North America - English

League of Legends Community

League of Legends Community > League of Legends > General Discussion
Help Riot improve matchmaking! Looking for examples of bad matchmaking

View Poll Results: How many games do you expect to play before you reach your true Elo?
< 10 Games 171 1.59%
11-40 Games 798 7.41%
41-80 Games 1,145 10.64%
81-120 Games 1,710 15.89%
121-160 Games 649 6.03%
161-200 Games 904 8.40%
200-399 Games 1,668 15.50%
400-599 Games 1,055 9.80%
600-799 Games 444 4.13%
800-999 Games 166 1.54%
1000+ Games 2,052 19.07%
Voters: 10762. You may not vote on this poll

First Riot Post
 
Comment below rating threshold, click here to show it.
Pentastik ?? Member
This user has referred a friend to League of Legends, click for more information This user has accepted the summoners code, click for more information
03-08-2012

retarded matching system.. most player feels bad with this ugly matching

can we expect new system soon or not?

 
Comment below rating threshold, click here to show it.
Kshatriya ?? Senior Member
This user has referred a friend to League of Legends, click for more information This user has accepted the summoners code, click for more information
03-08-2012

Chances are Matchmaking won't see a change until S2 ends. That's where my money is at.

 
Comment below rating threshold, click here to show it.
Jekyll79 ?? Senior Member
This user has accepted the summoners code, click for more information
03-08-2012

I am 1500+ elo in solo que. When I do normal premades with friends who are a lot lower elo than me(1200 or less) the games are stacked against us. I had one game where my 1200 elo friend was laning against a 2200 player...that one lane lost us the game before 8 minutes in. This is not isolated, lately all my premade games have been like this to some degree to where I can't get a decent game with a premade team of 4 to 5. In these games I am only slightly better than the other laner/jungler but my friends are way worse than the other laners who are anywhere from 200 to 1000 elo higher.

 
Comment below rating threshold, click here to show it.
EMeta ?? Senior Member
This user has referred a friend to League of Legends, click for more information This user has accepted the summoners code, click for more information
03-08-2012

Quote:
Originally Posted by Lyte View Post
Do not use bots or third-party tools to find out Normal Elos in SR or Dominion. They are not even close to accurate.

To answer the question, we have not decided if we are changing punishments for queue dodge, just discussing it at the moment.
Regarding the Dominion semi-hidden Elo, could you be more specific on its current limitations? As another poster said above, it's all we have. I assume you don't mind players having access to it or you'd hide it like you do SR normal Elo.

 
Comment below rating threshold, click here to show it.
Devare ?? Senior Member
This user has accepted the summoners code, click for more information
03-08-2012

Quote:
Originally Posted by Lyte View Post
The point is not to just punish players who queue dodge. We are interested in why players are queue-dodging, and then determining if we can fix the source of the problems. We are simply discussing whether the current penalties for queue dodging is appropriate while we are investigating the causes of queue dodging.
It's simple why players queue dodge.. They queue dodge 95% of the time because they don't like the team set up because 1 person or more picked a "troll" pick or a "bad" pick and they'd rather wait 5+ minutes in hopes of getting a better team comp. The remaining 5% dodge because someone (generally 1st pick in ranked) took a role or champion that they wanted to play which in turn leaves them to play support majority of the time; which they don't want to do so they dodge. Spend 1-2 weeks in the ~1k-1500 elo range and you will see this on a daily basis.

 
Comment below rating threshold, click here to show it.
Just a Nick Name ?? Senior Member
This user has accepted the summoners code, click for more information
03-08-2012

For start I just want to tell you that after community has been *****ing for over a year that matchmaking sucks, you finally decided to do something! But I guess better late than never.

Even thou I am not really happy with you taking so long to react, I appreciate your effort for doing so at last.

Here is my honest opinion of your matchmaking system and my idea how to improve it.

For start it is a team based game, and as such placing random people into a group they are forced to play in is wrong.
Judging their skill level based on their performance in totally random team is wrong.
Expecting people to be completely polite to other team mates who are forced to play with you as well as you are forced to play with them is wrong.

What you created is easily described as "putting 5 rabid cats into a small bag" and expect them to behave and work together. Well, it won't work, and you can see it by doing above mentioned experiment, or simply monitor players matched by your system.

