The rating system monitors a player’s performance through their winning probability per game, calculated from the rating difference between the two players — a 100-point advantage gives you roughly a 60%–40% winning expectation, and a 200-point advantage roughly 69.5%–30.5%. It’s non-zero-sum, meaning your rating gain doesn’t have to equal your opponent’s rating loss. The system also tracks a rating deviation for each player, a measure of how uncertain that rating is: players who’ve stayed active have more stable ratings, while players who’ve been inactive for longer accumulate greater uncertainty and see bigger rating movements when they return. The size of a rating change also depends on which rating tier you’re in, so the same result can affect players differently depending on their rating level. Your rating remains provisional until you’ve completed 50 rated games.