Prediction Markets
The same Forge Scale that rates protocol security also rates the integrity of individual prediction markets. A market-integrity score (300-1000) measures how exposed an average trader is to informed flow, manipulation, ambiguous resolution, and venue risk — using the same composite formula as the protocol BRI. Higher is safer.
Every dimension measures a structural property of a market — the shape of its flow, the clarity of its question, the trustworthiness of its resolution source. A low score flags where a careful trader should look harder. It is never a claim that a specific account, or the market itself, has broken a rule.
The Same Formula
Full breakdownScore = 300 + 700 × Π(Dᵢ / 100)^wᵢ
Multiplicative — one catastrophic dimension cannot be hidden by high scores elsewhere
Markets and protocols are scored by the same multiplicative composite: each dimension Dᵢ is 0-100 (higher = cleaner), combined as a weighted geometric mean and mapped onto the 300-1000 Forge range. Only the six axes below are specific to markets.
How a market is scored
Ingest
We pull each market's question, outcomes, order flow, traded volume, and declared resolution source from the venue, alongside calendar and news context for the entities the question names.
Signal
Six independent scorers each measure one structural property of the market — the shape of its flow, the cleanliness of its trades, the trustworthiness of its resolution path, its event exposure, its question clarity, and the health of its venue.
Compose
Each dimension is scored 0-100 (higher = cleaner) and combined as a weighted geometric mean, then mapped onto the 300-1000 Forge range — the same composite formula and tiers the protocol BRI uses.
Surface
Every market shows its composite score, its Forge tier, and a plain-language top risk driver naming the single weakest dimension — so a trader knows where to look harder before taking a position.
The Six Market Dimensions
Compare with protocolsEach dimension is an independent axis of market-integrity risk, scored 0-100 where higher is cleaner and safer. Below is what each one measures and what moves it.
How level the playing field is between the average trader and anyone who may be acting on information others do not have. We look at how much of the recent flow lands in concentrated, unusually well-timed positions ahead of known events.
Higher = a fairer, more symmetric market. A lower score flags more flow that looks well-informed — a heads-up to look harder, not a claim anyone broke a rule.
Broad, evenly distributed participation; positions that build gradually rather than spiking right before an outcome is known.
A few wallets driving most of the volume with sharply well-timed entries clustered just ahead of resolving information.
How clean the trade flow looks at the microstructure level. We screen for wash-trading proxies, the same wallet taking both sides, single-wallet volume dominance, and coordinated wallet clusters over a fixed recent window.
Higher = cleaner order flow. A lower score flags microstructure patterns associated with manipulation — it does not adjudicate intent.
Many independent participants; balanced taker/maker flow; no single wallet dominating recent volume.
Self-trading patterns, extreme one-sided taker pressure, or one wallet responsible for an outsized share of 24h volume.
How cleanly and verifiably the market can be settled. We score the declared resolution source by reputation, whether the criteria point to a single named authority versus an ambiguous source, and whether an explicit dispute path exists.
Higher = a clearer, more trustworthy settlement path. A lower score flags ambiguous or single-source resolution.
A named, reputable authority (official data, tier-1 outlets) as the resolution source, with an explicit dispute mechanism.
Resolution that hinges on operator discretion, an anonymous source, or criteria that point to no verifiable authority.
How exposed the market is to a scheduled, market-moving event in the near term. We extract the entities the question names, match them against upcoming calendar events, and weight by how close the nearest catalyst is.
Higher = a calmer horizon. A lower score flags a near-term scheduled event that can move the market discontinuously and widen information gaps.
A clear runway with no imminent scheduled catalyst tied to the question's entities.
A known, calendar-scheduled event (election, ruling, data release, FOMC) landing within days of the trade.
How unambiguous the market's wording and rules are. We assess question structure, whether a specific resolution date is stated, and the presence of fuzzy language that leaves room for disputed outcomes.
Higher = clearer rules with less room for a disputed outcome. A lower score flags wording that could be read more than one way.
Precise wording, a specific resolution date, and rules that admit only one reasonable interpretation.
Vague timing ('soon', 'this year'), hedging language ('officially', 'deemed', 'expected'), or an over-long, convoluted question.
The health of the platform every market on it depends on. This platform-wide signal covers stablecoin peg stability, acute regulatory developments affecting the venue, and the operational standing of its settlement infrastructure.
Higher = a healthier venue. A lower score reflects platform-level stress that affects every market on the platform, not this market alone.
A stable settlement currency, clean regulatory standing, and healthy, well-maintained settlement infrastructure.
Peg instability in the platform's stablecoin, acute regulatory pressure on the venue, or settlement-infrastructure stress.
Weight Distribution
Weights sum to 100%. Information Fairness carries the most weight because information asymmetry is the dominant risk a retail trader faces in a prediction market.
How to read a market score
It tells you where to look harder before you take a position — not whether to trust or distrust anyone.
Each market surfaces the single lowest-scoring weighted dimension in plain language, so you know what is pulling the score down.
When a dimension cannot consult live evidence, it falls back to a conservative baseline and is marked as such on the market detail page.
Every line of evidence describes a measurable structural property.