Causal Assessment of Unified Societal Effectiveness (CAUSE 13.0)
CAUSE 13.0 is our proprietary zero-based causal index that measures the true long-term impact of world leaders, governments and influential entities on ordinary people’s lives.
Every dimension starts from absolute zero — points are earned only with concrete, cited evidence. The system applies a Dynamic Harm Exponent (^7→^10): long-term negative consequences and human suffering weigh exponentially heavier than positive outcomes. Special focus is placed on uncovering hidden affiliations, dangerous political connections, secret agreements and domino effects that affect society, security and freedom.
This is the most critical and voter-oriented version of the index, built to reveal the full picture before elections and important decisions.
Purpose of the Index
- Zero-based scoring: every dimension starts at 0.00. Every point must be EARNED with specific cited evidence. No evidence = no credit.
- Separates genuine causal impact from external factors with extreme precision.
- Dynamic Harm Exponent (^7→^10) — suffering weighs exponentially heavier than benefit. The system actively and autonomously searches for and heavily penalizes ANY long-term negative consequences on ordinary people.
- Never allows economic or short-term successes to offset ethical failures, authoritarianism, human rights violations or suppression of freedom.
- Tracks Domino Effect chain-reactions: initial harm → propagation → secondary → tertiary harm across domains and generations.
- Automatically generates new nodes, satellite clusters and meaningful connections in the Network graph.
- VOTER PROTECTION (CAUSE 13.0): Searches for ALL hidden affiliations — funding trails, mentors, party donors, foreign links, oligarch connections — and penalizes connections to harmful figures. Bad connections are searched FIRST.
- Produces a fully reproducible 0–100 score with every value traceable to sources.
Automatic verdict classification
The final CAUSE 13.0 score maps to one of six verdict tiers. Each tier represents a distinct category of leadership effectiveness, from the highest achievement to the most severe failure.
Recommendation for resignation or international investigation
Final Formula (CAUSE 13.0)
Geometric Mean (GM)
Weighted product of all 31 utility values. If any U_i falls below 0.20, the GM is automatically multiplied by 0.30 — preventing cherry-picked excellence.
Utility Function U(x)
Bayesian MAUT transform: U(x) = (1 − exp(−k·x)) / (1 − exp(−k)) with k = 4.8. Compresses extreme scores and rewards consistent performance across all dimensions.
CAUSE 13.0: Dynamic Harm Exponent (^7→^10)
The Human Suffering multiplier uses a Dynamic Harm Exponent that scales from ^7 to ^10 based on severity, making harm devastatingly more impactful than benefit. Power table (at ^7): 0.9^7 = 0.478 (52% wiped), 0.7^7 = 0.082 (92% gone), 0.5^7 = 0.008 (99% gone), 0.3^7 = 0.0002 (99.98% destroyed). This single mechanism ensures no amount of positive achievement can offset serious harm to people.
CAUSE 13.0: Domino Effect Tracking
Chain-reaction cascade tracking across domains and time. Traces initial harm → propagation → secondary → tertiary harm. A single policy failure in economics can cascade into social instability, institutional erosion, intergenerational poverty, and cultural damage — each step multiplying the total impact.
CAUSE 13.0: Affiliation Penalty (Voter Protection)
If the entity is connected to persons or organizations with a CAUSE score below 40, the AffiliationPenalty_Multiplier reduces the score. Range: 0.35–1.0. The system searches 10 categories of hidden connections: funding/donors, mentors/protégés, business ties, family connections, party affiliations, lobbying ties, foreign government links, oligarch connections, secret agreements, revolving-door. Bad connections are searched FIRST.
CAUSE 13.0: Protégé Shadow Deduction
A flat deduction of 0–45 points subtracted AFTER all multipliers. If the entity is a protégé, ally, or mentee of a harmful figure, this deduction ensures the connection is always visible in the final score. The result is clamped at ≥ 0. Example: protégé of a convicted corrupt leader = −25 to −45 points.
