And MM can't determine how one person is going to play on any given day much less how 12 will play together and eventually how 24 will play together in a game. You somehow have to figure out how to define 'skill'/'performance' and that is a huge issue as there aren't any really good answers to the questions - a crap ton is ship/upgrades/captain points skills and how good the player is AT THE TIME OF PLAY. After all, WG has pretty much have all the data regarding over-pushes (die early, consistently) and snipers (live to the end and probably lose more often than not.) Shouldn't that be incorporated into the MM algorithm to even each team's odds? OK, but I find it hard to believe that MM can't be tweaked to account for player performance issues that you describe. Gun accuracy is certainly skewed by class of ship, and does not encompass bad/good DD play, which also determines a lot of matches. PR seems skewed by being on the winning team. This is a problem, but being realistic I do not know how a fix would manifest: if one went on WR it would help, but the more even the spread on “balanced teams” will then start to decrease the spread on WR. For many excruciating minutes they collectively got a total of 3 hits through countless single turret shot sequential firing. But both were sub 20% main gun accuracy BB mains. Despite bad tactics early, the two potatoes found themselves in perfect position to finish off the red BB and thus win the match. In a classic “snatching defeat from the jaws of victory” match recently 2 full HP BBs ( Colorado & Gneisenau) had a crossfire on a 1/3 HP Scharnhorst. Sadly, there are players that would appear to “not zoom in” when shooting longer range shots. The mix is similar when considering Ship WR, Ship PR, and even main gun accuracy. The losing side always has 5-6 sub 47% win rate guys, while the winning side has usually only 1-2.

At least, that's how it feels.įor kicks, many months ago, I looked up stats on several matches. Assuming the players don't all suck that bad (I can hit 100k when properly supported), I'm forced to conclude that MM is building 'bad' teams. That tells me the other players are REALLY bad. I can't speak for how the red stats lose when we streamroll them, but on my end, I can routinely put out poor numbers (say, 20k) and still be in the top 4.

World of warships matchmaking meme