Why doesn't LoL have MMR (Matchmaking Rating)?

Comment below rating threshold, click here to show it.

Krshna

Senior Member

07-12-2011

Quote:
Originally Posted by Laraki View Post
I think there would be plenty of truth to the part of this post about the "MMR" if there were a truely good and accurate way to rate such a thing in league of legends.

The only things i know of in league of legends that gives you some idea of how skilled a person is really, is a combination of their total games played (which shows you their overall time spent/experience in the game), looking at their tendancy towards wins/losses (shows you their effectiveness and to a degree their ability to team play) and to actually watch that player play to see how they function strategically, and i certainly dont know a way that that can be applied for all the players in League of Legends easily in an automated system...

And as far as hard elo resets "causing chaos" ... i reject this notion completely and will continue to do so as it is based on nothing as far as i can tell -> durring the reset period the best indicator of skill we have would be total games played, and i really dont believe it would be "chaos" for a person with a thousand games played and a past elo rating of 600, to be matched up with a person who also has a thousand games played to be matched with a person that had a past elo rating of 1700 for one game, the result of which would affect their new elo rating and help to place them into a better match making situation next game.....

I do also think the soft-reset serves very little purpose and is a mockery to the competitive nature of the game as you say, but i am hopefully it is Riot's way of testing the waters for Elo resets, though to my knowledge they havnt commented on it yet.
Yeah, to be honest I was shocked, I was sure they would do a hard reset... I agree that it makes a mockery of the spirit of competition. If you are just going to let some people start high, just because they were high at some point in the past...


Comment below rating threshold, click here to show it.

replayssoon

Member

07-12-2011

Quote:
Originally Posted by Krshna View Post
Yeah, to be honest I was shocked, I was sure they would do a hard reset... I agree that it makes a mockery of the spirit of competition. If you are just going to let some people start high, just because they were high at some point in the past...
How does it make a mockery of competition? It is actually the opposite. There would be no competition generated from a hard reset. Only high Elo players duo queuing with each other to get back to there normal place in the rankings, while all the low level people cry about how they are cheating.

A hard reset is not some magic wand that is going to make you better at this game. It is not going to make so you will magically jump from 1200 to 1600 rating. All a hard reset is going to do is force me and players like me to stomp a bunch of people and basically embarrass them to get back to my normal rating.

Also everyone is so focused on high Elo player and saying there are not many of them, well a 1600 ranked player is going to crush 1200 ranked players as well. A 1200 ranked player should be able to beat 800 ranked players. If you hard reset you throwing all of these into a group with no plan to separate the skill levels to attempt to maintain some sort of balance.

A soft reset lets you stop these complete face roll games, and still allow for changes in the ladder by opening up different players to higher levels of play.


Comment below rating threshold, click here to show it.

MathMage

Senior Member

07-12-2011

Player 1 is 1700 Elo
Player 2 is 700 Elo

Both players have their Elo reset to 1200, but Player 1's MMR is 1700 while Player 2's MMR is 700.

Player 2 wins a game against some other MMR 700 player; his Elo goes up. Player 1 loses a game against some other MMR 1700 player; his Elo goes down.

Player 1 now has a lower Elo than player 2, even though he is demonstrably a better player.

This system does not work.


Comment below rating threshold, click here to show it.

Laraki

This user has referred a friend to League of Legends, click for more information

Senior Member

07-12-2011

Quote:
Originally Posted by replayssoon View Post
How does it make a mockery of competition?
In my opinion it makes a mockery of competition by limiting competition, the soft reset would limit the number of different opponents you can face more so than the hard reset, hard reset is more geared toward opening up the matchmaking system for a wider range of possible pairings and that makes things more competitive, and a soft reset would maintain "last seasons" rank place-holdings. From my perspective its hard to see that and think of good competition.