Scale & Duration Factors
PopulationScale_F, DurationMultiplier, GlobalSpillover_F, and IntergenerationalLegacy_F ensure that actions affecting large populations over long time spans are appropriately weighted.
How CAUSE 13.0 automatically generates nodes and connections
When CAUSE 13.0 is calculated, the system automatically generates new entities and relationships in the Network graph:
This enables the graph to grow logically and organically around multiple centers of influence.
Hidden Affiliation Discovery (CAUSE 13.0)
We don't just score the person — we score who they are connected to and what shadow they carry. CAUSE 13.0 introduces two powerful new mechanisms that expose the full underbelly of any candidate before an election. Bad connections are searched FIRST, before any positive achievements.
If a leader is connected to persons or organizations with a known CAUSE score below 40, their own score is multiplied by 0.35–0.55 depending on the severity and number of connections. A clean politician with no bad connections receives 1.0 (no penalty).
If the entity is a protégé, ally, or mentee of a figure with a low CAUSE score, a flat deduction is subtracted from the final score after all multipliers. This ensures the “shadow” of a harmful mentor is always visible.
10 Hidden Connection Search Categories
The LLM is instructed to exhaustively search ALL of these categories for every entity analyzed. Discovered connections are flagged, scored, and displayed in the dossier.
Full description of all 31 dimensions
Each dimension is scored independently using Bayesian analysis with 95% credible intervals. CAUSE 13.0 uses 31 dimensions capturing democratic integrity, human rights, geopolitical stability, population scale, and intergenerational effects.
Core Leadership Dimensions
U1–U10 — foundational measurements of leadership effectiveness
Promise Fulfillment & Narrative Integrity
Measures how well public promises are kept and how consistent words are with actions.
LLM task: Extract all measurable commitments from speeches, interviews and manifestos. Compare with official data and fact-checks. Calculate fulfillment rate and narrative gap.
Causal Outcomes & Value Creation
Real contribution to economic and social outcomes (not just raw numbers). Uses Dynamic Synthetic Control to isolate leader's effect.
LLM task: Identify key KPIs and run causal analysis.
Stakeholder Wellbeing Impact
Change in quality of life for citizens, employees, customers and environment.
Leadership & Organizational Capital
Quality of team management and institutional capital.
Innovation & Future Readiness
Ability to create the future.
Ethics, Governance & Sustainability
Honesty, transparency and long-term sustainability.
Resilience & Risk-Adjusted Performance
Ability to withstand shocks.
Opportunity Cost & Missed Alternatives
What was lost compared to alternative scenarios.
Personal Integrity Consistency
Consistency between words and actions over time.
Societal Reception & Critical Accountability
How society and independent sources perceive the leader.
Democratic & Rights Dimensions
U11–U15 — democratic integrity, human rights, and geopolitical stability
Democratic Integrity & Institutional Resilience
Strength of democratic institutions under the leader.
LLM task: Analyze changes in V-Dem indices, judicial independence, election integrity.
Freedom of Speech & Information Integrity
Level of media freedom and information flow.
LLM task: Analyze Reporters Without Borders index, number of journalists imprisoned, censorship cases.
Human Rights & Minority Protection
Protection of human rights and minorities.
LLM task: Analyze Amnesty International, Human Rights Watch reports, number of political prisoners, discrimination cases.
Geopolitical Stability Contribution
Contribution to or damage of global stability.
LLM task: Analyze impact on international relations, sanctions, alliances.
Civilizational & Moral Legacy Impact
Long-term moral judgment by historians and future generations.
Scale, Duration & Legacy Dimensions
U16–U25 — population scale, intergenerational effects, and civilizational impact
Population Affected Scale
How many people were directly or indirectly affected.
Duration of Impact
Distinguishes short-term vs multi-generational effects.
Intergenerational Legacy Projection
Projects impact on children and grandchildren (education, debt, environment, values).
Global Spillover & Contagion Effect
How actions in one country affect others globally.
Cultural & Ideological Propagation
How ideas and values spread (positive or toxic).
Environmental & Planetary Impact
Carbon footprint, biodiversity loss, long-term climate contribution.
