Skip to content
Canli Capital

Measurements

Active ownership human gate audit

Active ownership human gate audit is one of 87 measurements this engine publishes in full. It is regenerated from a real run rather than transcribed, and it is shown here with its own claim boundary so the number and its limits arrive together.

The measurement

Schema
canli.alphac-active-ownership-human-gate-audit.v1
Generated at
2026-08-26T03:25:48.497500+00:00
Author
Arhan Canli
Stage
PROSPECTIVE_PRE_LABEL_PRE_RETURN_GATE_AUDIT
Public path
/glassbox/active_ownership_human_gate_audit.json
Content hash
sha256:668e65f17c9e83bb5118db75f3489b4ecfbf5c1bd351f41c07f529673bc88f90

Governance

Labels opened
no
Return data opened
no
Return hypotheses spent
0
Existing point thresholds changed
no
Audit may rescue known outcome
no

Frozen design

Rows
48
Machine predicted positive
8
Machine predicted negative
40
Item4 extracted
45
Item4 unresolved
3
Ownership sole candidate
14
Ownership unresolved by frozen machine rule
34

Years

First
2010
Last
2025
Count
16

Rows per year

2010
3
2011
3
2012
3
2013
3
2014
3
2015
3
2016
3
2017
3
2018
3
2019
3
2020
3
2021
3
2022
3
2023
3
2024
3
2025
3

Point gate reachability

Operationally reachable
yes
Not guaranteed to pass
yes

Precision

Threshold
0.950000
Minimum true positives
8
Maximum false positives
0

Interpretation. With 8 frozen predicted positives, the point gate passes only at 8/8; one false positive fails.

Recall

Threshold
0.800000
Human positive denominator known pre label
no
Maximum human positives that can still pass
10

Interpretation. The denominator is determined only by the independent labels. With eight frozen predicted positives, eleven or more human-positive rows necessarily make the 80% point-recall gate fail.

Ownership exact

Threshold
0.900000
Minimum exact rows
44
Maximum mismatches
4
Interpretation
At least 44/48 exact outcomes are required; 4 mismatches are permitted.

Statistical establishment audit

Confidence level one sided
0.950000
Method
exact Clopper-Pearson binomial lower bound

Best case precision

Successes
8
Trials
8
Point estimate
1
Lower bound
0.687656
Establishes threshold
no

Best case whole packet accuracy

Successes
48
Trials
48
Point estimate
1
Lower bound
0.939497
Establishes 0 95
no

Minimum all success denominators

Precision predicted positives
59
Recall human positives
14
Ownership rows
29

Conclusion. The 48-row packet can support a frozen feasibility point result but cannot, by itself, statistically establish 95% precision at one-sided 95% confidence. Passing point estimates must not be described as confidence-bound validation.

Decision. KEEP_FROZEN_48_ROW_POINT_GATE_FOR_RETURN_FEASIBILITY_AND_REQUIRE_DISJOINT_CONFIRMATORY_ACCURACY_BEFORE_SLEEVE_ADMISSION

Required interpretation

If point gate passes. May proceed only to the already-governed return preregistration stage. Report raw confusion counts, point metrics, and exact confidence bounds.

Before sleeve admission. Freeze a disjoint independent confirmatory corpus sized by the relevant positive denominators and require one-sided 95% exact lower bounds to meet the declared precision, recall, and ownership thresholds.

If point gate fails. The frozen classifier is data-gated. Do not tune it on these labels or open returns for this identity.

Source bindings

Generator

Path
scripts/audit_active_ownership_human_gate.py
Sha256
4599222242309956d625f4e934bc968c42e092367dedeae1915c350bd605f0cb

Protocol

Path
docs/design/FEASIBILITY_ACTIVE_OWNERSHIP_13D_ITEM4_V3.md
Sha256
017eac943055deaff83261e6b67e853d78e687d4790a6d78c4a05ecbdef0d988

Gate audit protocol

Path
docs/design/ACTIVE_OWNERSHIP_HUMAN_GATE_AUDIT.md
Sha256
bb75b1ab2f47dbeb2a3f317537665399a06217a9dafd43ebb6d6f600d2cfa32e

Frozen labels

Path
artifacts/feasibility/active_ownership_13d_item4_v3/frozen_human_labels.csv
Sha256
fefb4bdb5a1ad05e691e6a76b5dd98b08066dab88ede3054893342b93a46a137

Document audit

Path
artifacts/feasibility/active_ownership_13d_item4_v3/document_audit.parquet
Sha256
f8b2843b194d2b5d9825963af40bdc1e5e732a9e247e957ca828a3c961fc174c

Feasibility result

Path
artifacts/feasibility/active_ownership_13d_item4_v3/result.json
Sha256
6443ae40f358330bf9ca4512d09aebc834b962385de703483a2c2f2e795901c5

Blind packet manifest

Path
artifacts/labeling/active_ownership_13d_item4_v3_blind/manifest.json
Sha256
6e3e4480f7c248d6d7d42e3ab03a05e6fb20450e46313514f4e976b152bc3057

Check it yourself

Every figure above is read from /glassbox/active_ownership_human_gate_audit.json, the artifact the engine wrote. Nothing on this page is typed by hand: the page is generated from that file, so a figure that moves in the artifact moves here and a figure that is not in the artifact cannot appear here at all.

All 87 measurements · The glass box · The research library