Crucible Rating System · Season 1 · Full Technical Guide
The Crucible Rating (CR) system is a modified Elo rating system, the same foundational model used by FIDE chess, FACEIT, and many competitive platforms worldwide. Every team has a single numerical rating that goes up when they win and down when they lose. The size of the change depends on how likely that outcome was before the match started.
The core idea is simple: beating a stronger team earns more CR than beating a weaker one, and losing to a stronger team costs less than losing to a weaker one. This keeps the leaderboard accurate over time and rewards teams that challenge themselves.
Before any CR changes, the system calculates the expected probability of each team winning. This is based purely on the difference between the two teams' current CR ratings using the standard Elo formula.
The result is a number between 0 and 1; think of it as a win probability. A value of 0.75 means the system expects that team to win 75% of the time given their ratings.
A 200 CR gap makes the higher-rated team a 76% favourite. A 400 CR gap makes them a 91% favourite. Equal ratings give each team exactly a 50% expected score.
Once the expected scores are known, the actual CR change is calculated. The actual outcome is either 1 (win) or 0 (loss). The difference between the actual outcome and what was expected, multiplied by the K-factor, determines how much CR moves.
The key insight is (Actual − Expected). If Alpha wins (as expected at 76%), the surprise factor is small: 1 − 0.76 = 0.24. If Beta wins the upset, their surprise factor is huge: 1 − 0.24 = 0.76. This is what makes upsets worth so much CR.
Notice how an upset is worth three times as much CR as a routine win. When both teams use the same K-factor, the CR one team gains is the CR the other loses. When a provisional team (K=50) plays an established team (K=32), the provisional team moves further: in an even match, the provisional winner gains 25 while the established loser drops 16.
The K-factor controls the maximum CR swing per match. A higher K-factor means ratings change faster; a lower one means they are more stable. The Crucible uses two values depending on how established a team is.
The Dominance modifier is a multiplier between 1.0 and 1.2 that grows with the margin of the win. For match requests it is calculated automatically from the confirmed series score, so no one enters it by hand. A one-round win is 1.0, a sweep is 1.2, and the scores in between are spaced evenly. Results recorded directly by staff use the dominance staff enter.
Casual matches don't change CR, so dominance only matters for ranked ones. Misreporting the score is treated as a false report under the rules.
CR is always a whole number; results are rounded to the nearest integer after each calculation. CR cannot drop below 0 regardless of how the formula resolves. This means a team at 5 CR that loses a match will drop to 0, not into negative values.
Here is a complete end-to-end calculation for a match between two established teams (K=32) where the lower-rated team pulls off an upset with a dominant performance.
Vortex gains 29 CR for pulling off the upset with a dominant performance, nearly four times what Storm would have gained for the expected win at standard dominance. Storm drops 29 CR, a significant penalty for losing to a lower-rated opponent convincingly.
The Crucible operates in seasons, and CR is reset at the end of each one.