Skip to content

About

Deterministic, training-free engine that reads where a multi-sensor window sits relative to the system's own history. Returns closeness and agreement as a trust signal, not a forecast.

Topics

Resources

Stars

3 stars

Watchers

0 watching

Forks

Latest commit

 

History

11 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 

Repository files navigation

State Shape: PREREG_12 sealed result

Sealed: 9 Oct 2026, 3:09 PM MT, before any held-out data was scored. Manifest SHA-256: cdfb47280069b86ad4714aec5cf042e1dcda620f014d87c3605b5c292fcfbd1b

48 MIT-BIH ECG records, 480 held-out slices, scored against per-record skill.

Result Value
Closeness (d_rel) vs per-record skill, rank correlation ρ −0.75
Agreement (purity) vs per-record skill, rank correlation ρ +0.72
Held-out slices 480 across 48 records

All pre-registered outcomes are reported. Both trust signals passed. A third test failed: combining two leads was not better than the best single lead. A fourth, expected to show no gain in prediction, showed none. State Shape gives a trust signal, not a forecast. Validated on ECG only so far.

Data: MIT-BIH Arrhythmia Database (PhysioNet).

About

Deterministic, training-free engine that reads where a multi-sensor window sits relative to the system's own history. Returns closeness and agreement as a trust signal, not a forecast.

Topics

Resources

Stars

3 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors