Honestly the fact that it is 496 - 500 despite the fact they had two duos might actually mean it is working pretty well since it managed to balance out individual skill-rating with duo-skill-rating. Yeah, i think your kind of right about that. Would be interesting to know how much the skillrating is raised trough a duo. Or maybe it is not?Also its quite strange that the matchmaking has to do something like this. Would have been better to divide the duos, right? Unless the Skill rating-difference of one or two players were just way to big. So the question should be why these players even had to play in that rating-"class". The matchmaker used to work harder to spread out duo's. However, based on our data, skill rating is generally a better indicator of potential performance than whether someone queued up in a duo. So the matchmaker puts the greatest priority on making sure the average skill rating on each team is as even as possible. It also tries a bit to make sure the standard deviation of skill rating of each team is pretty close. In addition to these, the matchmaker will also do it's best not to put any more than +1 of a class on a team. Though, players can work around that themselves with character swapping. Something we've decided to leave in, based on community vote and because we think it's ok if players/teams who are skilled or knowledgeable enough to recognize bad compositions can resolve that themselves. I've said it many times and I'm sure people still won't believe it, but the match maker has no idea if you have won or lost your previous match. All it knows is your current skill rating, your class, and whether you're in a duo queue or not.Thx for the Information, the mechanic sounds reasonable as long as the data does not say anything else about the advantages trough duos. I also like the fact that this should make people able to play duo with more fun, even if their skill is quite different.