How Skilled Wallets finds copyable traders
From raw on-chain trades to a single 0-100 score: every step of the pipeline explained, so you know exactly what you are paying for.
What the platform does
Daily on-chain ingestion
The full Polymarket market catalog and complete per-wallet trade histories are synced every day from public APIs.
Bot and market-maker filtering
Exclusion filters remove high-frequency bots, institutional liquidity and maker-heavy wallets with no forecasting edge.
Excess Hit Rate (EHR)
Measures whether a wallet systematically buys contracts below their true probability — the core skill signal.
Risk-adjusted metrics
Sharpe, Sortino and Calmar analogs plus max drawdown, computed on each wallet’s monthly PnL series.
Composite 0-100 scoring
Each metric maps to a 0-100 score through transparent curves; the final score is their weighted average.
Actionable wallet signals
A thesis, top categories, suggested sizing and approximate timing derived from each wallet’s historical behavior.
Product
What you see in the dashboard
Wallet ranking
Sort by final score, EHR, Sharpe or volume; filter by score range; paginate through the full universe of skilled wallets.
Wallet detail
Fourteen metrics per wallet, a monthly PnL chart, first/last seen dates, and the trade activity behind the numbers.
Signal card
A concise thesis for each wallet with its strongest market categories, suggested sizing (low/medium/high) and timing guidance.
Common questions about the methodology
The details behind the pipeline, filters and scoring.
Why do you exclude profitable market makers?
Their profit comes from capturing spread across thousands of small trades, not from predicting outcomes. Copying their positions manually loses the spread income and keeps the risk.
How often are scores recomputed?
Daily. The pipeline re-ingests fresh market and trade data and recomputes every metric and score from scratch each day.
What does a missing metric mean?
It means there was not enough data to compute it — for example, a Sharpe analog needs at least three months of activity. We show "no data" instead of a misleading zero.
Is the scoring formula public?
Yes in substance: the final score is a weighted average of per-metric 0-100 scores produced by linear interpolation curves based on published prediction-market research.
See the ranking for yourself
Create an account and pick a plan — full access to every wallet, metric and signal.