"ELO Hell" math proves it doesn't exist

AlienPrimate·Sep 26, 2015·14 votes·2,476 views

I am getting really sick of hearing people complain about bad teams and trolls at lower ranks. I am not saying that feeders and trolls don't suck because I hate them too, but they shouldn't be an excuse for you being stuck in bronze.

Bronze feeder/troll math for those of you who can't do math:

Which team will get the feeder/troll

You's on your team: 1 Team mates: 4 Enemies: 5 Randoms who can be trolls: 9

Chance of being on your team assuming you are not the troll: 44% Chance of being on enemy team: 56%

Duo's on your team: 2 Team mates: 3 Enemies: 5 Randoms who can be trolls: 8

Chance of feeder on your team: 38% Chance of feeder on enemy team: 63%

If you play solo ranked in this so called "elo hell" there is a 12% greater chance that the trolls will be on the enemy team benefiting you. If you play duo, that increases it to 25% chance.

Stop complaining about "elo hell." Statistically speaking, feeders actually help you a lot more than they hurt you. Over a short period of time trolls, feeders, and unskilled team mates can hurt you, but if you play a while the odds will eventually work out to be close to the math I provided. My conclusion is that it is your own fault if you are stuck in bronze for a long period of games because you are the only factor in every game you play.

43 Comments

SorryBoysImLesSep 26, 201511 votes

That is what we call a piss-poor system.

"If you're losing a lot due to trolls and AFKs, don't worry, eventually you're gonna get a bunch of easy face-roll games because the enemy team is gonna get them instead!"

Do people actually like winning in those scenarios? I come from games where, when scenarios such as that occurred, it was etiquette to have a member of the unaffected team to sit out and create and even playing field.

Can't do that on LoL, though, you'll just get banned for wanting a fun, fair match.

AlienPrimateSep 26, 20154 votes

I totally agree with you. This post was just to show the whiners out there that they either haven't played enough games or that they are actually the problem and not their team mates

Tails DollSep 26, 20153 votes

Honestly I think ELO hell does exist. In fact your math proves that. You need a 65% win rate to win series and promos if I remember correctly. Accordint to you, not even with duos I can reach that number if you just take into account trolls etc. You aslo have to account for straight out getting out played our out picked. For every 10 games I play, I only feel 1 is a fair game and I only "suck" in 1/2 and carry on 1/2. The rest are a tost of the coin to who ever gets the troll.

This makes climbing out of each division hell.

Hell, I've been dancing in and out of plat 2 and 3 becouse of what I just told you.

DeynaTaggerungSep 26, 20151 votes

his numbers ignore any ability you have to carry the game you're in. Bronze may have a lot of bad players but if you're better than them you get out pretty quickly. For example I went from B5 -S4 and then capped out for a while because the players around me were as good/better than me. I wasn't crazy better than the people in bronze has I climbed but it was possible nonetheless.

Tails DollSep 27, 20151 votes

The problem is that ELO hell does exist and that's a good thing. ELO hell is just "how skilled you are", and when you reach that point then you reach your ELO hell.

The problem is that with the current system you dont know when you reach your ELO hell. I've been following my LOLKing stats, and every time I get them up by 100, I get stuck in a losing streak where I lose 100.

This means that I never reach a stable ELO, and is frustrating because you know some of that ELO you gained was thanks to a troll. And some you lost too.

As far I can tell, I belong to plat 2. I cant climb any further than that, but I seem to reach that division fairly easily. How ever, I'm on the brigde of getting demoted to plat 4, thanks to unfortunate games where I get DC/Troll/flamers and the enemy team does not.

According to OP some time or another I will get enemies with trolls/dc/flamers that should even out my ELO, plus if I truly belong to plat 2 I should carry some games too.

The thing is, it does not feel satisfiying or rewarding. It feels like a toss of the coin and is frankly pretty dumb.

However, I did hear that they are changing the Ranked system for the next season, hopefully they will remove this stupid division system we got now.

