Why do people claim ELO Hell doesn't exsist?

Comment below rating threshold, click here to show it.

LogicalTautology

Senior Member

12-09-2012

Quote:
Originally Posted by Buffalax View Post
I think the problem that a lot of people have, PogoPogoPogoPogo included, is that they think of player skill as a sort of number that gets summed over a team in some way to determine who wins. The team with the highest total skill should win, they reason, so if your skill is higher than average for your elo, your elo should slowly rise.

The problem with this is that, regardless of elo, there are few teammates so bad they usually lose; if they were that bad, their elo would drop further, and you wouldn't get matched with them. What I've found instead is that a lot of low-elo players are relatively bad at playing but surprisingly good at carrying. That is, they've honed a particular playstyle enough to snowball into a raging death machine if things go their way, but if they get countered or make an early mistake, they stick to their rigid, aggressive tactics and quickly become worse than useless. I think Asellas's comment is rather telling:

At higher elos, this isn't normal; one may carry some games very well, but unless one's assists are through the roof, going kill-less is usually a sign that someone made a serious mistake. In elo hell, though, this sort of "bipolar" performance seems fairly common.

So what's the upshot? When so many players are walking a fine line between awesome and abysmal, the winner stops being a question of "which team has the most total skill?" and starts being a question of "which team has the bigger snowballs?". In this setting, it doesn't matter if you get a kill or two, keep up good cs, and take down your tower, because the winning team was already decided when one of the adcs went 6/0 by 15 minutes. If anything, playing a cautious and flexible game can slightly decrease your chances of winning, because your team has one fewer potential snowball. This also explains why gold-and-above-level players tend to deny that elo hell exists; if you're so good you can not only adapt to but reliably dominate changing circumstances, you can snowball against an inflexible player nearly every time, and elo hell becomes easy to escape.

Sorry for the wall of text - I've been pondering this for a while. Since putting my finger on what I think the problem is, I've made an effort to 1)always choose a champion well suited to 1v1 and 1v2, 2)never play support if I can help it, and 3)focus on helping my teammates win their lanes as much as I focus on winning my own, and I seem to be slowly but steadily regaining my lost elo.
Let us assume that every player is 'matched up' against the opposing player in the same role. So there is my matchup, plus 4 other matchups.
Let us assume that in X% of games, the non-me matchups go 0-4. I will lose 95% of these games.
In the same X% of matches, by definition, the non-me matchups go 4-0. I win 95% of those games.
Let us assume that in Y% of matches, the non-me matchups go 1-3. I will lose 75% of those games.
In the same Y% of matches, the non-me matchups go 3-1. I win 75% of those games.
In the final Z% of matches, the non-me matchups go 2-2.

I will now argue several different things that all combine to prove that Elo Hell is a myth.
1) If you are even 50-100 Elo below your true Elo, you will win an absurdy high amount of the Z% of your games. Since you win 50% of the X/Y games, your Elo will rise until you win 50% of your Z games. As soon as you win 50% of your Z games, you are winning your matchup 50% of the time and have found your correct Elo. If you are at least 100 Elo below your true Elo, you will actually win more 1-3 matchups than you lose 3-1 matchups, making the Elo gain that much faster.
2) The one slight concession I will make to the 'Elo Hell' crowd: The lower the Elo bracket you are playing in is, the lower the percentage of Z games you will play in. This means that someone who is 10% better than the field at 800 Elo will take slightly longer to reach their Elo than someone who is 10% better than the field at 2000 Elo. This does not mean that you are being dragged down by bad teammates, though, as just as often as your teammates go 0-4 and cost you the game, they will go 4-0 and carry you.


Comment below rating threshold, click here to show it.

LogicalTautology

Senior Member

12-09-2012

Quote:
Originally Posted by Enconex View Post
The problem is, no one seems to realize just how bad it is at this elo. No matter how many games you play the system WILL NOT average you out to where you need to be. Nearly EVERY game there is someone who WANTS to lose the game, or someone who feeds so much it is not possible to carry the game. I don't care if your some pro from CLG or TSM, you cant carry a team who has fed the enemy so many kills.
Both math and anecdotal evidence say you are flat-out wrong. The math is undeniable and has been posted in this thread before, so I won't repeat it.

I have numerous examples of players (both pro and 1500-level) doing exactly what you say is impossible, though. The most prominent example off the top of my head is Chaox's Elo Marathon. He took a smurf and shot through the rankings. I don't think he lost a single game before he hit 1600, in fact.


