Win%
Calculated round win% from a multivariate logistic regression model, ±0.049
Swing
Win% minus 50%, how far above or below average the player performs
Raw
Swing relative to the lobby average, scaled by a factor of the team's round differential
Shift
Constant added that accounts for the team's performance, >1 indicates error
RAA
Rounds above average — how many rounds this player contributed above a lobby-average player