And this is how I see solution to this problem.

Instead of shoving people to play with each other even if they obviously can't bare each other, you make a group of 5 people before you match them against other 5 people. People are being matched with other players with similar elo and number of games.

That group, of 5 random people, has infinite time to chat about what they want to play, how and why. Everyone can do a background check of other players, and see their match history, stats, and so on... During that process, players unsatisfied with a match, can leave the team, without any penalty and rejoin another random group of 5 people.

Ofc, person that left can't be matched with other 4 people in group for another 30 minutes. To prevent awkward situations off being matched twice in a row with unwanted people.

Also you are able to "tag" a single person in a team you are matched with a "troll marker". Person that gets tagged with troll marker 10 times during 24 hours gets a 24 hours ban. One marker gets removed each hour. This way you are able to remove trolls before even being matched against other team. And trolls can't exploit it and report random people since they can't be matched with same people after they left for next 30 minutes, and are able to tag only one person per matched team.

So once random people team found themselves, and managed to get along and organize a team build, they are then being matched with equally strong opponents. They do so by clicking "i'm ready" button. Once all 5 are ready, matchmaking starts.
Equally strong opponents are team with as similar average elo and average number of wins as possible.

This way players are able to create team setup they think it is the best, and get matched against equal team that made also setup that they think it is the best. So no one can ***** about i got matched with idiots, team setup sucks, because you agreed that team is good.

Same can be applied for ranked games, except people don't make team straight away, but instead talk which role will be for who, and who can play what. Person with highest elo is leader, but that doesn't matter since people probably agreed who will get banned and why.
Once matched with opponent team you have a standard draft pick system.

People who leave game once they agreed to play with the team and are matched with opponent team gets a 24hours ban and reduced his elo by 100 if ranked game. This way people would think twice to leave the game once matched, and will they play with the team or not.

This way players are able to choose who to play with and let them organize before the match starts. Whole process would take some time at start, but people would get used to it really fast thus increasing the whole process.
This way to get in a game would be a bit longer, but games would be far better for both teams, and better balanced.

After all this is a team game, so let people choose who they play with. I'm quite sure that people would enjoy it way better if they get familiar with each other before the game actually starts.
Not to mention that Ignore list would actually do something, since you would see you are matched with a person from your ignore list, and after some time, most of vile and toxic players would be filtered out, because no one would play with them. So they are forced to assimilate to team environment or never play with people again.

Yes, i am willing to sacrifice 30minutes to get a team I want to play with, then get placed in a team with totally random idiots, and then suffer in game for the same time + lose elo and nerves!

Just my idea.

Cheers!

 
Comment below rating threshold, click here to show it.
SomaLOL ?? Senior Member
This user has accepted the summoners code, click for more information
03-08-2012

The problem with matchmaking is that we are playing a game based on team cooperation, and ranking the people individually, when this is not the proper idea of "elo", since giving an increased or decreased elo to someone based on the team performance instead of individual performance will never lead to their true elo until 1000+ games.
While ranking someone for their individual role is really difficult, since not only the KDA, CS or Turrets destroyed affects on this, but also the style of play, passive or aggressive, taking kills for someone or tanking, or just supporting.

Therefore and idea that comes to me would be letting the team qualify the performance of themselves, giving a score of "4,3,2,1" to each of the other members of their team (4,3,2 for duo queues), plus an overall performance of KDA on the match, this way, when a really bad player would be qualified by 1 (hopefully) by every other member of the team, and really good players would be qualified by 4 by them, and the performance given by the KDA could give a +1 to + 5 for better players or a -1 to -5 to bad or not that good players.... the KDA ratio would help to predict if the points given by the team are not off the board, like if a 10/0/4 player got qualified with 4,4,1,1 by 2 mad players, this would not affect in the plus given to their elo, or a suport that went 0/2/18 qualified with 1,1,2,1 just because he rolled suport (wich hopefully wont happen) would not have a reduce on his elo.

The problem of letting people qualify their own team would be the trollin, but this wont happen a lot just because 3 or more people trolling the same game is kind of hard, and duo queues wont have the chance of qualify themselves, also you are forced to choose a good player and a bad players.