Ah ok MathMage i see your point. Thing of it is Elo is a representation just of Wins/Losses in a set period of time, not a good demonstration of skill, so in our system you would expect the player who won his game to go higher and thats -> perfectly acceptable, because he's playing against much less skilled players, things would even out in this situation when -> because our previously 700 elo player is now being matched up against harder opponents -> he's loosing more games and working back to his deserved lower ranking (though hopefully higher than he was last season), and because our previously 1700 player was made to face less challenging opponents than he previously was, he's back to working his way up as he should be.

Its ok for a person who won his ONE game to be higher Elo at the start of competition than a person who LOST his ONE game. Differences in SKILL are maid clear by the players experience (represented here usually by total games played), the differences in WHO they are competing against AND as more games are played, by their respective Elo's as well.

And for those examples im just "assuming" that the player of 700 elo rating previously had WAAAAAAAAY less game experience (tons less overall games played) than the player of 1700 elo before the reset...

If they are even on games played in my opinion (which to me represents better than any other statistic, overall gameplay experience). Then the fact that the guy who used to have a 700 elo rating is WINNING his games compared to the guy who was at 1700 being matched up against similar players, would show how effective the elo reset was at evening out the playing field and giving credibility to Elo ratings as opposed to how they are now... (All hypothetical, just saying)


Comment below rating threshold, click here to show it.

Laraki

This user has referred a friend to League of Legends, click for more information

Senior Member

07-12-2011

Ya know, in a soft ELO Reset, since everyone isnt reset to the same "rank". you end up with people at different numbers than they are currently sure, but one guy will still be "ranked" above another guy, at the START of competition. Thats nonsense to me -> Because Elo ratings arent the only means we have to matchmake, nor are they the most suitable in my opinion.


Comment below rating threshold, click here to show it.

pelota

Member

07-12-2011

Guys Id just like to go back to the original topic. The main idea behind elo if you actually read up on it is to match players of roughly equally skill so they both have a 50% chance to win. It isnt some record of your wins or how well you did this season. So this MMR that you are proposing is actually Elo. I do however like the idea of keeping Elo as it is, but having it be hidden to determine who you play, and then have a seperate measure for how you did that season.

What do you guys think?


Comment below rating threshold, click here to show it.

HateYerGuts

Senior Member

07-12-2011

It's in HoN therefor you will be downvoted.


Comment below rating threshold, click here to show it.

Laraki

This user has referred a friend to League of Legends, click for more information

Senior Member

07-12-2011

Quote:
Originally Posted by pelota View Post
Guys Id just like to go back to the original topic. The main idea behind elo if you actually read up on it is to match players of roughly equally skill so they both have a 50% chance to win. It isnt some record of your wins or how well you did this season. So this MMR that you are proposing is actually Elo. I do however like the idea of keeping Elo as it is, but having it be hidden to determine who you play, and then have a seperate measure for how you did that season.

What do you guys think?

To my knowledge, the way Elo works in league of legens right now, it actually IS a measured record of your wins and how "well" you did this season. The idea of the machmaking system should certainly be to match you with the best opponents to give you a 50% chance to win at all times, but that's not how things work right now and the reason for that is -> Because Elo is a rating of how well you did this season in a team situation and not a fair representation of your skill.


Comment below rating threshold, click here to show it.

replayssoon

Member

07-12-2011

Quote:
Originally Posted by Laraki View Post
In my opinion it makes a mockery of competition by limiting competition, the soft reset would limit the number of different opponents you can face more so than the hard reset, hard reset is more geared toward opening up the matchmaking system for a wider range of possible pairings and that makes things more competitive, and a soft reset would maintain "last seasons" rank place-holdings. From my perspective its hard to see that and think of good competition.

Ah ok MathMage i see your point. Thing of it is Elo is a representation just of Wins/Losses in a set period of time, not a good demonstration of skill, so in our system you would expect the player who won his game to go higher and thats -> perfectly acceptable, because he's playing against much less skilled players, things would even out in this situation when -> because our previously 700 elo player is now being matched up against harder opponents -> he's loosing more games and working back to his deserved lower ranking (though hopefully higher than he was last season), and because our previously 1700 player was made to face less challenging opponents than he previously was, he's back to working his way up as he should be.