DeynaTaggerungSep 27, 20151 votes

Wait, how is playing with people of approximately your skill level elo hell? O.O

It seems that'd be the best place to be in this game because it's there that you can really be challenged. The whole point of the system is to put you in an environment where you can not only show how good you are, but also have a competitive environment to really enjoy close games. I think your perspective seems kind of skewed.

Most people would agree elo hell is being stuck on the low end of the ranked spectrum when you actually "deserve" to be higher because all your teammates are flamers, trolls, afks, etc. And we're saying that doesn't really exist, maybe to a small degree (1 tier or so) but not as bad as everyone makes it out to be.

I think if you approach this differently than you have been then League could be a lot more fun for you. LoL is a game, so if you don't win it's not the end of the world, see what you can learn from it and get ready for the next challenge.

Tails DollSep 28, 20151 votes

For me ELO hell is when you reac a point where you cant climb higher. No matter if you are bronze or diamond 1.

DeynaTaggerungSep 29, 20151 votes

But what I'm asking is how is that anything resembling hell? Shouldn't that be elo heaven where you actually have something like evenly matched games and get to improve as a player?

SmokedAlmondsSep 29, 20152 votes

{quoted}

But what I'm asking is how is that anything resembling hell? Shouldn't that be elo heaven where you actually have something like evenly matched games and get to improve as a player?

Its considered hell because until you reach that point you are winning a higher percentage of games. And from that point on you will never have much more than 50% winrate unless you DRASTICALLY improve. So you have someone who is used to winning 55-60% of their games now only winning 50%. Also when you reach that point you have less ability to make up for a weak teammate since you no longer have a significant skill advantage.

It also matters what your rank is when you hit that point. For example, if that point is anywhere near low silver you have tons of games decided by which team has a smurf, or which team has the new guy that is gonna drop to bronze 5 but just hasn't played enough games for that to happen yet.

DeynaTaggerungSep 29, 20151 votes

For example, if that point is anywhere near low silver

Heh, that'd be me(although I hope to improve as I have in the past).

I think rather than bemoan the fact that you're not losing 1 more game in every 20 it'd be better to just enjoy the game for what you can do. Because if League is just about winning and being able to stomp on your opponents then you're setting yourself up for disappointment and setting everyone else up for sad days.

SmokedAlmondsSep 29, 20151 votes

{quoted}

Heh, that'd be me(although I hope to improve as I have in the past).

I think rather than bemoan the fact that you're not losing 1 more game in every 20 it'd be better to just enjoy the game for what you can do. Because if League is just about winning and being able to stomp on your opponents then you're setting yourself up for disappointment and setting everyone else up for sad days.

Ya, I mean a lot of those are huge problems from an enjoyment standpoint. But they just don't have a significant impact long term impact on your rank.

AtaraxicOct 2, 20151 votes

Im way more sick of this horrifically flawed statistics example you use. just to make it simple, im sure we can agree that a coin flip is 50/50.

now go run your experiments and tell me statistically how many times 100 flips are 50/50.

Your 5 out of 9 base stats are correct but the real world results will not even be close.

man o man

AlienPrimateOct 4, 20151 votes

Statistically speaking, 8% of the attempts using 100 flips would be exactly 50/50. 68% are between 45 and 55. 95% are between 40 and 50.

Increase this to 1000 flips and 99.8% of attempts will be between 450 and 550.

My point of this is that statistics is not flawed in completely random occurrences. Every game you play it is completely random on who your 4 team mates are and who the 5 enemies are so please tell me how the results won't even be close. The more data you have the more accurate the statistics will become meaning the more games you play that have trolls in them the closer it will be to the statistics in my post.

AtaraxicOct 5, 20151 votes

False.. if you do this test in real world your nu bers will not be that consistent. the outlyers will be much more. and we are speaking of the outlyers when discussing elo hell.

KãtãrinãSep 26, 20153 votes

the math is finally revealed. Well unless you know you knew this all already LMAO.

