Marvel Rivals Confirms No Use of EOMM, Explains Matchmaking in Detail

Loki in Marvel Rivals (Credit: Netease)

Marvel Rivals developers have officially confirmed that the game does not use the controversial Engagement Optimized Matchmaking (EOMM) system, following heavy criticism and speculation within the player community.

Earlier this month, NetEase Games addressed concerns surrounding matchmaking by clarifying that Marvel Rivals’ competitive system does not rely on EOMM. Instead, the studio promised to break down how the ranking and matchmaking systems actually work.

That detailed explanation has now been delivered in a video posted on X (formerly Twitter), where Lead Combat Designer Zhiyong offered a closer look at the game’s approach to competitive balance, performance tracking, and player rankings.

The denial of EOMM is particularly significant because many competitive titles in the past have been accused of using such systems to artificially maintain a 50% win rate. According to Zhiyong, Marvel Rivals’ matchmaking is based on measurable factors like competitive scores and performance rather than hidden engagement metrics.

How the Ranking System Works

After every competitive match, a player’s score is recalculated using two major components: Base Points and Performance Points. These two values are combined in different proportions depending on the player’s current rank.

Base Points

Base Points are determined by the outcome of the match (win or loss), the player’s current competitive score, and the average score of the enemy team. Winning against stronger opponents yields more points, while losing against weaker teams results in heavier deductions.

For example:

• If both teams average around 3500 points, a win gives +20 and a loss deducts -20.

• If you beat a stronger team averaging 3700, you earn +25 but lose only -15 if defeated.

• Facing weaker teams averaging 3300 flips that balance, with +15 for a win but -25 for a loss.

This structure ensures that climbing the ranks reflects not just victories, but also the relative difficulty of those matches.

Performance Points

Performance Points reflect how well a player contributes within a match. However, given Marvel Rivals’ diverse roster of heroes, the game doesn’t simply tally up raw kills or assists. Instead, stats are normalized per 10 minutes, then compared against the average performance of that specific hero at the same rank.

For example:

• A Diamond 1 player spends 4 minutes as Loki and 6 minutes as Invisible Woman in a 10-minute match.

• Loki's performance value: 1200, Invisible Woman's: 1000.

• Average performance value for both heroes at Diamond 1: 1000.

• Coefficients: Loki (1.2), Invisible Woman (1.0).

• Time-weighted overall coefficient: 1.08.

• Compared to team averages (1.0, 1.0, 0.8, 1.15), resulting in a final adjusted performance coefficient of 1.11.

This means excelling with a character like Loki is measured differently from excelling with Invisible Woman, ensuring fairness across diverse playstyles.

Balancing Base and Performance Points

The game then combines Base and Performance Points, but the weighting changes by rank. At lower ranks, performance carries more weight, helping skilled players climb quickly. At higher ranks, base points matter more, rewarding consistency over individual standout performances.

For example, if a player earns 20 Base Points and 30 Performance Points in a win:

• In Silver (base point weight: 0.4, performance point weight: 0.6), the final score gained is 26 (20 0.4 + 30 0.6).

• In Celestial (base point weight: 0.7, performance point weight: 0.3), the final score gained is 23 (20 0.7 + 30 0.3).

This tiered system helps newer players rise faster while ensuring high-level competition demands sustained success.

Matchmaking Mechanics

Marvel Rivals’ matchmaking system is designed to balance fairness and queue times. According to Zhiyong, the algorithm considers server selection, team composition, and competitive scores. If matches take too long to find, the system gradually widens the acceptable rank range to secure players quickly while still avoiding extreme mismatches.

Importantly, team matchmaking is size-based. A group of six will only face another six, while three-player squads are matched against similarly sized teams when possible. However, as wait times increase, the system may mix groups into configurations like 4+1+1 versus 3+2+1.

Zhiyong admitted that while this helps reduce queue times, it can occasionally result in rank gaps or imbalanced team roles. For example, one side might have a full set of role-specialized players, while the other ends up forcing someone into an unfamiliar role.

Restrictions in Season 3.5

One of the most noticeable recent changes is the restriction on four-player and six-player teams in higher ranks. Starting in Season 3.5, these team sizes are banned at higher competitive levels because they made balanced matchmaking too difficult. NetEase explained that this temporary measure allows the team to refine the system further while avoiding unbalanced high-rank matches.

Addressing Role-Based Matchmaking

Some players have asked why Marvel Rivals doesn’t use role-based matchmaking. The developers argue that this would dramatically increase queue times without guaranteeing better balance. Because the distribution of player roles is often uneven, strict role-lock queues could lead to extended waiting while still failing to ensure fair skill matchups.

Instead, the system emphasizes flexibility, trusting players to adapt roles when necessary, even if it occasionally results in awkward team compositions.

Why Win and Loss Streaks Happen

Another hot topic in the community has been win/loss streaks. Many players assumed these streaks were engineered by EOMM to maintain engagement. Zhiyong strongly denied this, explaining that streaks are a natural outcome of probability and matchmaking constraints.

Statistically, even in a pure 50-50 environment, consecutive wins or losses are common. For instance, across 10 games, the chance of at least three wins or losses in a row is about 50%. Add in factors like facing the same players multiple times in a region or role imbalances, and streaks become even more likely.

This statistical perspective, the team argues, explains why some players feel matchmaking is "rigged," when in reality it’s the expected outcome of probability across large player bases.

You can watch the explanation from Zhiyong below:

With NetEase pulling back the curtain on its competitive system, will this transparency be enough to win over a skeptical player base? Let us know your thoughts in the comments.