Methodology
How we turn trade flow into actionable alerts.
Sight Whale tracks Polymarket trade activity, maps it back to wallets, and surfaces the most decision-relevant information. This page explains what we measure, what we do not measure, and the key limitations you should understand.
Data sources
Market activity
- Polymarket market metadata (events, markets, outcomes)
- Trade flow (size, price, side, timestamp) when available
- Publicly available on-chain data for settlement and attribution
Wallet context
- Historical activity and realized PnL where computable
- Aggregates by market, time window, and action type
- Optional display names (e.g., ENS or public aliases)
We do not collect private account data from Polymarket. We only use publicly observable information.
Signal selection
Alerts are triggered by a combination of trade size, wallet history, and the market context. We bias toward high-information moves.
Wallet scoring
Scores are based on measurable outcomes over time (e.g., realized PnL, consistency, drawdowns). The exact weighting is proprietary and may evolve.
Collections
Smart Collections group wallets by repeatable behaviors (e.g., high PnL, high score, specialized segments) to improve signal density.
Limitations & interpretation
Prediction markets can be thin, reflexive, and headline-driven. A large trade does not necessarily mean “inside information”; it can also reflect hedging, liquidity provision, or portfolio rebalancing.
Wallet performance metrics can be biased by survivorship, changing regimes, and incomplete data. Treat scores as directional context, not guarantees.
When data fields are missing from upstream sources, we fall back to best-effort normalization to keep pages and alerts consistent.