Method
Each item is scored as an action or observable pattern, not as a tribe or identity group. The model cares most about citizen harm, institutional durability, reversibility, and evidence quality.
Scoring Method
Scores are structured judgments, not objective measurements. Disagree by challenging the evidence, the weighting, or the counterargument.
Immediate Damage
Realized harm, institutional damage, human damage, scale/duration, and precedent.
Long-Term Damage
Realized harm, tail risk, institutional damage, human damage, and irreversibility over a 10-30 year horizon.
People-First Score
Citizen impact is a weighted score: human damage 36%, realized harm 20%, long-term damage 18%, institutional damage 16%, and irreversibility 10%. It puts ordinary people ahead of factional drama.
Confidence Adjustment
High-confidence long-term scores keep their full value. Medium-confidence scores are discounted to 86%. Low-confidence scores are discounted to 68%. The public ranking still shows the raw score, but uncertainty changes how aggressively the claim should be treated.
Uncertainty Bands
- High confidence: score range is plus or minus 5 points.
- Medium confidence: score range is plus or minus 10 points.
- Low confidence: score range is plus or minus 16 points.
Subscores
- Realized Harm: damage already visible or measurable.
- Tail Risk: catastrophic potential if repeated or normalized.
- Institutional Damage: harm to elections, courts, schools, policing, DOJ, watchdogs, federalism, civil service.
- Human Damage: deaths, injuries, learning loss, crime victims, mental health, household damage.
- Irreversibility: how hard the damage is to repair.
Score Bands
- 90-100: systemic, generational, or constitutional-scale harm.
- 80-89: severe harm or severe tail risk.
- 70-79: major governance or institutional harm.
- 60-69: serious but bounded or partially reversible harm.
- 50-59: meaningful but contested or mostly indirect harm.
- 30-49: symbolic, rhetorical, or trust damage unless paired with actions.
- 0-29: normal democratic conflict, ordinary policy dispute, or weak evidence.
Source Hierarchy
Preferred sources are primary documents, official data, reputable watchdogs, wire services, high-quality reporting, academic or think-tank work, then advocacy organizations with methodology caveats.
Claim Metadata
Each card has a stable claim ID, evidence-quality label, factual maturity, legal posture, scoring maturity, jurisdiction level, and source profile. These fields are defined in the data dictionary.
Revenue Firewall
Sponsors, memberships, paid briefs, and data licenses can fund the work. They cannot buy scores, source removal, suppressed counterarguments, or unlabeled content.
How To Beat A Score
Bring Better Evidence
The site should change when the record changes. The fastest correction is a specific source that moves a subscore, narrows uncertainty, or shows the claimed mechanism did not cause the harm.
- Challenge the fact.Show a source that contradicts the event, count, ruling, policy, or time range.
- Challenge the mechanism.Show that the harm came from another cause, or that the listed actor had weak control over it.
- Challenge the weight.Argue a subscore is too high or too low using citizen harm, duration, reversibility, and precedent.
- Challenge the confidence.Replace weak sources with primary documents, official data, court records, or better research.