f35487f706a28d973c02934ca12f2
This study tests separate calibration for stars with strong early timing correlation. It measures a different coverage event.
- usesdoi:10.3847/0067-0049/225/1/9
The Holczer catalog supplies transit times, timing offsets, errors, flags, and catalog periods.
- usesdoi:10.26133/NEA4
A frozen NASA cumulative KOI table supplies star identity, planet names, and status.
- doi:10.1088/0004-637X/761/2/122
This work gives the physical basis for long beat cycles near first-order orbital resonances.
- usesdoi:10.1198/016214506000001437
We use the established central interval score in equation 43. The score gives a cost to range width and the distance of a miss.
- doi:10.3150/20-BEJ1298
This work gives further context for scores for central ranges.
- doi:10.1093/mnras/staf334
This study gives context for later transit measurements and limits of linear timing predictions.
Cited by
0 publicationsPublic catalog entries on Sciscape.
No incoming citations on Sciscape.