Remlap1223Sep 26, 20152 votes

Or, and here's an idea, kinda outlandish, but stick with me here, maybe, oh, I dunno, BAN THE TROLLS SO NOBODY HAS TO DEAL WITH THAT SHIT?

ninjaIRL Sep 26, 20152 votes

Trolls, feeders, DCers, AFKers, flamers, griefers are indirectly the cause because some players can't seem to just shrug them off and focus on their own play, instead they get salty and tilt on team mate plays.

I get salty and annoyed as well but that shouldn't be a reason to forget the goal; to get better at the game.

There are some problems with how the promotional system works but it's not really that big of an issue. I went from Bronze II about 3 or 4 months ago and I'm ~~not~~ on in Silver IV slowly achieving higher ranks.

EDIT: Dyslexic typing.

Lotus BladeSep 26, 20152 votes

elo hell doesnt exist what exist is you lose one game you will keep losing till u drop 2 divisions then maybe u can get good teams and win

Magic MoSep 26, 20151 votes

Another thread misinformed about the true definition of "ELO Hell." People that cry about getting trolls in lower tiers aren't in ELO Hell. They're where they should be provided they're even in wins/losses (0-2% of games leeway either way).

Players that are incredibly negative in their win/loss ratio, meaning they were demoted quite a lot from what they were when they were first placed in the ranked system, are either:

Overplaced.
Underpracticed.
Stuck in ELO Hell.

Why the third option? Let me explain.

You win 7 games in a row. You get 12-15 LP per win, since your win/loss is terribad. That's 13.5 games on average, meaning you've gained roughly 95 LP. Not enough for a promotion.

You lose your momentum and drop back down to Earth. Only a little though. You lose 5 games. You lose 20-23 LP, since your win/loss is terribad. 21.5 average. That means you've lost about 107 LP.

tl;dr Your "reward" for going 7-5 is -12 LP when you're in ELO Hell. I know this because I've been in those fiery pits of darkness. ELO Hell is very real, and I'm tired of people that don't even play ranked telling me otherwise.

Maximum MordeSep 26, 20153 votes

{quoted}

Another thread misinformed about the true definition of "ELO Hell." People that cry about getting trolls in lower tiers aren't in ELO Hell. They're where they should be provided they're even in wins/losses (0-2% of games leeway either way).

Players that are incredibly negative in their win/loss ratio, meaning they were demoted quite a lot from what they were when they were first placed in the ranked system, are either:

Overplaced.

Underpracticed.

Stuck in ELO Hell.

Why the third option? Let me explain.

You win 7 games in a row. You get 12-15 LP per win, since your win/loss is terribad. That's 13.5 games on average, meaning you've gained roughly 95 LP. Not enough for a promotion.

You lose your momentum and drop back down to Earth. Only a little though. You lose 5 games. You lose 20-23 LP, since your win/loss is terribad. 21.5 average. That means you've lost about 107 LP.

tl;dr Your "reward" for going 7-5 is -12 LP when you're in ELO Hell. I know this because I've been in those fiery pits of darkness. ELO Hell is very real, and I'm tired of people that don't even play ranked telling me otherwise.

Keep up the 7-5 win pattern though and you will fix your mmr and start gaining more than you lose. So, I'd say #3 is really, they haven't played enough games combined with the the first two to get them into that position in the first place.

For CarthageSep 26, 20151 votes

Well, no. ELO Hell is the period of time in which you have to deal with your MMR sucking shit on buttered toast, which can lead to the trolls, feeders, and AFK's, but they are not innately the cause of it - they're usually a byproduct or the initial reason for the drop in MMR, not the reason you're stuck.

But it's a period of time in the end. However, you can get kicked back into ELO Hell by feeders and trolls and AFK's, and I know because I was in my promos for Plat 1 5 times already before having 5 consecutive troll-filled games that made my MMR plummet so quickly the Earth cracked and Cthulhu rose from the center of it once again to wreak havoc upon the human race. And that's dropped me back to Plat 5 twice now.

