There is only one big difference. All of those games - oposite to LOL, are based still on evaluation only one player (in meaning as a team) and not whole group of players. All of those games are making a matches against single players in 1vs1 formule. Even if there will be a basketball team, a player will be controling whole team. So he can be rated with just only traditional function of ELO system even withou changes.
Not quite even if we only use the examples Wikipedia gives us. " but is also used as a rating system for multiplayer competition in a number of video games" A lot of games are individual players with RNG teammates. Scrabbles up to 4 players, Diplomacy is "Best played with 7"
So we know the systems used for more then 1 player or 1 team and can be adapted to account for multiplayer.
So we know its used for more then just league and that any multiplayer game with inconsistent teammates via matchmaker will run into the exact same difficultys Riot would.
So can Elo systems work in that environment? Yes its something that can be tested and works.
Here you have a guy who built his own team MMR system that accounts for new teammates per game In the video he shows how just porting the algorithm for 1v1 or Team v Team Will fail to accomplish the principles of Elo systems. It works off wins and losses, Bad players go down good players go up and the system fixes itself over time. If you treat each team as 1 collective whole and adjust them as 1 as if it was a single player it throws off each individuals MMR gains/losses. For a multiplayer game, even if its team biased you need to account for the individual when they are not a constant on a team. Its done for team games, its also done for 1v1v1v1 games. To say it cant be used for a team game like league is provably untrue. We can make a system that accounts for individuals and still follows the Base principles Arpad Elo used even for a team game with rotating teammates.
Luck has more impact but ultimately there's a workable range of player skill distribution that we can show on a graph as a curve.
But difference is only a rating value of specific ELO system and it is based on same formula.
First up. and none of them follows Elo's original suggestions precisely!
O People have disagreements with Elo's Original formula. Its based on the same principles not the same algorithm. If you change the variables its no longer THE elo formula, if you keep the principles behind it the same how ever it is called an Elo formula. This is why even when heavily modified many people will say they are using a Elo system for their games. Even if its multiplayer or not chess. Modifying the variables or adding additional caculations are fine, it needs to obey the principle not copy and paste the algorythem. After all, not every game has a draw but chess does. You shape it for your game.
This information is worthless. First of all it becouse it is dated on 2012 year. For the secound it does not explain nothing.
Not worthless. It shows when riot officially stated they used a Elo system. Now its mot just me telling you riot used a Elo biased system and that you should just take my word. Here is the devs officaly calling it Elo and discussing that. If you read the article and compare it to modern League you find overlap in the systems. Let me show you.
"Will I see my Elo when the preseason starts?
No, you will be unranked, and you’ll need to finish provisional matches before you can see your Elo again. You’ll be playing against players of similar skill though because you’ll start your provisional matches with an Elo based on the soft reset.
If I’m an experienced player, will I still gain or lose only around 12 points after each game?
No, we’ll be resetting the system’s confidence in your rating. You’ll be gaining or losing similar amounts of Elo as when you first started playing ranked."
Note that when season 3 hit instead of gaining or losing similar amounts of Elo we started to gain LP and had a hidden MMR that we also gain and lose. Elo and MMR are what the Matchmaker used/uses to build game, according to the devs.
This link is relevant because I need to show the connection between MMR as a derivative system from Elo, we need to be able to look at it and ask does this follow the same principles Arpad used?
Back when it first came out. 2009 post by devs
Expectations they set for the system to players.
I want to show what to look for to determine if somethings comparable to Elo.
So 2009. http://forums.na.leagueoflegends.com/board/showthread.php?t=12029
"From there, the game is played. If you win, you gain points, if you lose, you lose points."
- https://support.riotgames.com/hc/en-us/articles/201752954-Matchmaking-Guide
"Everyone’s MMR starts at the same point when playing a queue for the first time.** It goes up when you win, and goes down when you lose. "
**
System ELO in traditional meaning can be applied to any game/team controled by only one player and it still does not matter if it will be football gme , American football game, basketball game, Major League Baseball game, table tennis, Scrabble becouse if whole team can be controled by one player - it still fit to system ELO description - and that is why the name of this system for those games - do not have be chang
It also is used for games like Leauge. Still fits the Description. Many games use it for their Matchmaker. The system is so flexible it can be applied in a lot of ways. Take a look at it being applied outside of games. https://www.youtube.com/watch?v=bLsM1z8lX_A
Can Elo work for team games like league with the right modifications? Yes. Again PLEASE! check this guy out he shows and tests it. https://www.youtube.com/watch?v=kB2UADNoRUA <- If you don't think it works for team games on an individual level please check this out. Shows the math. Shows the Graphs, Builds and tests systems.
So keep in mind Elo systems are not going to always resemble THE elo system but Learning about the OG elo system and why it works gives us the ability to judge if some one is using a Derivative system.
By looking on your question and by trying to relate them to things I wrote above - it is possible to write that this question make no sense because it is you - who do not understand logic and sense of applicability ELO system to specific environment.
No, See I know the principle of Elo system. I looked at his algorithm, I have looked at others, I checked a bunch of articles on it. I found some one take and build a Elo team system, I know multiple developers for games like league use Elo or Elo derivatives that adhere to the same core principles then adapt it to their game.
That Elo can be adapted to work its intended purpose for games like Leauge is a fact. Tested and proven. Sources for me saying that is provided in this post. That its used for games like Leauge is a fact. Source provided in 2 different posts from riot. 2009 and 2013 That there are many games working of derivative systems is a fact. (True skill, Glicko)
Thing is people seem to think the systems Fair games = actually fair games. It does not, that's not how it works. It fixes individual mismatches over time. It does not start correct. ****
Uncertainty and adjustment are Core to Elo. It requires games to be played. A lot of games, before your reasonably close to your skill. Your luck with teammates will have a pretty huge impact on this. It can easily double the # of games required compared to some one with good luck on teammates. Luck is a factor, but with the right skill curve Luck is only a factor in how fast you get to your skill, not how high you can go. At lower and higher Tiers Luck needs to be lowered. This is why K factor is so important.
Is it fair?
Yes/no. Depends on your opinion. If your bad your going to lose a lot more then you win because you have to get carried and carried hard. But that's a problem in every match you play, if you got better then you'd lose less often.
If your average? Bad news, your going to be counting on teammates or enemy's to be mismatched for wins. Your personal impacts going to be only as good as everybody properly placed at your level.
If your good? Great for you. going to pull more weight then the system expects so your opponents are easier for you to beat in every match.
Its fair in giving everybody the same potential to climb with skill. Its not fair because not all players are equals. Not every player has the time to put in high # of games. Not everybody improves at the same pace. And the game League of legends is inherently imbalanced.
Its competitive, it does divide ranked ladder by skill rather then just time spent. Fair is subjective to what we are judging it by.
Skill games are not fair unless the skill between players is close to equal. Mismatches are inherently unfair. What is Fair about Elo systems is how they treat all players equally. Its on your ability to win consistently. If you cant, that's on you. The system does not show favoritism. What is fair is that it does build more balanced matches off skill then other systems. Whats fair is that It will properly place players by skill if given enough time. In practice, your not going to have a fair match every game. The system accounts for and expects this. That's part of what makes it Elo or one of its derivatives.
By looking on your question and by trying to relate them to things I wrote above - it is possible to write that this question make no sense because it is you - who do not understand logic and sense of applicability ELO system to specific environment.
Posting that quote a second time because I want to show how its possible that we know ELO system works for team games like league. Did you watch This Link yet?
Contact me if you need any help translating it.
