Awaiting first refresh — totals will appear after the next refresh-champion-summary cron tick.
283,880 total matches crawled
Last 30 days.
How champion-roles are distributed across each metric (Platinum+, last 30 days, min 200 games). Dashed orange line = mean.
Bayesian-smoothed WR per champion-role.
Share of total games at this role.
Share of games where the champion was banned. Champion-level — bans aren't role-specific.
WR slope vs game length. Positive = scales late.
WR slope across ranks. Positive = better in high elo.
WR slope vs career mastery (ln championPoints).
Mean absolute deviation of WR across lane matchups. High = swingy by opponent.
Per-role z-score residual: WR in ≤20 min games vs what overall WR would predict. Positive = closes early-finishing games more than baseline implies.
Not enough data.
Per-role z-score residual: WR in ≥30 min games vs what overall WR would predict. Positive = scales late beyond baseline. Catches U-shape scalers the linear Scaling slope misses.
Not enough data.