Like, going up one rank can be a fluke. Two can be as well. Three is stretching that possibility, and if you're headed to ranking up four times before getting Spartan kicked back somewhere, that's not you "being where you belong", that's the game saying, "Hey, you're doing well. Let's give you this unranked player three consecutive times who hasn't played a game since Season 3 and only one game in Season 4. Have fun, later."

... So the endpoint I'm getting at is: it is a period time, not a permanent situation. ELO Hell is real - but probably more correctly named "ELO Purgatory".

Maximum MordeSep 26, 20152 votes

[{quoted}](name=For Carthage,realm=NA,application-id=cIfEodbz,discussion-id=xkbrZ04h,comment-id=000300000000,timestamp=2015-09-26T05:31:12.677+0000)

... So the endpoint I'm getting at is: it is a period time, not a permanent situation. ELO Hell is real - but probably more correctly named "ELO Purgatory".

I can buy that. I disagree but it's based on reasonable thought process so sure.

AtaraxicOct 2, 20151 votes

Wrong.
It gets hard the more games you play not easier to break out of the gravity well. Also, any evaluation of individual skill in such a team game is doomed to be ... poor.

AlienPrimateOct 4, 20151 votes

How is measuring individual skill in a team game doomed to be poor? Let's talk football for a bit. If 1 person on an 11 man team screws up the whole team suffers or if 1 person has an outstanding game or even just outstanding single play the whole team benefits. How much more do you think that can be said of a 5 man team?

If you can consistently win your lane that means you are more likely to consistently win games. Also in bronze if you can be a good team leader in telling your team where to go for objectives and pay attention to the map and spawn timers to warn team mates of threats you will have much better games.

Probably one of the biggest reasons for doing poorly as a team is lack of awareness, especially in bronze. As you stated, it is a team game so even if you are far better than your laning opponent you can still get destroyed by them if you don't pay attention to the map. If just one person keeps their eye on the map it can save the team from multiple deaths and get free kills assuming you use wards.

AtaraxicOct 5, 20151 votes

In football the center can consistently win his "lane". he can be the best player on the field. he can be recognized by his peers as the best player on the field.

all true do u agree?

Now that player can also be on an 0-16 team, yes?

How do you propose to have any system reward/recognize that player in rank based on his win loss?

A system cant, thats my point.

AtaraxicOct 5, 20151 votes

In football the center can consistently win his "lane". he can be the best player on the field. he can be recognized by his peers as the best player on the field.

all true do u agree?

Now that player can also be on an 0-16 team, yes?

How do you propose to have any system reward/recognize that player in rank based on his win loss?

A system cant, thats my point.

footnote: are Dan Marino and Barry Sanders "bronze" in foot ball rank? they would be base don win loss record.

but we know them to be two of the best ever. Right? what a poor way to rank them on w/l.

AlienPrimateOct 5, 20151 votes

They are consistently great players which I agree with but here is where your point is flawed. They played with the same team in every game, not a new, completely randomized team of different players with different skill sets against a randomized set of opponents who know nothing of each other.

AtaraxicOct 8, 20151 votes

I think it proves the opposite. You are evaluating them on their skill which is observed and on their individual stats compared to others of the same postilion. Thats fine but it is NOT based on their win loss record as LOL. If it were they would be very poorly ranked, and we all know that would not be correct to do it that way. Then add ranking the best left guard? exactly impossible based on win loss record to do it accurately or fairly.

should the hall of fame induction be base don win loss? def not and thank god it isnt.

SmokedAlmondsSep 26, 20151 votes

{quoted}

Another thread misinformed about the true definition of "ELO Hell." People that cry about getting trolls in lower tiers aren't in ELO Hell. They're where they should be provided they're even in wins/losses (0-2% of games leeway either way).

Players that are incredibly negative in their win/loss ratio, meaning they were demoted quite a lot from what they were when they were first placed in the ranked system, are either:

Overplaced.

Underpracticed.

