EOMM is a matchmaking framework that optimizes player pairings to maximize engagement and reduce churn. The concept originates from a 2017 academic paper using anonymized EA data. Three components drive the framework: a skill model, a churn prediction model, and a graph matching model. The optimization target is aggregate retention across the match pool.
How EOMM Works
The skill model estimates each player’s ability from match history. The churn prediction model flags early player churn indicators, including losing streaks and shortened sessions. The graph matching model then pairs players across the full pool to maximize predicted engagement in aggregate. The result is a population-level match assignment, optimizing for collective engagement across all active sessions.
EOMM vs. SBMM: Key Distinction
SBMM (Skill-Based Matchmaking) pairs players of similar ability to produce competitive fairness. EOMM targets a different objective: predicting which pairings keep the most players engaged across the longest sessions. A matchmaking system can incorporate both signals simultaneously. The two frameworks address different optimization goals and can coexist within a single title.
EOMM in Academic Research vs. Live Deployment
The 2017 Chen et al. paper demonstrated EOMM as a mathematical optimization framework, using anonymized EA data in a controlled study. Whether any studio has since deployed that framework exactly as described remains an open question. EA and other developers have publicly denied active use. Player-perceived patterns such as win streaks followed by losses can also emerge from normal statistical variance.
Benefits of EOMM as a Design Framework
As a design philosophy, EOMM can protect new players from early losing streaks. Early losing streaks are among the strongest drivers of churn before players invest in a title. Broader session lengths follow when more players encounter varied, competitive outcomes. Studios pursuing this approach target lower churn and higher player LTV at the population level.
EOMM in Live-Service Games
Live-service studios using retention-aware matchmaking still generate player frustration that surfaces in support tickets and community channels. “EOMM demon” describes a player whom the community perceives to be on a manipulated losing streak. Recognizing this shorthand helps support and community teams accurately read sentiment spikes. Studios can use Helpshift’s Care AI to monitor matchmaking-related frustration at scale, across 75+ languages.
If matchmaking frustration is surfacing in your support queue or community channels, see how Helpshift helps studios turn those signals into actionable retention intelligence.
