MMR (matchmaking rating) is essentially a measure of skill used to ensure fair matches in multiplayer games: gathering players with similar MMR helps to avoid vicious hits that can occur from random matches via server browsers. It’s not something games typically display to players – actual MMR ratings are often invisible data points that exist behind the scenes – but a group of Halo fans recently figured out how to track a player’s MMR. individual in Halo Infinite and discovered how playing the game in the casual modes can actually wreak havoc on your ranking experience.
The post explaining the process on Reddit states that while team performance and average MMR data is provided by 343’s servers, individual MMR data is not included. Free-for-all games, however, essentially count each individual player as part of their own team, so the data that team MMR data servers spit out can be linked to an individual player in an FFA game.
“This means that with the way data flows from Halo Infinite now, every time you play an FFA game, you can see what your MMR is from when you finish that game,” wrote redditor donutmonkeyman . “At the moment, that’s the only way we know of to find that value, because we don’t think that’s something 343 wants people to enter.”
This is all pretty arcane, and knowing your exact MMR isn’t particularly helpful when the game already ranks you between Bronze and Onyx, but its potential value as an analysis tool is where the discovery gets really interesting. . The post notes that some players on the Halo subreddit have complained that doing well in bot-based playlists artificially inflates their MMR, forcing them to be thrown into highly-skilled lobbies in ranked games they are not at all in. prepared. Apparently, the reverse can also happen: players can prevent their MMR from getting too high by starting matches in unranked playlists, making it easier to navigate the rankings.
Being able to see individual MMR allowed the group to confirm that this was actually happening: one of their friends is a regular Big Team Battle player who had a “horrible experience” when trying ranked play, and after upon reviewing his data, they determined that it was because his MMR was so high, despite not being a high rated player.
“Not only was he getting some of the most skilled lobbies as someone who wasn’t as familiar with those maps or those game types, but the TrueSkill system also expected him to ‘bag’ his teammates with lower MMRs,” donutmonkeyman explained. “If an average team MMR for him was 1600, he was on the high end, and the game expected him to perform extremely well. He also thought the game was a balanced game even though in reality this player had very little experience with the ranked playlist and the best way to play.”
Over a few weeks of ass kicking, his MMR started to drop from 2230 to around 1550, which led to better matches: he was improving in ranked play, but he wasn’t either. against professional-level players. But, as they noted, if he returns to Big Team Battle for a while, his MMR will climb again and he’ll be back where he started.
“This looks like a core element of the game that requires serious design thought, otherwise we’re concerned that the competitive integrity of ranked playlists will be compromised, and casual gamers interested in trying ranked will be turned away from the experience. very quickly,” donutmonkeyman wrote. “The current system, in our view, does not support the idea of players getting enjoyable matches in Ranked and Big Team Battles if they are interested in playing both.”
The analysis is imperfect, as MMR data only comes from FFA matches, and its link to CSR – the competitive skill ranking – which is visible to players and only applies to ranked matches is unclear. is not clear. Assuming these results are accurate, this indicates a potential problem for 343 Industries if not addressed. And if you just want to see where you stand against the rest of the playerbase, it’s really handy for that too.
I have contacted 343 Industries for comment and will update if I receive a response.