Will it resolve cleanly?
A 0–100 resolution-risk score per market — every factor visible and weighted in the open.
indexing Polygon + Ethereum live
See inside any prediction market before you trade it.
Resolution-trust analytics built from raw on-chain events — our own multi-chain indexer, not third-party APIs. Know whether a market will resolve cleanly, and who actually decides when it doesn't.
Research & analytics. Not betting advice, no outcome predictions.
live index telemetry
870,412
Total markets tracked
52,125Open markets
1,019
Disputes settled by an Ethereum UMA vote
Polygon → Ethereum
Polygon → Ethereum
41,039
Total markets scored
9,214Open markets scored
2
Chains reconciled end-to-end
What it answers
Three questions, per market.
Every answer is derived from on-chain events and shown with its inputs — no black-box scores.
Who decides if it's contested?
Cross-chain dispute-vote concentration: which Ethereum voters actually settle disputed Polygon markets.
top-5 share 71%4 pivotal
Is the price telling the truth?
Trading-layer signals: cross-market mispricing, late-buy spikes, thin liquidity, and volume concentration.
late-buy spikethin liquidityΣp 1.14deep liquidityconcentrated volume
The hard part
One market, two chains, one answer.
A dispute is raised on Polygon. The outcome is decided by a UMA DVM vote on Ethereum mainnet. We join the two on-chain, so a contested resolution can be traced to the wallets that actually settled it — something single-chain data tools can't produce.
PolygonDispute raised
market0x7c4d…b31a
eventDisputePrice
block#58,640,912
keccak(ancillaryData)
Ethereum L1UMA DVM vote
request0x1f09…44de
revealed4.28M UMA
voters27 distinct
→
4
pivotal voters decided it
Top-5 controlled 71% of revealed weight. HHI 2,140.
Dominance is not manipulation.UMA's DVM is a Schelling-point oracle — voters converging on the truthful outcome is indistinguishable from collusion using vote data alone. We report concentration, never intent.
What we won't claim
The limits are part of the product.
An analytics tool that overstates its own certainty isn't research. Here's exactly what this data can and cannot tell you.
✕
No betting advice, no predictions
We don't forecast outcomes or tell you what to trade. We measure how a market resolved and how contested that was.
✕
Dominance ≠ manipulation
Voting concentration is a Schelling-point pattern. Convergence on the truth looks identical to collusion in vote data. We never assert intent.
✕
Wallets aren't people
Concentration is measured over on-chain proxy wallets — not unique humans. One person may hold many; one wallet may serve many.
✕
Fills, not order books
Everything derives from executed on-chain trades. Mispricing is an implied-price deviation, not an executable arbitrage after fees and slippage.
Build on it
The whole dataset, free over GraphQL.
Every number on the dashboard is queryable. Keys are free — we hand them out in the Telegram group.
POSTapi.predictionxray.com/graphql
# riskiest open markets, with the factors behind each score { riskyMarkets(first: 5, openOnly: true) { totalCount nodes { questionId resolutionRiskScore factors { key normalized weight } disputeVote { pivotalCount top5Share } } } }
Free API key— ask in the Telegram group and you'll get one. Same group for feedback, bug reports, and schema questions.
Browsing the schema in GraphiQL at that host? Add your key as an x-api-keyheader in its header pane — requests without it return 401. That's the gate doing its job, not a broken API.
Read-only. Cursor-paginated connections. No auth needed for public aggregates.
Start with the riskiest open markets.
No signup, no wallet connect. The dashboard is read-only and open.