vacuity threshold result

Definition	The value of a certificate's statistic below which the certificate is uninformative, derived from the problem rather than tuned to the data. Chapter 9. Also vacuity.
Example	A certificate statistic at the maximum certifiable value, 1.83, is vacuous, and the derived threshold says so before the data is read.
Book	Data Mining as Observation, draft 0.2, commit `f3914f0`; entry id `vacuity-threshold`, kind result.
Status	measures [demonstrated]. Corrections: none recorded.
Defining equation	Book equation 9.3.
Assumptions and scope	The threshold is derived from the problem, the expected maximum of the distractors’ margins, and is not tuned to the data. A certificate at or below the threshold proves nothing about the object. It has not shown that no structure exists, only that this test could not certify one at this resolution. Tested under seal on retrieval, which is recognition with one candidate per query, and the transition has a finite width.
Prior art	none recorded
Evidence	`geometric-observation/claims/LEDGER.md:65`, `geometric-observation/experiments/GO3-certificate-vacuity-v3-NOTES.md:1-60`, `geometric-observation/claims/LEDGER.md`, `lean/DataMiningAsObservation/Vacuity.lean`
Reviewed	not yet reviewed; generated 2026-09-10 from records at the commits on the provenance page.

The statistic below which the certificate says nothing.

Equation

Book equation 9.3.

\[\hat\mu=\frac{\bar s_{\mathrm{true}}-\bar s_{\mathrm{distr}}}{\sigma_{\mathrm{distr}}},\qquad \mu_{\mathrm{crit}}=\mathbb E\Big[\max_{N-1}\mathcal N(0,1)\Big],\qquad \rho=\frac{\hat\mu}{\mu_{\mathrm{crit}}},\qquad \rho=1\ \text{vacuous}.\]

Conditions

The threshold is derived from the problem, the expected maximum of the distractors’ margins, and is not tuned to the data.
A certificate at or below the threshold proves nothing about the object. It has not shown that no structure exists, only that this test could not certify one at this resolution.
Tested under seal on retrieval, which is recognition with one candidate per query, and the transition has a finite width.

Conditions are curated in entries.toml rather than read from a record.

Ledger

measures. GO-3 [demonstrated]. The certificate’s vacuity threshold predicts where single-stage retrieval dies. geometric-observation/claims/LEDGER.md:65.

First stated

Volume 14, the GO-3 registration and its notes, geometric-observation/experiments/GO3-certificate-vacuity-v3-NOTES.md:1-60, DOI 10.5281/zenodo.21776291, and chapter 9 section 9.3 of Data Mining as Observation.

Measurements

Where the book states it	Numbers, as the book’s sources table records them	Source
chapter 9 section 9.3	the margin certificate, mu crit as the expected maximum of N minus 1 standard normals, rho, death at 0.948 within 6 percent, Spearman 0.991 vs 0.873, fourteen corpora, six gates, the v1 to v3 path, the standing correction	`geometric-observation/experiments/GO3-certificate-vacuity-v3-NOTES.md:1-60`; `geometric-observation/claims/LEDGER.md` row GO-3

Failures and corrections

none

Invariance envelope

none declared

Machine checked

lean/DataMiningAsObservation/Vacuity.lean, theorems maxOver_mono, rho_eq_one_iff, rho_lt_one_of_lt, at observation-data-mining f3914f0; what the check covers is stated in the book’s appendix C.

Used in

Data Mining as Observation chapters 8, 9.

certificate; rank certificate; recognizer.

Status

Generated 2026-09-10 by encyclopedia/generate.py; book at observation-data-mining f3914f0; the commit of every record is listed in the encyclopedia’s provenance.

← unit vectorvalidity index →