Comment below rating threshold, click here to show it.

PogoPogoPogoPogo

Senior Member

12-09-2012

Quote:
Originally Posted by EyeHearShapes View Post
No where did i claim i belong in a higher elo .
First, if you're not saying this and not claiming this, then you're not in Elo hell. You're at the Elo you belong.

Quote:
Originally Posted by EyeHearShapes View Post
You cant put a variable on a random action like a mistake. Since well its random . Saying theres a 50/50 split doesnt work with a random unpredictable action.
Just because something happens at a random chance doesn't mean we can't discern the odds of it happening. And even if we don't know the exact odds of it happening to one team or the other doesn't mean we can't logically deduce equal odds for both teams.


Do you know what the odds of drawing an ace of the top of a randomly shuffled deck?
NOPE. THE ACT OF SHUFFLING A DECK IS RANDOM AND UNPREDICTABLE THEREFORE WE CAN NOT ASSIGN A PERCENTAGE CHANCE TO THAT RANDOM, UNPREDICTABLE ACT HAPPENING.

Well, wrong. We actually can. There are 52 cards in a deck. Of these, 4 are aces. The odds of drawing an ace off the top of a completely randomly shuffled deck are 4 in 52, or 1 in 13, or roughly 7.7%.


Comment below rating threshold, click here to show it.

Enconex

Senior Member

12-09-2012

Quote:
Both math and anecdotal evidence say you are flat-out wrong. The math is undeniable and has been posted in this thread before, so I won't repeat it.

I have numerous examples of players (both pro and 1500-level) doing exactly what you say is impossible, though. The most prominent example off the top of my head is Chaox's Elo Marathon. He took a smurf and shot through the rankings. I don't think he lost a single game before he hit 1600, in fact.
Fair enough, if your someone like chaox, then with enough games played you will advance just because you are that good. But no 1500 elo player could do that unless they were just as good as Chaox/any other professional.

I'm telling you though, I AM better than many players in the 1500-1600 range. My stats prove it, and me playing with them prove it, their decision making just boggles me when I see them run in 1v4 trying to save a tower or initiate a fight when one of our players is at our base.


Comment below rating threshold, click here to show it.

EyeHearShapes

Senior Member

12-09-2012

Chaox's Elo Marathon. I'm not sure just how relevant this is . Chaox was playing his smurf on an EU server maybe Elo Hell is different there :P Just speculation. Also i did say that if you play well and give it a good try every game you'll climb out . I'm fairly certain Chaox is way more skilled then the ppl he was playing with at the 1000 elo mark.


Comment below rating threshold, click here to show it.

PogoPogoPogoPogo

Senior Member

12-09-2012

Quote:
Originally Posted by Buffalax View Post
I think the problem that a lot of people have, PogoPogoPogoPogo included, is that they think of player skill as a sort of number that gets summed over a team in some way to determine who wins. The team with the highest total skill should win, they reason, so if your skill is higher than average for your elo, your elo should slowly rise.
I don't think of skill in this way.

But if you don't think that the team with the highest overall skill or highest average skill wins has a higher chance of winning that directly correlates to their skill as a team, then why the hell are you playing ranked?

If the winner of a match is not in any way impacted by the ability of the individuals on each team to win a match, then the winner of any game (EVEN premade 5s matches) is a matter of an entirely random flip of a coin. Skill means literally nothing.

And if skill means literally nothing in LoL, then not only does Elo become a 100% meaningless measure of nothing (this means a 2500+ Elo player has a 50/50 chance of winning against a sub-500 Elo player, because the winners of matches is a flip of a coin), but it also actually means that you CAN NOT claim to be better than anyone at this game.

Why?

Because it's no longer a game of skill. It's a game of getting lucky coin flips. I can't possibly make the claim that I'm better at calling heads/tails on a coin flip then you because it's pure luck.

So in that regard, Elo hell still couldn't possibly exist because you can't be better than your teammates. No one can be better than anyone because skill means nothing and games are all truly decided by the flip of a coin.



Yet the professional players make smurf accounts and carry from 1200 to 1800 without losing a single game.


Comment below rating threshold, click here to show it.

