Skip to content

A data-driven audit of the 'Geopolitical Thermostat,' documenting how timed information disclosure regulates public attention to enable structural shifts in policy and capital flows.

Notifications You must be signed in to change notification settings

Leerrooy95/The_Regulated_Friction_Project

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

338 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

The Regulated Friction Project v8.5

A data-driven analysis of temporal correlations between friction events, policy shifts, and capital flows (2015-2026).

New to this project? See Glossary.md

In a rush? See consolidation_pattern_significance.md


Table of Contents


Quick Summary

This repository documents statistically significant correlations between high-visibility "friction" events (file releases, scandals, media cycles) and institutional compliance events (policy shifts, financial positioning, regulatory changes).

Core Finding: Friction events (document releases, scandals) predict institutional compliance events (policy shifts, financial moves) at a 2-week lag (r = +0.6196, p = 0.0004, n = 30 weeks).

What this means: When high-visibility friction events spike, institutional compliance events follow ~14 days later — suggesting shared exploitation of calendar-driven attention windows.


Understanding the Statistics

This research uses Pearson correlation (r) to measure relationships between event types.

r Value Interpretation
0.0 No relationship
±0.1-0.3 Weak
±0.3-0.5 Moderate
±0.5-0.7 Strong
±0.7-1.0 Very strong

Our finding (r = +0.6196): When friction events spike, compliance events follow approximately 2 weeks later. This correlation (p = 0.0004) has less than 0.05% probability of occurring by chance.

Why this matters:

  • A correlation this strong is reproducible — run the scripts in Run_Correlations_Yourself/ yourself
  • It suggests a structural pattern, not random coincidence
  • Separate analyses (Project Trident, cross-validation) confirm the pattern using independent datasets

Important caveat: Correlation ≠ causation. We document that these events cluster together, not that one causes the other. The claim is structural: the pattern exists and is statistically significant.


Key Statistics

Finding Value Status
Friction → Compliance correlation r = +0.6196 (2-week lag) ✅ Verified
Statistical significance p = 0.0004, n = 30 weeks ✅ Verified
Ritual → Policy proximity 50.7% vs. 19.9% baseline (2.5x) ✅ Verified
Project Trident significance p = 0.002 (Mann-Whitney U) ✅ Verified
Cross-validation (14-day periodicity) χ² = 330.62 (p < 0.0001, 2,102 events) ✅ Verified
December 2025 cluster 108 events in 12-day window ✅ Verified
Dec 22 signal types 5 (Friction, Geopolitics, Financial, Policy, Cyber) ✅ Verified
Event colocation Friction dates attract 20–42x more compliance than random dates ✅ Verified
January 2026 signal peaks 3 peaks (Jan 3-9, Jan 20-22, Jan 27-31), 1 trough (Jan 10-16) ✅ Verified
January 2026 event density 34 events: 12 friction, 19 compliance, 3 anchors ✅ Verified
Feb 1–19 compliance window 9 compliance events to 6 friction events in 19 days ✅ Verified

The Convergence Model

Original Hypothesis (Sequence)

Friction (t) → [creates window] → Compliance (t+14 days)

Revised Finding (Convergence)

Calendar Anchor (solstice, holiday, fiscal deadline)
        ↓
┌───────┼───────┐
↓       ↓       ↓
Friction  Policy  Financial
        ↓
Lagged Clustering (r = 0.6196, 2-week lag)

The raw data shows friction and compliance events cluster together rather than following a sequential cause-effect pattern. This makes the phenomenon more robust—it doesn't require coordination, just shared exploitation of the same calendar signals.

December 19-23, 2025: The Pincer Window

The Dec 19-23, 2025 window demonstrates convergent clustering:

Date Event Type Events
Dec 19 Friction Epstein Library release (DOJ)
Dec 19 Financial Bull & Bear sell signal (8.5)
Dec 22 Geopolitics China EU dairy tariffs (42.7%)
Dec 22 Financial BlackRock Bitcoin ETF "top theme"
Dec 22 Cyber/Intel CRINK nation-state threat analysis
Dec 22-23 Peak 6-8 friction + 9-13 compliance events/day
Dec 23 Infrastructure Redaction failures exposed (NYT)

Five independent signal types converged on the same window—not because one caused another, but because all actors respond to the same environmental signals.