Technological & Future Readiness Acceleration
How much the entity accelerated or slowed technological progress.
Institutional Erosion or Strengthening
Effect on courts, media freedom, civil society and democratic institutions.
Social Cohesion & Polarization Effect
Changes in social trust and polarization.
Historical Reputation & Moral Legacy
Long-term moral judgment by historians and future generations.
Media, Technology & Structural Power Dimensions
U26–U31 — media manipulation, digital rights, inequality, climate, AI ethics, and strategic dependency
Media Manipulation & Propaganda
Measures the extent to which media is weaponized to control public narratives, suppress dissent, or manufacture consent.
LLM task: Analyze media ownership concentration, state media influence, propaganda campaigns, social media manipulation, and bot networks.
Digital Privacy & Surveillance
Evaluates protection or erosion of digital rights, mass surveillance programs, and data privacy practices.
LLM task: Analyze surveillance legislation, data collection practices, encryption policies, digital rights protections, and whistleblower cases.
Wealth Inequality & Economic Justice
Measures changes in wealth distribution, economic mobility, and access to economic opportunity.
Climate Accountability & Environmental Stewardship
Evaluates climate policy commitments vs. actual emissions, environmental regulation enforcement, and stewardship of natural resources.
LLM task: Analyze Paris Agreement compliance, emissions trajectories, deforestation rates, renewable energy investment, and environmental regulatory changes.
AI & Technology Ethics
Measures responsible governance of emerging technologies including AI, biotechnology, and autonomous systems.
LLM task: Analyze AI regulation frameworks, algorithmic bias incidents, technology export controls, and ethical oversight mechanisms.
Strategic Dependency Creation & Power Lock-in
Measures how entities create systemic dependencies that lock populations into reliance on specific actors, resources, or frameworks — reducing autonomy and choice.
LLM task: Analyze infrastructure monopolies, resource dependency creation, institutional capture, vendor lock-in at national scale, and deliberate erosion of alternatives.
Global coefficients (strengthened)
CAUSE 13.0 introduces the Dynamic Harm Exponent (^7→^10) for the Human Suffering multiplier and Domino Effect chain-reaction tracking. Suffering weighs exponentially heavier than benefit. These coefficients adjust the raw geometric mean to account for causality, evidence quality, external context, human suffering, cascading harm, and scale of impact.
Causal Attribution Coefficient
Isolates the leader's direct causal contribution from external factors. Uses Dynamic Synthetic Control with a donor pool of comparable entities, placebo tests, and bias correction.
Evidence Quality Coefficient
Reflects source quality and triangulation. Requires multiple independent, high-quality sources for maximum confidence. Penalizes reliance on single-source or low-tier evidence.
Context Factor
Adjusts for global shocks (wars, pandemics, economic crises) that affected the evaluation period. Ensures fair comparison across different eras.
Legacy Projection Factor
Projects the long-term impact trajectory into the future using causal ML forecasting. Leaders whose policies have compounding positive effects receive upward adjustment.
Human Suffering Multiplier (Dynamic Harm Exponent ^7→^10)
CAUSE 13.0's most devastating component. Uses a Dynamic Harm Exponent (^7→^10) that scales with severity, making harm devastatingly impactful. Range: 0.02–1.0. Even a seemingly mild 0.9 becomes 0.9⁷ = 0.478, wiping 52% of the score. At 0.7 → 0.082 (92% gone). At 0.5 → 0.008 (99% gone). The LLM autonomously searches across 10 harm categories for any long-term negative consequences on ordinary people.
Domino Effect Multiplier
Chain-reaction tracking across domains and time. Traces initial harm → propagation → secondary → tertiary harm. Range: 0.05–1.0. Captures how a single policy failure cascades through economic, social, institutional, and intergenerational domains, compounding damage far beyond the original action.
Population Scale Factor
Scales impact by the number of people directly or indirectly affected. Formula: log10(affected_population + 1) / log10(global_population) × impact_depth.
