
Embark has begun testing a reworked matchmaking algorithm in ARC Raiders, one that takes into account how individual players behave in solo, duo and trio queues rather than treating every raider the same. The studio has paired the announcement with a caveat: finding a match may now take longer.
Starting today, we'll be testing adjustments to our matchmaking algorithm aimed at improving the quality of your matches. One thing to note: the algorithm may take a little longer to find a match, since it's now working harder to find you a better one.
The stated goal is to group players whose approach to a raid resembles that of the people they are dropped in with, and to do so per squad size. Because the system reads behaviour separately across queues, the same account can be sorted one way in solos and another way in trios, which Embark says should make for "more fair and enjoyable raids."
The queue-time trade
That extra sorting work is where the cost lands. A matchmaker with more constraints has a smaller pool of acceptable matches at any moment, and the wait grows accordingly. Embark has acknowledged that the effect will not be evenly felt, flagging that "there is room for improvement, especially during low-population hours, in certain regions, or in specific map conditions," and saying it is working to improve those cases.
Why not just split the modes
The change lands in the middle of a long-running community argument about whether Embark should separate PvE and PvP into distinct playlists instead of tuning matchmaking around player behaviour. The studio has kept the extraction shooter's mixed lobbies intact, which means a player sorted into a predominantly co-operative match can still run into someone hunting other raiders.
Embark has not given a timeline for how long the test will run, or said what metrics would decide whether the adjustment stays. For now the change is live for all players, and the practical question is how much additional queue time the community will accept in exchange for lobbies that are better sorted.
Join the conversation — sign in to comment.
Sign in