The Three-Layer Model

This repository is part of a three-layer analytical framework:

Layer Repository Focus Primary Finding
1. Attention Capture This repo Friction → compliance clustering r = 0.6196 (2-week lag)
2. Vacuum Creation DOGE_Global_Effects Aid cuts → instability r = 0.42-0.69, 3-12 month lag
3. Alternative Capture BRICS-NDB-LocalCurrency-DiD Alternative financial systems +25.5 pp local currency lending

Unified Thesis: Domestic chaos consumes attention while foreign policy vacuums emerge, which alternative systems then capture.


Repository Structure

The_Regulated_Friction_Project/
│
├── README.md                    # This file
├── Report.md                    # Executive summary and findings report
├── CITATION.cff                 # Citation metadata
├── Glossary.md                  # Key terminology definitions
├── SOURCES.md                   # 138 unique sources catalogued across all datasets
├── resistance_indicators.md     # Resistance indicator tracking
│
├── Core Analysis
│   ├── Repository_Synthesis.md                 # Three-layer framework overview
│   ├── Thermostat_Explained.md                 # Why the mechanism exists
│   ├── Claude's_Analysis.md                    # AI-assisted interpretation
│   ├── Grok_Analysis.md                        # Cross-verification with Grok
│   ├── China_State_Media_Null_and_Meanings.md  # Null finding: China state media shows no anticipatory signaling
│   └── Case_Study_David_Barnes_Detention.md    # Hostage diplomacy as human leverage dimension
│
├── December 2025 Focus
│   ├── CRUCIAL_Synthesis_Dec19_Convergence.md  # Dec 19-23 pincer window analysis
│   ├── FINANCIAL_RECEIPT_VERIFICATION.md       # Financial event verification
│   └── Main_Characters.md                      # Cabinet timing analysis
│
├── Policy & Implications
│   ├── How_This_Happened-A_Policy_Breif.md     # Regulatory citations, oversight questions
│   ├── Implications.md                         # China BRI expansion implications
│   ├── 'Transparency'_Timeline.md              # Document release history
│   └── Alternate_Mechanisms.md                 # Alternative explanations considered
│
├── Methodology & Transparency
│   ├── TRANSPARENCY_NOTE_FOR_2026_ANALYSIS.md  # Dataset inclusion/exclusion criteria
│   └── VERIFICATION_REPORT_Jan2026.md          # Complete independent statistical verification
│
├── 00_Quick_Breakdowns/                        # Executive-level summaries
│   ├── About_Me.md                              # Background on the author
│   └── Copilot_Executive_Synthesis_Feb2026.md  # Comprehensive repository synthesis (Feb 2026)
│
├── 01_Levers_and_Frictions/                    # Friction events timeline
│   ├── Epstein_Files_timeline.csv
│   └── Epstein_Files_timeline_updated.csv
│
├── 02_Anchors_and_Financials/                  # Capital flow data
│   ├── pep_banking_combined.csv
│   └── pep_banking_sentiment.csv
│
├── 03_Master_Framework/                        # Primary datasets
│   ├── MASTER_reflexive_control_2015-2025.csv
│   ├── MASTER_timeline_2015-2025_UPDATED.csv
│   └── updated_master_theory.csv
│
├── 04_Testing_and_Counters/                    # Validation datasets
│   ├── expanded_historical_backtest.csv
│   └── merged_backtest_counters.csv
│
├── 05_Geopolitical_Vectors/                    # Nation-state analysis
│   ├── CRINK_Analysis.md                              # CRINK axis integration
│   ├── Global_Election_Analysis.md                    # Allied election patterns
│   ├── Graham_Venezuela_Analysis.md                   # Graham's 54-day Venezuela escalation
│   ├── Graham_Venezuela_Posts_Timeline.csv            # Supporting timeline data
│   ├── January_2026_Parallel_Operations_Timeline.md   # Venezuela-Yemen parallel operations
│   ├── Venezuela_Privatization_Amnesty_Stack_Feb2026.md  # Venezuela compliance stack
│   └── thermostat_control_data.csv                    # Nation-state linkage data
│
├── 06_Visualizations/                          # Charts and images
│
├── 07_My_Previous_Epstein_Research/            # Pre-project archive (4 PDFs)
│
├── 08_How_It's_Possible/                       # Mechanism documentation
│   ├── 08_How_Its_Possible.md                         # Core mechanism analysis
│   ├── DOJ_Probe_Results.csv
│   ├── Phase_0_Maxwell_Pivot.csv
│   └── pincer_data.csv
│
├── 09_Silicon_Sovereignty/                     # Semiconductor, AI & infrastructure
│   ├── SILICON_SOVEREIGNTY_REPORT.md                  # Core report: compute-as-currency thesis
│   ├── Infrastructure_Consolidation_Pattern_Jan2026.md  # Oracle/MGX/PIF consortium analysis
│   ├── CRUCIAL-Cross_Verification_Check.md            # Cross-verification with llama2
│   ├── Coalition_Narrative_Map_2015-2025.csv          # Media coverage patterns (456 records)
│   ├── REFINED_supercomputer_geopolitics (1).csv      # Supercomputer geopolitics (906 records)
│   ├── Regulatory_Map_Data_CLEANED.csv                # Policy/compliance events (76 records)
│   └── VOCA_funding_timeline_clean.csv                # Victim services funding (667 records)
│
