Methodology

• Specific predictions — Not vague opinions, but measurable claims with clear outcomes
• Clear timeframes — Every prediction must have a deadline for resolution
• Verifiable sources — All predictions are linked to public statements (articles, tweets, videos)
• Binary outcomes — Predictions are scored as CORRECT or WRONG, no partial credit

We use a paper trading simulation to calculate PnL (Profit & Loss) for each prediction:

• Position sizing — $100 virtual stake per prediction
• Market odds — When available, we use Polymarket odds to determine potential returns
• Win rate — Simple percentage of correct predictions
• ROI — Return on investment based on simulated trades

Pundits are ranked by their paper PnL. To appear on the leaderboard:

• Minimum 3 resolved predictions — Ensures statistical relevance
• Primary sort: Total PnL — Who's made the most (or lost the least)
• Secondary: Win rate — Consistency matters

Predictions are resolved through a transparent process:

• Automatic matching — AI matches predictions to Polymarket outcomes when possible
• Manual review — Our team reviews edge cases and ambiguous outcomes
• Community input — Users can vote on prediction accuracy
• Appeal process — Disputed resolutions can be reviewed

Every prediction is cryptographically sealed and linked in a tamper-evident chain:

• Content Hash — Each prediction's claim, quote, source, and timestamp are hashed using SHA-256
• Chain Link — Each prediction links to the previous one, creating an unbreakable chain
• Tamper-Evident — Any modification to past predictions breaks the chain and is immediately detectable
• Public Verification — Anyone can verify a prediction's authenticity using its hash