Marvel Rivals is very keen to remind you all that, aside from that whole bot debacle, it's not trying to make you feel better after having a string of particularly bad games.
That's per the game's official X account, which posted: "Hi Rivals! As our game continues to thrive, we want to reiterate that Marvel Rivals does not use EOMM. We are currently working on a video to demonstrate our developer insights on the matchmaking and ranking system, which is expected to be released next week." This was also preceded by a video on the Rivals Assembled YouTube channel, explaining the studio's intent to tell-all in the future.
EOMM stands for Engagement Optimized Matchmaking—as opposed to Skill-Based Matchmaking, EOMM is designed to, as the name suggests, optimise engagement.
You can see here a study by the University of California, Los Angeles if you want all the nitty-gritty, but the long and short of it is that EOMM doesn't just take your skill into account, but also your recent win/loss streak in an effort to reduce "churn", i.e. players putting the game down.
The study in question found that a trio of losses had a 5.1% "churn" rate—almost twice that of mixed results, such as two losses and a win, three draws, etc. Interestingly enough, it also found that a trio of wins produced a higher "churn" rate; while EOMM tries to knock you off loss streaks, it's also not particularly jazzed about you winning all your matches, either.
When implemented poorly, this results in "swingy" games, with EOMM serving as the pendulum. Winning too many games? You get put on a team that's at a severe skill disadvantage. Losing too many games? You're given the keys to the steamroller and told to have fun. All in the service of keeping your play hours up.
As for why Rivals felt the need to correct the record, just a single jaunt over to the game's recent "Mixed" reviews shows the EOMM-irage in action, with several calling it out as a direct contributor to their dissatisfaction.
One reviewer writes: "This game is fun but EOMM actually just ruins the game. It's either stomp or be stomped." Another adds: "Every game is a steamroll for one side and your actions feel useless," again, blaming the spectre of EOMM.
A straightforward search in the game's subreddit sees plenty of highly-upvoted threads reinforcing these suspicions. Here's one. Here's another. Here's a third. Here's a thread where the executive producer of Marvel Rivals, Danny Koo, outright stated in a Discord message that EOMM is a myth two months ago. The response was largely disbelief.
There's a couple of possibilities here—the least likely is that Marvel Rivals' devs are simply lying, mostly because there's no real benefit to doing so. The second is that there is some sort of EOMM-adjacent matchmaking at play that doesn't technically fit the definition. The third is that human beings are generally garbage at recognising probability.
I'm not saying that Rivals' matchmaking is perfect—I'm not familiar enough to make that call—but if you've ever felt a great sense of injustice at a 95% shot missing in XCOM, astonishment at rolling three natural ones in a row in D&D, or angry at pulling a terrible hand in Balatro, you've felt this yourself. You feel like the universe is conspiring against you, but improbable doesn't mean impossible.
Improbable things happen to us at an individual level all the time—on a micro scale we can feel hard done-by, especially if we're telling ourselves that a certain story (like EOMM existing) is the truth. On a macro-scale, however, you could simply be one of the (un)lucky ones. NetEase is the only player with the bigger-picture numbers, and I'll be interested to see if its explanations help quell doubts.
I understand why players have them, though. This is the game that grabbed names from your Steam friends list to populate bot games, after all.