├── 10_Real-Time_Updates_and_Tasks/             # Ongoing analysis and daily tasks
│   ├── README.md
│   ├── 2026_January/                                  # January 2026 daily updates
│   ├── 2026_February/                                 # February 2026 daily updates
│   └── Tasks/                                         # Standing monitoring tasks (14 trackers)
│
├── 11_Protest_Dynamics_and_Funding/            # Protest analysis
│   ├── Protest_Funding_Audit.pdf
│   └── README.md
│
├── 12_The_Media_Firewall/                      # Media coverage patterns
│   ├── 1789_Symbolism_Analysis.md                     # 1789 Capital / Gulf SWF funding network
│   ├── Analyzing Geopolitical and Media Control.pdf
│   └── README.md
│
├── 13_State_and_County_Analysis/               # State-level forensics
│   └── arkansas_infrastructure_forensic_audit.md      # Arkansas Act 373/548, PSC analysis
│
├── Statistical Verification
│   ├── Control_Proof/                                 # Core correlation data
│   │   ├── master_reflexive_correlation_data.csv      # Weekly friction/compliance indices
│   │   ├── MASTER_reflexive_control_v2.csv
│   │   ├── reflexive_control_scraped_data.csv
│   │   ├── thermostat_control_data.csv
│   │   └── correlation_results.txt
│   │
│   ├── Run_Correlations_Yourself/                     # Independent verification suite
│   │   ├── README.md                                  # Folder guide and correlation reference
│   │   ├── requirements.txt                           # Python dependencies (pandas, numpy, scipy, statsmodels)
│   │   ├── run_original_analysis.py                   # Reproduce r = 0.6196, p = 0.0004, Mann-Whitney p = 0.002 (pre-2026 data)
│   │   └── Wrong_Correlations/                        # ⚠️ Archived scripts that used wrong datasets or excluded data
│   │       ├── README.md                              # Explanation of what went wrong
│   │       ├── reproduce_updated_correlation.py       # ⚠️ DEPRECATED (used 2025-only datasets — produced inflated r = 0.6685)
│   │       ├── original_correlation_test.py           # (used relative paths — was correct)
│   │       ├── reproduce_original_correlation.py      # (used hardcoded paths to wrong datasets)
│   │       ├── independent_statistical_verification.py # (used hardcoded paths to wrong datasets)
│   │       ├── run_original_analysis.py               # (was correct — archived copy)
│   │       └── DISCREPANCY_ANALYSIS.md                # Methodology comparison (still valid)
│   │
│   └── Project_Trident/                               # Temporal correlation study
│       ├── PROJECT_TRIDENT_CASE_STUDY.md
│       ├── The_Trident.md                             # Three-prong mechanism
│       ├── DATASET_REFERENCE.md
│       ├── Veriify_Trident_Analysis.py                # Verify ritual timing analysis
│       ├── anchor_events_parsed.csv                   # 70 anchor events
│       ├── project_trident_final_dossier.csv          # 118 dossier entries
│       ├── ritual_events_parsed.csv                   # 51 ritual events
│       ├── temporal_correlations_analyzed.csv          # 338 temporal pairings
│       ├── Best_Data_For_Project_Trident/             # Ritual timing, fund flow datasets (8 files)
│       ├── Claude_Code_Analysis/                      # Q1 2026 Privatized Integration research
│       │   ├── Privatized_Integration_Networks_Q1_2026_Synthesis.md  # Master document
│       │   ├── Phoenix_Settlement_Portfolio_and_New_Gaza.md          # Reference appendix
│       │   ├── LEAD_ANALYST_REVIEW.md                               # Independent verification
│       │   └── README.md
│       └── Copilot_Opus_4.6_Analysis/                 # Lead Researcher statistical verification
│           ├── README.md                              # Transparency notice, methodology, work log
│           ├── Statistical_Tests/                     # 9 runnable Python robustness scripts
│           ├── Findings/                              # Active analysis — provenance, backfill guide
│           ├── Verification_Reports/                  # Prediction tracker
│           ├── Consolidation_Analysis/                # Cross-domain consolidation assessment
│           ├── FaaS_Signal_Analysis/                  # SuperGrok signal verification + January 2026 signal map
│           ├── Influencer_Narrative_Timing/            # Media Firewall narrative timing analysis
│           ├── Archive/                               # Previous analysis kept for transparency
│           └── Datasets/                              # Local copies of original pre-2026 CSVs (23 files)
│
└── New_Data_2026/                              # January-February 2026 datasets
    ├── 2026_Analysis.md                               # Correlation methodology and findings
    ├── Additional_Anchors_Jan2026_Final.csv
    ├── Biopharma.csv
    ├── BlackRock_Timeline_Full_Decade.csv
    ├── CRINK_Intelligence_Dataset_Final_Verified.csv
    ├── High_Growth_Companies_2015_2026.csv
    ├── Infrastructure_Forensics.csv
    └── Timeline_Update_Jan2026_Corrected (1).csv

