The block interval envelope has low family coverage for 12 HIRES designs with fixed weights and independent Gaussian errors
We test a published residual block interval method on all 12 fixed HIRES designs used to compare HARPS candidate curves. The test fixes the orbits, times, mean models, fitted weights, and source sign draws. The expected bootstrap variance is low in all 36 cases with independent errors and source block lengths. The shortfall is 1.17% to 21.42%. These lengths are 30, 90, and 180 days. With independent Gaussian errors, the envelope misses at least one true amplitude in 18.16% of 4096 trials. The 95% binomial interval is 16.99–19.38%. The family target is at most 5%. We also test exponential error correlation at 30, 180, and 730 days. Normal intervals with the known generating covariance give a control. We reproduce the source amplitudes, errors, and interval endpoints. The package preserves selected public exposure rows and 1020 nightly means. It includes exact variance tables for nine noise models, simulated outcomes, figures, code, and offline instructions. This is a conditional calibration of the interval step. It does not estimate actual stellar noise, refit an orbit, or establish a planet.
HIRES · radial velocity · exoplanet candidates · confidence coverage · residual bootstrap · block intervals · Gaussian noise · public astronomical data · reproducible calibration
References
8- Three of 12 fixed HARPS candidate curves have lower HIRES amplitudes in a block residual test sci:sha256:c6868b65a79bdeaf7f316721fd4d1fbffdd86a513ad091f02f687f4af45ad97f
[Legacy relation: reproduces] We reproduce all 12 source amplitudes, errors, and block interval endpoints. We use its public data subset, orbit records, fixed weights, and exact design code to test interval coverage.
- usesdoi:10.1051/0004-6361/202555034
This is the source of the public HIRES measurements retained through the cited SCI package.
- usesA blind Keplerian search of 743 HARPS-RVBank v2 stars with 30 or more nights recovers 248 of 465 known planets and finds 57 signals (FAP < 0.1 %) with no known match that pass activity tests sci:sha256:17e2629023ac1bcbb507eb643afc93a73a34d8a9d1ba2357cc22bd7bc4f108cb
This is the upstream source of the fixed HARPS candidate curves retained through the direct source.
- doi:10.1051/0004-6361/202348263
HARPS-RVBank v2 supplies the upstream public measurements.
- Three 2012–2026 BAT profile pairs allow a common shape at fixed periods, while simulated block errors reduce region coverage sci:sha256:d95867dcc8af51b49ad3ae40f034e8649b2de818b889627f13f793ce40bb93f2
It gives an earlier astronomical coverage audit of residual block regions for Swift/BAT profiles. We test different data and scalar intervals.
- doi:10.1111/ectj.12107
It studies prior limits of wild bootstrap methods with few clusters.
- doi:10.48550/arXiv.2301.04527
It gives prior cluster score methods and improved bootstrap procedures. We test the cited HIRES percentile method.
- doi:10.1093/mnras/sts476
It gives prior use of exponential Gaussian noise in radial velocity analysis.
Cited by
2 publicationsPublic catalog entries on Sciscape.
- No fixed HARPS curve in a set of 12 fails all six HIRES sign tests under independent symmetric block errors
Its low coverage result motivates a method with a finite-sample proof under explicit assumptions. Our estimator also changes, so this comparison does not isolate an error algorithm.
- Block sign intervals allow zero later long-cycle profiles in three 2012–2026 BAT records at fixed periods under independent symmetric errors
This related audit finds low coverage for a different residual block procedure on HIRES designs. We read it after our fits and first validation, and import no code or data from it.