Stuck in ELO Hell.

Why the third option? Let me explain.

You win 7 games in a row. You get 12-15 LP per win, since your win/loss is terribad. That's 13.5 games on average, meaning you've gained roughly 95 LP. Not enough for a promotion.

You lose your momentum and drop back down to Earth. Only a little though. You lose 5 games. You lose 20-23 LP, since your win/loss is terribad. 21.5 average. That means you've lost about 107 LP.

tl;dr Your "reward" for going 7-5 is -12 LP when you're in ELO Hell. I know this because I've been in those fiery pits of darkness. ELO Hell is very real, and I'm tired of people that don't even play ranked telling me otherwise.

If you have a net loss in LP from going 7-5 its because you have been losing more than you have been winning and were on your way down.

The arbritrary sample size of 12 games you chose to focus on is far to small to mean much of anything. And part of the reason the system is the way is, is to reduce the amount of promotions and demotions from luck streaks. Play 100 games and go 60-40 and you will climb.

I urgot about itSep 26, 20151 votes

Thank you! I didn't want to explain what elo hell was actually

Horse GrenadesSep 26, 20151 votes

{quoted}

Another thread misinformed about the true definition of "ELO Hell." People that cry about getting trolls in lower tiers aren't in ELO Hell. They're where they should be provided they're even in wins/losses (0-2% of games leeway either way).

Players that are incredibly negative in their win/loss ratio, meaning they were demoted quite a lot from what they were when they were first placed in the ranked system, are either:

Overplaced.

Underpracticed.

Stuck in ELO Hell.

Why the third option? Let me explain.

You win 7 games in a row. You get 12-15 LP per win, since your win/loss is terribad. That's 13.5 games on average, meaning you've gained roughly 95 LP. Not enough for a promotion.

You lose your momentum and drop back down to Earth. Only a little though. You lose 5 games. You lose 20-23 LP, since your win/loss is terribad. 21.5 average. That means you've lost about 107 LP.

tl;dr Your "reward" for going 7-5 is -12 LP when you're in ELO Hell. I know this because I've been in those fiery pits of darkness. ELO Hell is very real, and I'm tired of people that don't even play ranked telling me otherwise.

OP is just uninformed on the definition of ELOHell

I urgot about itSep 26, 20151 votes

To me elo hell is when you win you get 3-10 and lose 15-20 It's elo hell cause your elo is so low it's basically hell

BudupopsSep 26, 20151 votes

well considering u r in bronze, ur post is INVALID

AlienPrimateSep 27, 20151 votes

Where did you get that from? I'm actually silver 2 right now and steadily climbing now that I'm playing ranked more. I was in bronze 4 at the beginning of season 4 and it took me all of about a week to get out of bronze using 4 champions a lot of people considered to be troll picks or worthless.

http://www.lolking.net/summoner/na/46058999#ranked-stats

BudupopsSep 27, 20151 votes

booo

AtaraxicOct 8, 20151 votes

Then u are not stuck. others are.

I just won like 22 out of 25 games. (i do not know the actual figures but u get the gist) my best streak ever, went from b5 to b4. its just the way it is. There is a gravity well effect in the code. it is MUCH harder to climb once you are stuck. and thats is elo hell.

In statistic there are always outliers, and if you are one of them u are in elo hell, if you are not then u just cant get it.

RueianOct 4, 20151 votes

I wont argue about elo hell. but the concept that in solo queue people's ideals do not mesh well is entirely valid. both teams have an equal chance of people not working together and causing poor choices to happen. One person might want to go in, and someone else might not want to go in. These clashes are the real reason why people get so frustrated, its not that they are stupid or bad, but because they made choices that just overall negatively influenced the game or as a team their choice didn't line up with eachother and you end up having a rift in the team.

solo queue players are not part of your hive mind. and someone has to lose the game.