What's New (v8.5 - February 2026)

January 2026 Signal Analysis

Full-month signal map identifying three friction-compliance peaks and one trough, with 34 verified events. The 2-week lag holds across all major friction-compliance pairs. Signal escalates across the month rather than cycling at steady state. Signal strength is rated 1–10 based on event density, media saturation, friction-compliance temporal proximity, and structural significance.

Peak Dates Signal Key Events
Peak #1 Jan 3–9 9/10 Maduro capture + Saudi Yemen takeover (STC dissolution)
Trough Jan 10–16 4/10 Cooldown; no kinetic friction
Peak #2 Jan 20–22 9/10 Free America Walkout (450+ events, 50 states) + TikTok deal + Board of Peace signed
Peak #3 Jan 27–31 10/10 Epstein files (3.5M pages) + Warsh Fed Chair + Paris exit + government shutdown

See Project_Trident/Copilot_Opus_4.6_Analysis/FaaS_Signal_Analysis/january_2026_signal_analysis.md

Media Firewall Narrative Timing

New analysis mapping influencer narrative pushes against policy compliance events. Key findings:

  • Tucker Carlson's "NATO is dead" narrative (Jan 6–8) precedes TikTok deal + Board of Peace (Jan 22) by 14–16 days — matching the 2-week lag
  • Consistent structural silence on financial architecture (MGX, Silver Lake, Gulf sovereign flows) across Dec 2025–Jan 2026
  • Candace Owens departure (March 2024) functions as boundary marker — anti-Israel is the one narrative the firewall won't tolerate because Israel is structurally necessary to the Vendor-State model
  • Jan 30 case study: Carlson's Epstein coverage directs anger toward intelligence agencies, not financial intermediaries — the Warsh Fed Chair nomination executes under cover of maximum Epstein friction

See Project_Trident/Copilot_Opus_4.6_Analysis/Influencer_Narrative_Timing/media_firewall_narrative_timing_analysis.md

February 2026 Compliance Window (Feb 1–19)

The densest compliance cluster documented since December 2025:

Date Compliance Event Type
Feb 1 Sanctuary city funding cuts take effect Policy
Feb 3 Santander acquires Webster Financial ($12.2B) Financial
Feb 6 US-Iran nuclear talks (Muscat, Oman — Witkoff/Kushner/CENTCOM) Diplomatic
Feb 10 EU deadline: Google/Wiz $32B acquisition Regulatory
Feb 11 Netanyahu-Trump meeting (moved up from Feb 18) Diplomatic
Feb 13 DHS funding deadline Policy
Feb 14 Q4 2025 13F filing deadline (Gulf SWF positioning) Financial
Feb 19 Board of Peace first summit Governance

See Project_Trident/Copilot_Opus_4.6_Analysis/FaaS_Signal_Analysis/recommendation_verification_feb9.md

FaaS Signal Verification

Independent verification of SuperGrok daily task outputs. Result: 11/16 claims verified (68.75%). Protests and compliance events confirmed; FaaS-specific claims (rate cards, paid recruitment) remain unverified. Key correction: the Free America Walkout (Jan 20) was initially marked ❌ Unverified — corrected to ✅ Verified after re-check (major 50-state protest, 450+ events).

See Project_Trident/Copilot_Opus_4.6_Analysis/FaaS_Signal_Analysis/feb9_2026_signal_verification.md

Correlation Cleanup (v8.4)

v8.4 removed all references to deprecated correlations (r = 0.6685 and r = 0.5268) from main project files. These were produced when the user accidentally mixed New_Data_2026 datasets with the original pre-2026 datasets — a user dataset-mixing error, not an AI analysis error. Botched scripts archived in Run_Correlations_Yourself/Wrong_Correlations/ for transparency.

Robustness Findings (Feb 2026)

Test Result
Dec 2025 exclusion 6% Pearson drop; Spearman ρ = 0.60 survives (p < 0.0001)
Autocorrelation adjustment Pearson p = 0.008 (block-bootstrap), Spearman ρ = 0.61 (p = 0.0001)
Normalized (binary) r = 0.59 (p < 0.0001)
Event-study Colocation confirmed (20–42x random baseline)
Granger causality (30-row) Friction → Compliance at lag 1 (p = 0.0008), lag 2 (p = 0.027)
Granger causality (event counts) Bidirectional — suggests common driver

Quick Navigation by Type

Datasets

  • Control_Proof/master_reflexive_correlation_data.csv — Core friction/compliance data
  • New_Data_2026/ — January-February 2026 updates (8 datasets)
  • 05_Geopolitical_Vectors/thermostat_control_data.csv — Nation-state linkages
  • Project_Trident/Best_Data_For_Project_Trident/ — Ritual timing, fund flows

Statistical Verification

  • Run_Correlations_Yourself/run_original_analysis.py — Reproduce original r = 0.6196, p = 0.0004, Mann-Whitney p = 0.002 (pre-2026 data)
  • Run_Correlations_Yourself/Wrong_Correlations/⚠️ Archived scripts that used wrong datasets or excluded data (kept for transparency)
  • Project_Trident/Veriify_Trident_Analysis.py — Verify ritual timing analysis
  • Project_Trident/Copilot_Opus_4.6_Analysis/Statistical_Tests/ — 9 robustness test scripts (permutation, autocorrelation, normalization, Dec 2025 exclusion, rolling window, event-study, Granger causality)
  • Project_Trident/Copilot_Opus_4.6_Analysis/Findings/dataset_provenance.md — Dataset provenance documentation