Its ok for a person who won his ONE game to be higher Elo at the start of competition than a person who LOST his ONE game. Differences in SKILL are maid clear by the players experience (represented here usually by total games played), the differences in WHO they are competing against AND as more games are played, by their respective Elo's as well.

And for those examples im just "assuming" that the player of 700 elo rating previously had WAAAAAAAAY less game experience (tons less overall games played) than the player of 1700 elo before the reset...

If they are even on games played in my opinion (which to me represents better than any other statistic, overall gameplay experience). Then the fact that the guy who used to have a 700 elo rating is WINNING his games compared to the guy who was at 1700 being matched up against similar players, would show how effective the elo reset was at evening out the playing field and giving credibility to Elo ratings as opposed to how they are now... (All hypothetical, just saying)
NO ONE DOES A FULL LADDER RESET, no company willingly sets up players to get stomped. Which is what you are advocating. This is akin to taking every single baseball player in America from little league+rec leagues+high school+minor leagues+major leagues and just throwing them in the same league to play. How does that make it more competitive?

Quote:
Originally Posted by Laraki View Post
To my knowledge, the way Elo works in league of legens right now, it actually IS a measured record of your wins and how "well" you did this season. The idea of the machmaking system should certainly be to match you with the best opponents to give you a 50% chance to win at all times, but that's not how things work right now and the reason for that is -> Because Elo is a rating of how well you did this season in a team situation and not a fair representation of your skill.
So you are saying that my rating is not a fair representation of my skill? OK so that means that Jiji really is not good at this game, right? Cause his Elo is not a fair representation of his skill. I mean that is what you are saying right?


Comment below rating threshold, click here to show it.

MathMage

Senior Member

07-12-2011

Quote:
Originally Posted by Laraki View Post
In my opinion it makes a mockery of competition by limiting competition, the soft reset would limit the number of different opponents you can face more so than the hard reset, hard reset is more geared toward opening up the matchmaking system for a wider range of possible pairings and that makes things more competitive, and a soft reset would maintain "last seasons" rank place-holdings. From my perspective its hard to see that and think of good competition.

Ah ok MathMage i see your point. Thing of it is Elo is a representation just of Wins/Losses in a set period of time, not a good demonstration of skill, so in our system you would expect the player who won his game to go higher and thats -> perfectly acceptable, because he's playing against much less skilled players, things would even out in this situation when -> because our previously 700 elo player is now being matched up against harder opponents -> he's loosing more games and working back to his deserved lower ranking (though hopefully higher than he was last season), and because our previously 1700 player was made to face less challenging opponents than he previously was, he's back to working his way up as he should be.

Its ok for a person who won his ONE game to be higher Elo at the start of competition than a person who LOST his ONE game. Differences in SKILL are maid clear by the players experience (represented here usually by total games played), the differences in WHO they are competing against AND as more games are played, by their respective Elo's as well.

And for those examples im just "assuming" that the player of 700 elo rating previously had WAAAAAAAAY less game experience (tons less overall games played) than the player of 1700 elo before the reset...

If they are even on games played in my opinion (which to me represents better than any other statistic, overall gameplay experience). Then the fact that the guy who used to have a 700 elo rating is WINNING his games compared to the guy who was at 1700 being matched up against similar players, would show how effective the elo reset was at evening out the playing field and giving credibility to Elo ratings as opposed to how they are now... (All hypothetical, just saying)
The problem here is that you haven't actually reset anything meaningful. Elo, in this example, would not measure your skill, and it would not measure your performance in Season 2 (because it doesn't take Strength of Schedule into account); it would measure whether you improved or regressed from Season 1. That's a pretty random statistic to hang your Elo on.