Explore every Worlds draft.
Explore picks, bans and role trends, with the evidence behind each champion.
Model method, freshness and coverage
Final delivery is not a guarantee of accuracy. S15 Worlds was used for candidate acceptance and calibration, not an independent blind test; S16 is unvalidated. Historical references compare usage, not future forecast errors.
Models and filters
Filters update charts, tiers and the champion tableForecasts and match references
Upper bar: forecast. Lower bar: match reference. Pick rates use role segments.
Forecast top 20 and match reference
All champions
Champion tiers
Gray outer frame: fewer than 1 expected pick over the tournament. Red: ban rate ≥10% and at least twice the pick rate. Both marks may apply. Tiers are display thresholds, not measured strength.
Explore every champion
| Champion | Total pick | Ban | BP presence | Mean pick position | Expected picks | Reference for selected metric |
|---|
Evaluation and its limits
Assumptions, weights and limitations
Scope and effect type are separate; ultimate/basic, stage and nonlinear duration are encoded. Professional response and a mechanism prior use a 50/50 blend. Prior weights were calibrated within an S15 error budget, not established as causal values. S15 BP RMSE is limited to 1.10 times the previous model; pick/ban RMSE to 1.15 times. Practice has a 6% mixture weight without comparable pre-S15 practice validation. Player-champion win rate/KDA are excluded.