AMYS GRAVEOct 5, 20151 votes

Chance of being on your team assuming you are not the troll: 44% Chance of being on enemy team: 56%

what is the variance on this number? with 10s of millions of players daily if the odds of being on the other side of 50% is 1 in 10,000 then that means there are 1000 people who are getting shafted by statistics. understand that 56% is only an average and this "mathematical" explanation is fundamentally flawed. 56% is the average for the entire community, not for an individual.

AlienPrimateOct 5, 20151 votes

With you asking about the variance of the number, it is obvious that you either didn't read the whole thing or don't understand it. No offense to you if you don't understand it as we all have things we are bad it, like for me, I am a terribly slow reader. Let me try to make this clear for you.

Q: Are you a troll?

A: No, therefore your team has 4 people who can be a troll and the enemy has 5 people who can be a troll; if there must be 1 and only 1 troll in the game which is what this whole post is about then your team has a 4/9 chance to get the troll and the enemy team has a 5/9 chance

These statistics are not looking at the whole community. It is only looking at a random sample of 9 individuals over and over again, so it is the opposite of what you said. If you take every game you have ever had with a troll and assume that it has been a fairly large number of games as some people make it seem with their complaints, then the statistics should look pretty close to this. The community as a whole doesn't matter for these numbers. If you look at the whole community on a big picture, then the chance of either team getting a troll is 50% when you are not accounted for in the statistics.

AMYS GRAVEOct 5, 20151 votes

https://en.wikipedia.org/wiki/Binomial_distribution

you are treating the problem as a binomial distribution with p=4/9 and q=5/9. therefore the mean is np and the variance on that is npq.

lets say you play 1000 games that have a troll in them, then the average # of games where the troll is on your team is 444 but the variance on that number is still 246! having an equal number of games with trolls on your team compared to the other team isn't even outside your margin of error after not 1000 games, but 1000 games that have a troll in them. Interestingly that fact does not change even if you play more games because the variance grows as n just like the mean.

11% more people (who don't troll or afk) are climbing compared to falling because of afks but that does mean that no one is falling and that anyone who claims otherwise is lying or wrong.

I'm a graduate student in the sciences, so i guess you could say that "About binomial theorem I'm teeming with a lot o' news"

SmokedAlmondsOct 6, 20151 votes

{quoted}

https://en.wikipedia.org/wiki/Binomial_distribution

you are treating the problem as a binomial distribution with p=4/9 and q=5/9. therefore the mean is np and the variance on that is npq.

lets say you play 1000 games that have a troll in them, then the average # of games where the troll is on your team is 444 but the variance on that number is still 246! having an equal number of games with trolls on your team compared to the other team isn't even outside your margin of error after not 1000 games, but 1000 games that have a troll in them. Interestingly that fact does not change even if you play more games because the variance grows as n just like the mean.

11% more people (who don't troll or afk) are climbing compared to falling because of afks but that does mean that no one is falling and that anyone who claims otherwise is lying or wrong.

I'm a graduate student in the sciences, so i guess you could say that "About binomial theorem I'm teeming with a lot o' news"

I'm not going to pretend that the math isn't over my head. But that doesn't accurately depict the situation. Because if you play 1000 games, 1000 of them aren't going to have trolls. Personally I've played over 5000 games and I'm not sure how many trolls I've seen but its probably around 10. Ill admit that seems to be below average, but its hard to tell because so many people consider things like building bork over bloodthirster to be trolling.

It also doesn't consider that matchmaking is going to correct, at least some of that imbalance. When you lose because of w/e external factor you presumably didn't get worse at the game, but your rank went down. So in successive games you will have a slightly increased chance to win. This obviously also can happen in reverse with wins you got carried or afks and such on the opposing team.

RyzeReworkOct 5, 20151 votes

i had a sliver match were i had a bronze 2 on our team and laned against a gold 3 if thats balanced then idk what balance means

AlienPrimateOct 6, 20151 votes

This is the reason that I don't think you should be allowed to duo with anyone who isn't in the same tier with you. The bronze was probably duo with a silver on your team and the gold was probably duo with a silver on the enemy team.