Duration Multiplier
Distinguishes short-term vs multi-generational effects. Positive impact: 1 + (years / 50). Negative impact: 1 − (years / 30).
Global Spillover Factor
Measures how actions in one country affect others globally — contagion effects, cross-border consequences, and international ripple effects.
Intergenerational Legacy Factor
Projects impact on future generations — children and grandchildren — through education, debt, environment, and cultural values.
Affiliation Penalty Multiplier (NEW in 13.0)
CAUSE 13.0 — VOTER PROTECTION. If the entity is connected to persons or organizations with a CAUSE score below 40, this multiplier penalizes them. Range: 0.35–1.0. The system searches 10 categories of hidden connections: funding/donors, mentors/protégés, business ties, family connections, party affiliations, lobbying ties, foreign government links, oligarch connections, secret agreements, and revolving-door relationships. Bad connections are searched FIRST, before any positive achievements.
Protégé Shadow Deduction (NEW in 13.0)
CAUSE 13.0 — VOTER PROTECTION. If the entity is a protégé, ally, or mentee of a figure with a low CAUSE score, a flat deduction of 0–45 points is subtracted from the final score AFTER all multipliers. This ensures that being the protégé of a harmful figure always has a visible impact on the score, regardless of other factors. Example: a political protégé of a convicted corrupt leader receives -25 to -45 points.
Workflow in the application
From data input to final output, every CAUSE 13.0 assessment follows this rigorous 7-step process. Each step is automated and auditable.
Input Data
The user submits a query — a leader, government, corporation or influential entity to evaluate.
Automatic Collection via Grok API
Automated gathering from 20+ source categories: news archives, government data, economic databases, social media, Wikipedia, fact-checkers, and academic research.
LLM Processing
LLM extracts measurable claims from sources. Each claim is verified against multiple independent sources and classified as confirmed, disputed, unverified, or refuted.
Causal Modeling
Dynamic Synthetic Control separates the leader's actual impact from external factors like economic cycles, global events, and inherited conditions. Only causally attributed outcomes count.
Bayesian Scoring + All 31 Dimensions and Penalties
Each dimension is scored using Bayesian analysis with Beta(2,2) priors. Raw scores are transformed through the MAUT utility function. All penalties, guardrails, and global coefficients are applied.
Automatic Creation of Nodes, Satellites and Connections
New nodes, satellite clusters and meaningful connections are automatically generated in the Network graph based on impact scale, penalties, and causal relationships.
Output JSON + Radar Chart + Full Audit Trail
The final output includes a structured JSON result, a radar chart of all 31 dimensions, and a complete audit trail with LLM call counts, token usage, and computation time.
Safety by design
CAUSE 13.0 is designed with safety as a first principle. Every assessment includes built-in protections against misuse, bias, and harm.
No guilt assertions
We never assert guilt, criminality, or corruption as fact. Every finding is evidence-grounded with explicit confidence levels.
Source transparency
Every conclusion traces back to specific source evidence. Users can verify any claim by following the evidence chain.
Limitations disclosed
Alternative explanations, data gaps, and assessment limitations are always presented alongside findings.
Ethics guardrails
Hard mathematical constraints prevent ethical failures from being masked by strong economic performance.
Private individual protections
Non-public figures receive stricter protections by default. Hypothesized relationships are visually distinct from confirmed ones.
Adversarial validation
A dedicated red-team pass challenges all findings, checks for bias, and penalizes hidden manipulation or narrative gaps.
Technical transparency
Open parameters
- •All 31 dimension weights are published on this page
- •Utility function shape (k = 4.8) is fixed and documented
- •Guardrail thresholds (U < 0.20 → GM × 0.30, Ethics < 0.40 → CAUSE × 0.35) are hardcoded
- •Every coefficient includes human-readable reasoning
Audit trail
- •Each assessment records LLM call count and token usage
- •Computation time is tracked and displayed
- •Score before and after guardrails is always shown
- •Radar chart of all 31 dimensions included in every output
- •Assessment limitations are explicitly listed per evaluation