My impression is that matchmaking returns a lot of uneven teams that led to non-competitive matches. I made this thread because a/ I want to see if others also believe that's the case, and b/ I wanted to test it objectively, using Control scores (which I play the most). So, first up, do people find matchmaking generates even teams or not? Second, my test. My belief is that competitive matches are usually reflected in score, with odd exceptions. I made a few categories to try to put in numbers what I consider even. I believe good matchmaking would result in plenty of the first three categories and few of the last ones. - [i]Close[/i] matches I defined as having a difference in score of 1 000 or less. These matches could easily have gone either way. - [i]Competitive[/i] matches were up to 2 000. - [i]Mostly Competitive[/i] matches were up to 3 000. - [i]Decisive[/i] matches had a difference up to 4 500. At this point, matches are no longer very close. - [i]Uneven[/i] matches had a difference up to 6 000. - [i]Blowouts[/i] had differences past 6 000. These probably frustrated losers and bored winners; should be rare. My test was simply to record, to the nearest hundred, the difference in score between winning and losing teams in 40 consecutive Control matches from a random chunk of my Crucible history. I then found the average difference, and slotted in all the scores into my categories to find out how many matches were close, and how many weren't. Results: [b]The average difference between the teams was 4 600 points[/b]. The breakdown into categories: [b] - Close: 2 - Competitive: 6 - Mostly Competitive: 6 - Decisive: 8 - Uneven: 8 - Blowouts: 10[/b] In other words, 18 of 40 (45 %) of matches were decided by more than 4 500 points, which is a difference in scores of a fifth or more. One quarter of matches ended in blowouts. 20% of matches were highly competitive (a max difference of 2 000 points). I think this suggests that Crucible matchmaking currently needs some real work to return better team parity. It's not accounting for skill well, if at all. Notes: A few things. First, I'm [i]not[/i] a great Crucible player - I freely admit that. My K/D ratio is 1.0 and change, and I win around 51 per cent of my Control matches, almost all of which are with random players. I don't carry a team or sink it. I'm in there to get stuff for PvE, but I do try to help the team. I don't believe my own contribution skews the results. Second, this isn't a huge sample. For a sample this size, the margin of error is something like 15 %, plus or minus. Additionally, I'm playing from Asia, which may have some impact (players often play with people outside their country, meaning latency; it's not as popular a genre here; etc.).