9K Network / Methodology
The Glass Box
9K Network publishes its methodology. Every intelligence report includes a source list, a methods log, and a confidence rating. If we are wrong, we say we are wrong. We do not inflate confidence. We do not hide our process.
This page is the public contract behind that promise. Intelligence is judged by decision-grade utility, every claim is traceable to a documented source tier, and every prediction is testable — misses are reported, not buried.
Three Pillars
Field
Physical intelligence — surveillance, reconnaissance, on-site verification. The world that screens can’t see.
Net
Digital intelligence — OSINT, network reconnaissance, data correlation, pattern analysis. The world that hands can’t touch.
Broker
Intelligence delivery — structured packages, encrypted delivery, destruction after confirmation. The world that doesn’t stay.
No-Storage Doctrine
Intelligence is manufactured on demand, delivered, and destroyed. 9K Network retains no intelligence inventory. No stored packages. No archived reports on private servers. No databases of collected data. What is delivered to the client is the only copy. What is published on this site is the only record.
Source Reliability Scoring
Every source is rated:
- ›Confirmed — verified across multiple independent sources.
- ›Probable — single reliable source, not yet cross-verified.
- ›Unverified — single source of unknown reliability.
- ›Deprecated — source has produced errors in previous reports. Downgraded or eliminated.
JM/OS — Behavioral Modeling Engine
The Judgment/Subject operating layer models any organization, government, or individual as a set of decision engrams — structured behavioral units that capture what a subject is likely to decide next and why. Each engram is composed of:
Subject
The organization, institution, or actor under analysis.
Decision Pattern
The recurring choice structure the subject tends to follow.
Predicted Response
The forecast decision the subject is expected to make under stress or signal.
Sources
The public-record sources (tiered below) grounding the prediction.
Status
PENDING, CONFIRMED, MISSED, or RETIRED — the test outcome over time.
Accuracy
The engram-level confidence score, refined after each post-mortem.
JM/OS Framework Flow
Judicator — Cross-Referencing Engine
Every engram is validated through the Judicator: a cross-referencing engine that weighs an observation across seven analytical layers before it is admitted into a report.
Sourcing Protocol
Tier A · 3.0× — Foundation
Official government records, court filings, and regulatory documents. Highest source weight in the Judicator.
Tier B · 2.0× — Corroborated
News organizations and academic publications. Mid-weight, deduplicated against the public record.
Tier C · 1.0× — Supplementary
Supplementary public sources that contextualize the record.
Tier D — Direct Observation
Direct observation (verified) — if applicable.
Excluded
Classified documents · Leaked private data · Surveillance data · Hacktively obtained material— categorically excluded from 9K Network intelligence. The organization works entirely from the open record.
Judicator Score
Score = Σ(weight × verified) / (1 + redundancy_penalty)
Each observation is weighted by its source tier (A 3.0×, B 2.0×, C 1.0×), confirmed across the seven analytical layers, and penalized for redundant corroboration.
Publication Threshold
3.0 — no exceptions
No engram is published below a Judicator score of 3.0 regardless of pressure, urgency, or salience.
Confidence Rating System
Every report carries a confidence rating from 0% to 100%.
- ›90–100% — High confidence Multiple confirmed sources. Cross-referenced. Field-verified where applicable.
- ›75–89% — Moderate confidence Multiple probable sources. Partially cross-referenced.
- ›60–74% — Low confidence Limited sources. Not cross-referenced. Intelligence presented as preliminary.
- ›Below 60% — Insufficient confidence Intelligence not published. Additional collection required.
Corrections Protocol
When 9K Network publishes an error, we document:
- ›The original claim and the report it appeared in.
- ›The cause of the error.
- ›The correction.
- ›The updated confidence rating.
- ›The systemic change to prevent recurrence.
Corrections are not weaknesses. Corrections are credibility engineering. A methodology that never corrects itself is a methodology that can't be trusted. The corrections log is published alongside the work it corrects.
Post-Mortem Process
When a prediction misses, a 200–300 word post-mortem is written and published. Each post-mortem documents:
- ›What was predicted
- ›What actually happened
- ›What data was missing
- ›What assumption failed
- ›How the methodology was refined
Core Principles of Execution Intelligence
Signal over noise
Intelligence is measured by decision-grade utility, never by volume of output.
Source-tier discipline
Every claim is traceable to a documented tier; classified and leaked material is categorically excluded.
Decision latency is the metric
The speed and freedom with which an institution converts signal into action determines its fate.
Predictions are testable
Every projection carries a status and an expiry date — and is scored when the future arrives.
Failure repeats in archetypes
The same structural patterns recur across domains; the databank exists to name them.
Post-mortem accountability
The methodology evolves only through documented, published misses.
Execution Intelligence Doctrine
The full body of doctrine, databank models, almanac case studies, and index patterns lives in the Research Repository.
Key Metrics
DLI — Decision Latency Index
Days to institutional response. Measures how fast an organization or government converts signal into action — lower latency means more responsive; higher latency means more inertia-bound.
EI — Execution Intelligence
Predicted outcome accuracy. The realized accuracy of 9K Network forecasts, scored when each projection’s expiry date arrives.
Glass Box Principle
“All sources linked. All methodology documented. All corrections published.”