How WoW PvP Rating Works: CR and MMR Explained

CR and MMR in one paragraph

Every rated bracket gives you two numbers. Current rating, usually shortened to CR, is the visible score on your character, the one that unlocks ranks and appears on leaderboards. Matchmaking rating, or MMR, is the score the game uses to decide who you face and how much your result is worth. It is not tracked on your character sheet or on the ladder, but the post-match scoreboard of a rated game shows it, so you can read it after the fact in every bracket. You control CR only indirectly: you win games, MMR moves, and CR follows it.

Almost every question players ask about rating comes from that gap. Winning a game and gaining a couple of points, losing far more than that for a single defeat, climbing quickly on a fresh character and then stalling: all of it is the same system trying to move your CR to where your MMR already is.

What the game is actually measuring

Rated PvP uses an expectation model. Before a match starts, the system compares the matchmaking ratings on both sides and forms an expected result. What you are rewarded for is not winning, it is beating the expectation.

Win against a team the system expected you to beat and you have confirmed what it already believed, so the adjustment is small. Win against a team it expected to beat you and the model was wrong about you, so the adjustment is large. Losses invert the same logic: a loss to a much stronger opponent costs little, and a loss to a weaker one costs a lot.

In an evenly matched game the movement is modest, in the region of fifteen points, and it grows or shrinks from there with the gap between the two sides. Specialist rating guides also describe an additional multiplier that starts from the fourth consecutive win or loss, so long streaks move both numbers faster than the individual results suggest. Blizzard does not publish the formula, so treat every specific number as an observed pattern rather than a rule.

The block above is the fastest way to see the model at work. Watch the size of the gains and losses next to the ratings they happened at, and note how rarely two players at the same rating move by the same amount.

Why CR chases MMR and then stops

The design intent is that CR ends up mirroring MMR. Blizzard has treated CR rising above MMR as a defect rather than a feature: when a bug once let personal ratings climb past matchmaking ratings, it was hotfixed out.

That gives your climb two distinct phases, and knowing which one you are in explains most of your point gains.

Catch-up. While your MMR sits above your CR, the game is trying to close the distance. Wins pay more than they normally would and losses take less, which is why a strong player on a new character can climb several hundred rating in an evening.

Equilibrium. Once CR reaches MMR the subsidy is gone. Now you gain roughly what you lose, and further progress requires beating opponents at your own level often enough to push MMR up first. The plateau most players hit is not a hidden cap, it is the moment the catch-up phase ends.

Why you and your team mate gained different amounts

Rating is personal even in a team bracket. Every participant carries an individual current rating and an individual matchmaking rating, and the adjustment is calculated for each player separately after the same result.

Two players who queue together and win the same game will therefore often see different numbers. The one whose matchmaking rating sits furthest above their current rating gains the most, because that player has the most catching up left to do. The one whose two numbers already match gains the least. Nothing about the match itself changed between them.

This also explains a pattern that looks unfair from the outside. A player returning to a bracket after a strong previous season can climb faster than a team mate who has been grinding all along, not because the game is favouring them, but because the matchmaker is still holding a higher opinion of them than their visible rating reflects.

Where your rating starts each season

Every player begins a new season at 0 CR, regardless of what they finished the previous one with. A fresh character at max level enters matchmaking around 1500 MMR, which is why the first games of a season feel far harder than a rating of zero would suggest: your visible number starts at the bottom while the matchmaker places you in the middle.

Those opening games work like a placement phase. Your CR sprints upward because it is far below your MMR, and the rating you settle at is the system's first estimate of your level. Blizzard does not publish how MMR carries between seasons, so the safest assumption is that your early opponents reflect your recent history rather than your empty rating.

Two practical points follow from that. Those first games are the cheapest rating you will earn all season, so it is worth playing them with a prepared setup rather than treating them as a warm-up. And a bad opening run is not a season ending problem: while your current rating sits below your matchmaking rating, the system is still paying you extra to correct itself.

Every bracket keeps its rating separately

Rating is not one number on your character. Each bracket tracks its own CR and MMR, and two of them go further and track a separate rating for each specialization.

The practical consequence is that switching specialization in Solo Shuffle or Blitz does not move your progress with you. You start a second ladder entry from scratch while the first one stops improving. Each specialization is then evaluated on its own ladder at the end of the season, so splitting a season's games across two of them can leave neither one high enough to qualify for anything.

Solo Shuffle: six rounds, one verdict

Solo Shuffle puts six players through six rounds, rotating the teams so that everyone plays with and against everyone else. Your rating change is not decided round by round in a normal lobby. The system looks at the whole lobby, at the matchmaking ratings involved and at which side won each round, and compares your final record against what it expected you to do.

