ELO Clarification Please

12
Comment below rating threshold, click here to show it.

Zakonel

Senior Member

10-04-2010

So from what I understand, ELO is suppose to be your 'skill level' in the game. Thus, with this skill level, you are matched with players of the same skill or close to it in the matchmaking system.

Am I right so far?

So lets say Experienced Player Bob has been playing LoL for a very long time, even since the beta. He starts Season 1 at 1200 ELO like everyone else.

First game, he does amazing! But granted, he has a leaver on his team so his team loses. Bob loses ELO.

Second game, Bob does amazing! But of course, the two tanks on his team did not have a clue what they were doing. Bob loses ELO.

Example: Chogath (the main tank) stopping mid teamfight to type in chat "I am out of mana", while his team died, him shortly after. (Yes, that actually happened)

Third game, Bob does amazing! But sadly, it was free Karthus week. Karthus fed (insert carry here) and the team was stomped. Bob loses ELO.

So, if I am understanding this correctly... How does ELO have anything to do with our skill level? No matter how well 1 person does in a ranked game, ELO is solely based off of wins and losses.

I understand the law of large numbers quite well. I also understand there are 1001 different reasons why a person can lose a game, no matter how well they play, or their skill level.

I am also not saying the Matchmaking system is broken in any way, shape, or form. I believe it is working very well.

Can someone help me understand the logic behind this?
And why ELO is not based off of the actual performance and skill of the Summoner themselves?

Surely the brilliant minds at Riot can come up with some way to determine individual performance on a per game basis?


Comment below rating threshold, click here to show it.

Swipe Knife

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

Recruiter

10-04-2010

Clarification: Elo is a name, not an abbreviation.


Comment below rating threshold, click here to show it.

Foxflare

Senior Member

10-04-2010

To be honest, I think it's more like:

Bob is super experienced and pro and played since beta! He tries out Season 1, starting at 1200 Elo!

First game, he does amazing! There's a leaver on his team. =( But he carries the leaver anyway! Bob gains Elo.

Second game, Bob does amazing! The two tanks on his team did not have a clue what they were doing. Bob adapts to an early game strategy and wins anyway! He gains Elo.

Third game, Bob does amazing! But sadly, it was free Karthus week. Karthus fed -enemy carry- and he had a much harder time. But he manages to drop the carry with good positioning, pinging, and being overall better! Bob gains Elo.

Because it matters what Elo Bob truly belongs at. If he's a 1300-1400 player, you're probably right. If he's a 1600+ player, a 1700+ player, what I described is more realistic.


Comment below rating threshold, click here to show it.

Zakonel

Senior Member

10-04-2010

Quote:
Originally Posted by Leralyth View Post
To be honest, I think it's more like:

Bob is super experienced and pro and played since beta! He tries out Season 1, starting at 1200 Elo!

First game, he does amazing! There's a leaver on his team. =( But he carries the leaver anyway! Bob gains Elo.

Second game, Bob does amazing! The two tanks on his team did not have a clue what they were doing. Bob adapts to an early game strategy and wins anyway! He gains Elo.

Third game, Bob does amazing! But sadly, it was free Karthus week. Karthus fed -enemy carry- and he had a much harder time. But he manages to drop the carry with good positioning, pinging, and being overall better! Bob gains Elo.

Because it matters what Elo Bob truly belongs at. If he's a 1300-1400 player, you're probably right. If he's a 1600+ player, a 1700+ player, what I described is more realistic.
This is very true and also very possible. But lets say Bob should be in the 1300-1500 Elo ranking.

This means, he is an above average player, but not planning on winning any WCG tournaments.

Bob knows he doesn't belong in the 1100-1200 ranking, as he often offers very helpful advice to his team. Item builds, adapting strategies, warding encouragement, dragon, buffs...etc.

All of which most of his teammates seem to be totally oblivious to.


Comment below rating threshold, click here to show it.

Politburo

Senior Member

10-04-2010

Although you can be dragged down by bad players/players having a bad game, Theoretically, it will eventually even out to where you should be...it just might take a really long time...


Comment below rating threshold, click here to show it.

LolGuno

Senior Member

10-04-2010

(This was clarified a million times already)

"why ELO is not based off of the actual performance and skill of the Summoner themselves?"
- Do you have a system you can perhaps show us that would be more reasonable that the current system? I'm assuming you would like it to be based off KDR + takedowns. LOL is a team game!; you can play Janna every game, and have like a 1/4/20 average score.

EXPECT there to be leavers/noobs in Solo queue.. but hey, if you ARE good then your team should have the odds of having a noob/leaver from 4 players; the other team having 5 players

TL;DR
If you ARE ACTUALLY good.. you will eventually (might take awhile, or a short time) get to the ELO you should be in.
If you recognize the term "ELO HELL" and cannot get out of it, you have to face the fact that you ARE BAD.


Comment below rating threshold, click here to show it.

Zakonel

Senior Member

10-04-2010

Quote:
Originally Posted by Fufufufufu View Post
Although you can be dragged down by bad players/players having a bad game, Theoretically, it will eventually even out to where you should be...it just might take a really long time...
Again, the Law of Large numbers comes into effect. And this may very well work.

But my main question is:

Isn't there some way to rate a summoner on individual performance on a per game basis?

I am not saying it would be an easy task by any means...it would be quite hard...

But I know the capable minds at Riot can do it.


Comment below rating threshold, click here to show it.

LolGuno

Senior Member

10-04-2010

Quote:
Originally Posted by Zakonel View Post
But my main question is:

Isn't there some way to rate a summoner on individual performance on a per game basis?
Shouldn't your ELO prove your individual performance? Like he just said, if you ARE good, you will eventually carry up to the ELO you should be in. You may have a TERRIBLE game once and be at 1600+ ELO, but that probably doesn't mean you should fit into a lower ELO


Comment below rating threshold, click here to show it.

Swipe Knife

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

Recruiter

10-04-2010

Quote:
Originally Posted by Zakonel View Post
Again, the Law of Large numbers comes into effect. And this may very well work.

But my main question is:

Isn't there some way to rate a summoner on individual performance on a per game basis?

I am not saying it would be an easy task by any means...it would be quite hard...

But I know the capable minds at Riot can do it.
Short answer: No. There are an absurd number of variables to weigh in a game of this sort. Creep kills, APM, tower kills, damage dealt, damage taken, and k/d/a are all different aspects of in-game productivity, but does not aptly show many other aspects, such as proper ward placement, skills used in an effective manner, etc. Also, basing Elo on the above variables would turn the game into an Elo farm-fest with players ignoring their team to kill more creeps to boost their Elo, etc.


Comment below rating threshold, click here to show it.

Zakonel

Senior Member

10-04-2010

Quote:
Originally Posted by Swipe Knife View Post
Short answer: No. There are an absurd number of variables to weigh in a game of this sort. Creep kills, APM, tower kills, damage dealt, damage taken, and k/d/a are all different aspects of in-game productivity, but does not aptly show many other aspects, such as proper ward placement, skills used in an effective manner, etc. Also, basing Elo on the above variables would turn the game into an Elo farm-fest with players ignoring their team to kill more creeps to boost their Elo, etc.
You used ward placement as an example. Hell, if someone even BUYS a ward, their Elo should jump 100 points automatically.


12