This is just an Idea I hope you can read it Lyte.

 
Comment below rating threshold, click here to show it.
Lyte Lyte's Avatar ?? Lead Designer of Social Systems

Follow RiotLyte on Twitter
This user has referred a friend to League of Legends, click for more information This user has accepted the summoners code, click for more information
134 of 173 Riot Posts
03-08-2012

Quote:
Originally Posted by SomaLOL View Post
The problem with matchmaking is that we are playing a game based on team cooperation, and ranking the people individually, when this is not the proper idea of "elo", since giving an increased or decreased elo to someone based on the team performance instead of individual performance will never lead to their true elo until 1000+ games.
While ranking someone for their individual role is really difficult, since not only the KDA, CS or Turrets destroyed affects on this, but also the style of play, passive or aggressive, taking kills for someone or tanking, or just supporting.

Therefore and idea that comes to me would be letting the team qualify the performance of themselves, giving a score of "4,3,2,1" to each of the other members of their team (4,3,2 for duo queues), plus an overall performance of KDA on the match, this way, when a really bad player would be qualified by 1 (hopefully) by every other member of the team, and really good players would be qualified by 4 by them, and the performance given by the KDA could give a +1 to + 5 for better players or a -1 to -5 to bad or not that good players.... the KDA ratio would help to predict if the points given by the team are not off the board, like if a 10/0/4 player got qualified with 4,4,1,1 by 2 mad players, this would not affect in the plus given to their elo, or a suport that went 0/2/18 qualified with 1,1,2,1 just because he rolled suport (wich hopefully wont happen) would not have a reduce on his elo.

The problem of letting people qualify their own team would be the trollin, but this wont happen a lot just because 3 or more people trolling the same game is kind of hard, and duo queues wont have the chance of qualify themselves, also you are forced to choose a good player and a bad players.

This is just an Idea I hope you can read it Lyte.
The problem with individual performance metrics is that it is completely against the spirit of the game. Individual performance metrics encourage players to be selfish and pad their own statistics in the hopes that they get more bonus Elo, or reduced Elo loss. With your specific suggestion, very few support players will be voted as the 'best' player on the team. For many players, seeing a stat line of 0-2-10 is not a positive that ever overrides the perception of seeing a carry going 10-2-2.

Being a 'skilled' player in MOBAs is not about being a great individual player--it is about your ability to contribute to your team's win. It is about your ability to be an awesome leader for your team, or an awesome teammate. I would rather see team-oriented metrics where we encourage the team as a whole to perform better to get small adjustments to their Elo gains/losses as a team--not to any individual player.

 
Comment below rating threshold, click here to show it.
The Crow ?? Senior Member
This user has accepted the summoners code, click for more information
03-08-2012

Quote:
Originally Posted by Lyte View Post
The problem with individual performance metrics is that it is completely against the spirit of the game. Individual performance metrics encourage players to be selfish and pad their own statistics in the hopes that they get more bonus Elo, or reduced Elo loss. With your specific suggestion, very few support players will be voted as the 'best' player on the team. For many players, seeing a stat line of 0-2-10 is not a positive that ever overrides the perception of seeing a carry going 10-2-2.

Being a 'skilled' player in MOBAs is not about being a great individual player--it is about your ability to contribute to your team's win. It is about your ability to be an awesome leader for your team, or an awesome teammate. I would rather see team-oriented metrics where we encourage the team as a whole to perform better to get small adjustments to their Elo gains/losses as a team--not to any individual player.
What about using individual performance metrics without telling anyone you are doing so.

Maybe even just using it as a small percentage of the total calculation.

 
Comment below rating threshold, click here to show it.
Lyte Lyte's Avatar ?? Lead Designer of Social Systems

Follow RiotLyte on Twitter
This user has referred a friend to League of Legends, click for more information This user has accepted the summoners code, click for more information
135 of 173 Riot Posts
03-08-2012

Quote:
Originally Posted by The Crow View Post
What about using individual performance metrics without telling anyone you are doing so.

Maybe even just using it as a small percentage of the total calculation.
Players are smart Given that Elo adjustments in Ranked are visible, players will figure it out.