No. We just added number of wins as an additional "consideration" in the match making system.

I check the stats from that game, it looks normal. While the number of wins did have a larger difference than usual, the skill (elo) matching part are really close. Also, if you take a look of all other games you have played while the new MM system is on, you will found out the system is actually working effectively.

The key thing is this, when the system can't find suitable players with matching number of wins, it will try to expand its search range for number of wins, and concurrently, to find someone you can play with based on skill.

Excerpt from "Match-Making" systems.

http://na.leagueoflegends.com/learn/...ay/matchmaking

"Each player is assigned a number that represents his relative skill level, which is determined by the outcomes of his previous games".

So there is the potential for experience or raw data to play more of a part in match-makings systems than complex equations of skill?

Edit: The number of wins spectrum will increase rather than the skill spectrum.

I looked your records for games you have played since we turned on the system.
You have a net gain in ranked solo Elo, you also have a net gain in normal Elo.
The leavers you encountered were punished by LeaverBuster, and all other stats looks good.
The system(s) looks like are working you.

Number of games don't have anything to do with equal matches, how would using wins work any better?

If match-making already has a problem matching equally skilled players, any data from that system would be corrupted and the pool used would have to be small and recent using the new system which again would be terribly inaccurate.

This thread asked for feedback and mine is this:

Skill is not relevant to number of games played, match-making would do better as randomized off win:loss because this system corrects itself through trial and error over a longer time period than any other but almost assuredly.

I can provide a brief description but if you randomize a pool of games then a heap of players are 50:50 win/loss.

one hitting a win should not make it 51:50 but 49:51 in the system and matching them against a 49:51 player and winning, relevant to teams etc should result in 50:50 for the one exceeding odds and not give them a bonus rating because really all that happened is two equal players met, one beat the other, the player that player is matched with in future may have had the same meeting and player 1 beating player 2, then player 1 meeting a player 3 who bet player 4 and beating that player 3 shouldnt result in a bonus, merely redemption.

This way, adapting and learning is more at play than lucky snowballs, it would be more about fighting those who appear to be better at random stages, still getting points for beating the less skilled and maintaining your rating, not underestimating your opponents or the underdogs or those supposedly less skilled opponents because they'll all be fighting to beat you because there wont be a pedestal system, merely a struggle to maintain and gain.

But how is that acceptable that there is a level 28 there? Even if a level 28 player is 2K Elo rated he should not be matched with/against under ANY circumstances.

@Lyte

I'm in a game with a pantheon and a yi that are feeding on purpose, they look like (yet another) duo que. They literally just started running into mid tower from the start of the game.

I didnt get paired with these players before your matchmaking change, so I'm curious to know how exactly does everything appear to be normal on your end, it sure as hell doesn't on mine.

The game in my history is the one where I play miss fortune if you want to check it out.

p.s.
This duo que of feeders is refusing to surrender (im pretty sure they are duo que), which makes the game all the more fun, right?


edit. They're kill trading now so they're worth gold again. I'll attach a screenshot shortly Lyte, this is probably going to be a 1 hour loss (was 40 mins).

Here you go, enjoy it.

Why not? it is not ranked game, and if his skill is at that level that equals to his peers, then he should be matched.

Why?

2 separate ELOS.

If you're at 2k elo in non-ranked play and the level 28 is, assuming the pool to get to 28 is large enough (it is, only takes 3 games) then its a fair matchup.

Don't show them mercy if you realise they are lower level.

Don't look at the game after and base judgements off level and number of wins.

Look at their score and look at the teams and look at damage dealt and everything that happened in that one, solitary game, that's all that's relevant and that's all that you should be using to improve, learn and get better.

Don't look for excuses, look for their weakness, flaws and look at your own failures, misconceptions or mistakes, along with your teams possible flaws or errors, think about strategies or communications, items, roles or ideas that you could have mentioned, suggested or even performed yourself and above all, win.