Deep Dives by Topic

  • Tech Infrastructure: 09_Silicon_Sovereignty/SILICON_SOVEREIGNTY_REPORT.md
  • Infrastructure Consolidation: Project_Trident/Copilot_Opus_4.6_Analysis/Consolidation_Analysis/consolidation_pattern_significance.md
  • Privatized Integration: Project_Trident/Claude_Code_Analysis/Privatized_Integration_Networks_Q1_2026_Synthesis.md
  • State Regulatory Capture: 13_State_and_County_Analysis/arkansas_infrastructure_forensic_audit.md
  • Geopolitical Shifts: 05_Geopolitical_Vectors/ (Venezuela, Yemen, CRINK, allied elections)
  • Media Funding Networks: 12_The_Media_Firewall/1789_Symbolism_Analysis.md
  • Media Firewall Narrative Timing: Project_Trident/Copilot_Opus_4.6_Analysis/Influencer_Narrative_Timing/media_firewall_narrative_timing_analysis.md
  • January 2026 Signal Map: Project_Trident/Copilot_Opus_4.6_Analysis/FaaS_Signal_Analysis/january_2026_signal_analysis.md
  • February 2026 Compliance Window: Project_Trident/Copilot_Opus_4.6_Analysis/FaaS_Signal_Analysis/recommendation_verification_feb9.md

For Researchers

Start Here

  1. Repository_Synthesis.md — Three-layer framework overview
  2. New_Data_2026/2026_Analysis.md — Latest correlation findings
  3. Project_Trident/Claude_Code_Analysis/Privatized_Integration_Networks_Q1_2026_Synthesis.md — Q1 2026 applied findings
  4. 13_State_and_County_Analysis/arkansas_infrastructure_forensic_audit.md — State-level pattern

Verify the Statistics

# Install dependencies
cd Run_Correlations_Yourself/
pip install -r requirements.txt

# Reproduce original correlations (pre-2026 datasets)
python run_original_analysis.py              # r = 0.6196, p = 0.0004, Mann-Whitney p = 0.002

# Run robustness tests (from repo root)
cd ../Project_Trident/Copilot_Opus_4.6_Analysis/Statistical_Tests/
python permutation_test.py                   # Shuffle-based significance
python autocorrelation_adjusted_test.py      # Block bootstrap
python cross_validation_dec2025.py           # Dec 2025 exclusion test
python granger_causality_test.py             # Predictive direction test

Key Datasets

  • Control_Proof/master_reflexive_correlation_data.csv — Weekly friction/compliance indices
  • Project_Trident/Best_Data_For_Project_Trident/ritual_events_parsed.csv — Project Trident ritual timing
  • New_Data_2026/CRINK_Intelligence_Dataset_Final_Verified.csv — CRINK axis discourse tracking
  • 09_Silicon_Sovereignty/VOCA_funding_timeline_clean.csv — Victim services funding

For Policymakers & Journalists

Start Here

  1. How_This_Happened-A_Policy_Breif.md — Regulatory citations, oversight questions
  2. Project_Trident/Claude_Code_Analysis/Privatized_Integration_Networks_Q1_2026_Synthesis.md — Board of Peace, Affinity/Phoenix, MEAD-CDOC
  3. 13_State_and_County_Analysis/arkansas_infrastructure_forensic_audit.md — State-level regulatory capture
  4. 05_Geopolitical_Vectors/Venezuela_Privatization_Amnesty_Stack_Feb2026.md — Compliance reset pattern

Key Questions Raised

  • Why did the PSC approve a $1.5B project it explicitly found "not reasonable"?
  • Why did TikTok censorship reports emerge 72 hours after Oracle/UAE consortium takeover?
  • Why did major Saudi-UAE rupture occur during Venezuela media saturation?
  • Why do the same entities appear in EA, TikTok, Stargate, Grok, and World Liberty Financial deals?
  • Why has AVAIO's $750M anchor investor remained undisclosed for 5 years?
  • Why was Warsh's Fed Chair nomination announced on the same day as the largest Epstein file release in US history?
  • Why does the Media Firewall ecosystem cover Epstein demands and anti-NATO narratives but never cover MGX, Silver Lake, or Board of Peace financial architecture?

