We maintain a dataset of 164,000+ prediction markets scored for dispute risk. Of those, 1,654 were actually disputed through the UMA oracle system. Here's what the data shows about which markets get disputed, which don't, and the patterns that predict it.
The Numbers
Between November 2023 and April 2026, across Polymarket and UMA-resolved markets:
- 164,000+ markets in our database
- 1,654 were disputed (1.0% base rate)
- Highest-risk markets: 19.6% dispute rate (8.2x baseline)
- Lowest-risk markets: 1.5% dispute rate (0.6x baseline)
Disputes aren't random. They cluster around specific question types with predictable characteristics.
Notable Disputed Markets
"Biden senile during the debate?" — Risk Score: 75%
Why it was disputed: Resolution required that Biden "exhibits clear signs of senility." This is a subjective medical judgment with no defined criteria. What constitutes "clear signs"? Who decides? There's no agreed-upon clinical threshold in the resolution text. The UMA oracle couldn't resolve it cleanly.
Pattern: Subjective human judgment with no measurable threshold.
"Was Barron involved in $DJT?" — Risk Score: 75%
Why it was disputed: Resolution required "a preponderance of evidence" of involvement. What counts as evidence? What's the threshold for "preponderance"? This is legal language being applied without a legal process. Two disputes were filed.
Pattern: Legal/evidentiary standard without a defined arbiter.
"Will Gaza ceasefire deal be reached before December 31 2024?" — Risk Score: 75%, Disputed 2x
Why it was disputed: "Ceasefire deal" has no universally agreed definition in international relations. A temporary pause? A written agreement? An announced framework? Implemented on the ground? Two separate disputes were filed because different stakeholders interpreted "deal" differently.
Pattern: Geopolitical term with no precise definition.
"Will there be a ceasefire in Ukraine before March 2026?" — Risk Score: 75%, Disputed 2x
Why it was disputed: Same ambiguity as Gaza. "Ceasefire" is not a binary state in active conflicts. The market was disputed twice with different interpretations of what constitutes a ceasefire.
Pattern: Recurring — ceasefire questions are dispute magnets regardless of which conflict.
"Did Israeli intelligence have advanced knowledge of the Attack?" — Risk Score: 85%
Why it was disputed: "Advanced knowledge" is inherently subjective. How much knowledge? How far in advance? From what source? Does a general warning count? Does intercepted chatter count? The question can't be cleanly resolved by any single data source.
Pattern: Intelligence/knowledge claims without specific, verifiable criteria.
The Three Dispute Patterns
Across all 1,654 disputes, three patterns dominate:
Pattern 1: Subjective Threshold (45% of disputes)
The resolution criteria require someone to judge whether something "clearly" or "significantly" or "materially" happened. Without a numeric threshold or authoritative arbiter, different people reach different conclusions.
Trigger phrases: "clear evidence," "significant," "major," "material," "reasonable person"
Pattern 2: Ambiguous Key Terms (35% of disputes)
The core question uses a term that reasonable people define differently. Geopolitical terms are the worst offenders: "ceasefire," "regime change," "deal," "crisis," "involvement."
Trigger phrases: Any geopolitical or social term used without an explicit definition in the resolution criteria.
Pattern 3: Source Reliability (20% of disputes)
The resolution depends on a source that may not publish clearly, may contradict itself, or may not exist at the time of resolution. "According to media reports" is a classic source problem.
Trigger phrases: "according to reports," "widely reported," "confirmed by sources," "media consensus"
Markets That Were Never Disputed
For contrast, here are categories with near-zero dispute rates:
| Category | Example | Dispute Rate | Why |
|---|---|---|---|
| Sports outcomes | "Will France win the World Cup?" | <1% | Single authority (FIFA, NBA, etc.) |
| Central bank decisions | "Fed rate cut in June?" | <1% | FOMC press release is unambiguous |
| Election results | "Will X win the election?" | 2-3% | Official certification process |
| Binary financial data | "BTC above $100K on Jan 1?" | <1% | CoinGecko/CMC price at specific timestamp |
The pattern: a single, authoritative, publicly accessible source that produces an unambiguous YES/NO answer at a specific point in time.
What This Means for Traders
If you're trading Polymarket:
- Avoid markets with subjective resolution criteria — or at least size down. Our data shows these get disputed at 3-8x the baseline rate.
- Be especially careful with geopolitical markets — "ceasefire," "regime change," and "deal" are the most disputed terms in our dataset.
- Prefer markets with authoritative resolution sources — sports, financial data, regulatory decisions.
- Check dispute risk before entering — we score every market and send alerts via Telegram.
The Full Dataset
We're not publishing the full list of 1,654 disputed markets (it's part of our scoring model's training data), but we do make dispute risk scores available for every open market via our Telegram bot (@OracleManglebot) and API.
If you want to check a specific market's dispute risk, message the bot or check the live scores.
OracleMangle has scored 164K+ markets for dispute risk. Join the free Telegram bot to check any market before you trade.