PogoPogoPogoPogo

Senior Member

12-09-2012

Quote:
Originally Posted by EyeHearShapes View Post
Chaox's Elo Marathon. I'm not sure just how relevant this is . Chaox was playing his smurf on an EU server maybe Elo Hell is different there :P Just speculation. Also i did say that if you play well and give it a good try every game you'll climb out . I'm fairly certain Chaox is way more skilled then the ppl he was playing with at the 1000 elo mark.
EXACTLY

Chaox is better than 1000 Elo players. He climbed out. Regardless of how poorly his teammates played, he climbed out. Why? Because while his teammates were 1000 Elo players, his enemies were also 1000 Elo players.

Why would a team with four 1000 Elo players and Chaox not beat a team with five 1000 Elo players every single time?


Comment below rating threshold, click here to show it.

EyeHearShapes

Senior Member

12-09-2012

Quote:
Originally Posted by PogoPogoPogoPogo View Post
First, if you're not saying this and not claiming this, then you're not in Elo hell. You're at the Elo you belong.


Just because something happens at a random chance doesn't mean we can't discern the odds of it happening. And even if we don't know the exact odds of it happening to one team or the other doesn't mean we can't logically deduce equal odds for both teams.


Do you know what the odds of drawing an ace of the top of a randomly shuffled deck?
NOPE. THE ACT OF SHUFFLING A DECK IS RANDOM AND UNPREDICTABLE THEREFORE WE CAN NOT ASSIGN A PERCENTAGE CHANCE TO THAT RANDOM, UNPREDICTABLE ACT HAPPENING.

Well, wrong. We actually can. There are 52 cards in a deck. Of these, 4 are aces. The odds of drawing an ace off the top of a completely randomly shuffled deck are 4 in 52, or 1 in 13, or roughly 7.7%.
Do u know the odds of me posting a new post on forum. Nope because i'm a random person with random action and my mind makes my decisions based on my emotions and thoughts . Pls pogo put a number on this .


Comment below rating threshold, click here to show it.

PogoPogoPogoPogo

Senior Member

12-09-2012

Quote:
Originally Posted by EyeHearShapes View Post
Do u know the odds of me posting a new post on forum. Nope because i'm a random person with random action and my mind makes my decisions based on my emotions and thoughts . Pls pogo put a number on this .
Just because I can't put a number on it doesn't mean a number doesn't exist. I don't have absolutely all the information.

But Riot does. Riot could put a number on the odds of you making a post on their forum. They've got all the data. They see how often you log on to the website, how often you read threads, how often you create new threads, how often you make new posts.

Riot has literally all the information, and they can tell you the odds of you making a new post on any given day. They can tell you whether or not you're more likely to post in the early afternoon or in the wee hours of the morning. They can tell you whether or not you're more likely to post on the weekday or the weekend.

They have ALL the information needed to calculate odds of you making posts. I don't.

But I have all the information needed to calculate the odds of drawing an ace out of a deck of cards, no matter how randomly it's shuffled. Even if you modify the deck. Even if you replace all the red cards with black cards from another deck. If you tell me the total number of cards in the deck and the total number of aces, I can tell you the odds of drawing an ace.

The only reason I can't put a number on your odds of making a post is because I don't have all the information.


Comment below rating threshold, click here to show it.

EyeHearShapes

Senior Member

12-09-2012

Quote:
Originally Posted by PogoPogoPogoPogo View Post
Just because I can't put a number on it doesn't mean a number doesn't exist. I don't have absolutely all the information.

But Riot does. Riot could put a number on the odds of you making a post on their forum. They've got all the data. They see how often you log on to the website, how often you read threads, how often you create new threads, how often you make new posts.

Riot has literally all the information, and they can tell you the odds of you making a new post on any given day. They can tell you whether or not you're more likely to post in the early afternoon or in the wee hours of the morning. They can tell you whether or not you're more likely to post on the weekday or the weekend.

They have ALL the information needed to calculate odds of you making posts. I don't.

But I have all the information needed to calculate the odds of drawing an ace out of a deck of cards, no matter how randomly it's shuffled. Even if you modify the deck. Even if you replace all the red cards with black cards from another deck. If you tell me the total number of cards in the deck and the total number of aces, I can tell you the odds of drawing an ace.

The only reason I can't put a number on your odds of making a post is because I don't have all the information.
What if a random action happens in my mind and decide never to post on forum again .
In any given game on SR there are ten different people playing . Each player actions are dictated by their mind . The minds of ten people or what they're thinking and basing there decisions on is a near impossible number to scale . Given that s1 can be thinking anything at any time.