This is why a 3-3 finish is not automatically neutral. Three wins in a lobby where you had the lowest matchmaking rating beat the expectation. The same three wins in a lobby where you were the strongest player fall short of it. The scoreboard at the end of the lobby lists a matchmaking rating for all six players, so you can check which of the two situations you were in instead of guessing from the rating change alone.

Incomplete lobbies are handled proportionally. When a player leaves, gains and losses are scaled linearly to the number of rounds that were actually played, and leaving affects both MMR and CR.

Compare the spread here with the 3v3 block earlier. Solo Shuffle results cluster differently because a six round record produces finer gradations than a single win or loss.

Blitz and rated battlegrounds: team results, personal ratings

Battleground Blitz is a solo queue mode with a team result. Rating is stored per specialization as in Solo Shuffle, but the unit of measurement is the match: your side wins or loses, and the adjustment follows the matchmaking ratings on both teams.

Rated battlegrounds work the same way with a premade group and one rating per character. In both modes personal performance inside the match does not change your rating. Damage, healing and objectives affect whether you win, and nothing else.

That distinction is worth stating plainly, because it is the single most common misconception carried over from other competitive games. The scoreboard is not an input. A healer who kept the team alive for twenty minutes and a player who spent the match dead receive the same adjustment from the same defeat, and the only lever either of them has over their rating is the win.

The one place performance enters the system indirectly is Solo Shuffle, where your own play across six rotating team compositions decides how many rounds you take. There is still no scoring of your damage or healing; it is simply harder to hide a weak game when you are the only constant across every round.

What each rating threshold unlocks

Ranks are fixed. Unlike a cutoff, a rank is a rating you either reach or do not, and it is the same number in every rated bracket this expansion.

Reaching a rank unlocks its seasonal rewards, and dropping below the rating afterwards does not take them back. This is the opposite of how title cutoffs behave, where only your standing at the end of the season counts. If you are chasing a rank, the goal is to touch the number once. If you are chasing a title, the goal is to be above a line on the last day.

Rating inflation and why the same number means less in month three

Rating is not a fixed scale. Blizzard applies inflation to matchmaking ratings over the course of a season and has adjusted the weekly rate mid-season more than once. In October 2024 it raised the rate in Solo Shuffle, 2v2 and 3v3, stating that the goal was to bring arena ratings closer to the rewards then available in Battleground Blitz. The exact rate in any current season is not published.

Two things follow. First, a rating achieved in the opening weeks represents a harder result than the same rating achieved near the end, because the whole ladder has drifted upward in between. Second, and more usefully, waiting is not a strategy but neither is panicking: as the ladder inflates, the rating needed for a fixed rank becomes progressively easier to hold, while the rating needed to stay above a percentage based cutoff keeps moving.

If your goal is a rank, later in the season is the cheaper time to reach it. If your goal is a title, see how cutoffs are calculated, because inflation works against you there.

Common misreadings of the rating system

Assuming a win should always pay. A win against a lower rated opponent confirms the expectation and pays accordingly. The size of the reward is set before the game starts, not by how the game went.

Blaming a hidden cap. There is no rating wall. What feels like one is CR arriving at MMR, which removes the catch-up bonus you had been climbing on.

Chasing MMR by dodging. Leaving a Solo Shuffle lobby moves both MMR and CR, and incomplete lobbies are scored on the rounds that were played. Dodging is a rating decision, not a free reset.

Reading one game as evidence. Single results are noisy by design, because the model is measuring you against an expectation that changes every match. Ten games tell you something, one does not.

Treating specialization ratings as one rating. In Solo Shuffle and Blitz each specialization is its own ladder entry and is judged on its own at the end of the season. Two half seasons on two specializations are not added together.

Comparing ratings across brackets. 1800 in Solo Shuffle, 1800 in 3v3 and 1800 in rated battlegrounds describe three different populations. Rank rewards are aligned, difficulty is not.

How to check where you actually stand

Your rating on its own says very little. Three views make it meaningful. Start with the rating distribution for your bracket, which converts your number into a position among everyone playing it, on the EU 3v3 distribution or the US 3v3 distribution, switching the bracket to the one you play. Then look at recent rating activity for your region, EU or US, to see the size of the gains and losses happening at your level right now. Finally, if a title is the objective, check the current cutoff for your bracket and specialization, because that line moves independently of your progress.

The top of the ladder is useful for one specific purpose: it shows what the ceiling of the current season looks like in absolute numbers, which is the context that makes your own rating readable.