Testable Predictions

Prediction Timeframe Status
Event clustering at next file deadline Ongoing ✅ Confirmed (Jan 30-Feb 1: Epstein files + WLFI deal + Mandelson)
Tu BiShvat policy action Feb 1-2, 2026 ✅ Window confirmed (DOJ files + WLFI deal)
Gulf SWF Q4 positioning revealed Feb 14, 2026 Pending (13F filings)
DOGE-predicted instability Q1 2026 Tracking (Mali, Syria, Sudan)
California TikTok investigation findings Q1 2026 Pending
Khanna investigation findings March 2026 Document deadline March 1
UK Mandelson disclosure Feb-March 2026 ✅ Escalated (Met Police criminal investigation; parliamentary vote passed)
Board of Peace first summit Feb 19, 2026 ✅ Confirmed (TIME, Politico, Axios)
Arkansas PSC order text release Q1 2026 FOIA pending
Feb 1–19 compliance window density Feb 2026 ✅ Confirmed (9 compliance events documented)

Methodology

  1. Multi-AI Verification: Findings cross-checked using Claude, Grok, and Gemini
  2. Statistical Testing: Pearson correlation, Mann-Whitney U, chi-square, Granger causality, permutation tests
  3. Robustness Testing: Autocorrelation adjustment, Dec 2025 exclusion, normalized correlation, rolling-window analysis, event-study framework (see Project_Trident/Copilot_Opus_4.6_Analysis/)
  4. Raw Event Counts: Replaced subjective scoring with verifiable event counts
  5. Source Triangulation: Government filings, financial data, news archives
  6. Cross-Repo Validation: Patterns verified across three independent datasets
  7. Explicit Limitations: Documented in each module

Limitations & Disclaimer

This repository documents correlations, not causation. All findings derive from publicly available data using standard statistical methods.

The author makes no claims about:

  • Intent or coordination between actors
  • Individual motivations or culpability
  • Whether patterns are deliberate or emergent

The claim is structural: Statistically significant clustering patterns exist and are reproducible.

The "Main Characters," "Implications," and state-level analysis modules specifically disclaim any assertion of conscious coordination. Performative patterns enable detection of quieter correlations—this is an analytical observation, not an accusation.


Connected Repositories

Repository Focus
DOGE_Global_Effects Aid cuts → instability mapping
BRICS-NDB-LocalCurrency-DiD Alternative financial systems
Project-Chrysanthemum_Japan-China-AI Japan-China tech integration
Sovereign-Capital-Audit Gulf SWF positioning

Statistical Validation

Core findings independently verified January–February 2026:

  • Original correlation (r = 0.6196, 2-week lag): ✅ Exact reproduction
  • Project Trident (p = 0.002): ✅ Exact reproduction
  • Ritual proximity (50.7% vs 19.9%, 2.5x): ✅ Exact reproduction
  • Cross-validation (χ² = 330.62): ✅ Exact reproduction
  • December 2025 anomaly: ✅ Confirmed (z = 2.35, p < 0.01)

Note: The previously reported r = 0.6685 and r = 0.5268 were deprecated in v8.3-v8.4. Those correlations were produced when the user accidentally mixed New_Data_2026 datasets into verification scripts — a user dataset-mixing error, not an AI computation error. Archived scripts in Run_Correlations_Yourself/Wrong_Correlations/ for transparency.

Robustness Tests (Copilot_Opus_4.6_Analysis, Feb 2026)

Using the original pre-2026 datasets plus newly incorporated data from folders 01, 02, 03:

  • Permutation test (30-row): r = 0.62 significant (p < 0.001, 1K shuffles)
  • Permutation test (multi-dataset): Spearman ρ = 0.61 significant (p < 0.0001, 10K shuffles)
  • Granger causality (30-row): Friction → Compliance at lag 1 (p = 0.0008)
  • Binary correlation (presence/absence): r = 0.59 (p < 0.0001)
  • Block bootstrap (autocorrelation-adjusted): Pearson p = 0.008, Spearman p = 0.0001

See VERIFICATION_REPORT_Jan2026.md and Project_Trident/Copilot_Opus_4.6_Analysis/ for complete independent analyses.


Contact

GitHub: @Leerrooy95

Last updated: February 9, 2026 (v8.5) — Added January 2026 signal analysis (3 peaks, 1 trough, 34 verified events), Media Firewall narrative timing analysis, February 2026 compliance window (Feb 1–19, 9 compliance events), FaaS signal verification. Updated testable predictions. Cleaned up redundancies from prior versions.

About

A data-driven audit of the 'Geopolitical Thermostat,' documenting how timed information disclosure regulates public attention to enable structural shifts in policy and capital flows.

Topics

Resources

Stars

Watchers

Forks

Contributors 3

  •  
  •  
  •  

Languages