Boards›Gameplay

The REAL problem with ELO/MMR/Dynamic Queue

Gabelous··4 votes·1,209 views

TL:DR

  • Doing anything other than solo screws your MMR.
  • System actively punishes you for using dynamic queue system, regardless of where you are and who you're help is.

The REAL reason dynamic queue is fair is because it secretly PUNISHES you for not going solo. So let's consider the ELO system, the system used today by nearly all competitive games to establish player skill. Now to be fair the system was originally created for CHESS players, so its accuracy is only really for games where it is one player against other player. Generally if you win as much as you lose (50% win-rate) the system judges you "where you should be". And logically if you lose more your ranking drops, but if you win more your ranking goes up. Unfortunately this doesn't work with Rito. And now my example...

I was Bronze 5, and have a coworker who was Bronze 4. We duo'd a lot and had a 7-win streak. Also when I wasn't duo'ing with him I was duo'ing with another bronze 5. So how's that data come out?

Overall Winrate : 48% Winrate last 20 games : 75%

Now under the normal ELO system this would mean I would be "placed higher" since while my winrate is 48% my past history shows "improvement". Normally you gain/lost 15LP or so for "evenly matched games". My entire experience in bronze 4 so far? A win nets me only 10 - 12 LP, while a loss loses me 22 - 24 LP. My present example I was at 95LP, and after losing Bronze 3 series earlier I would be granted another win. When this happened in Bronze 4 promos I only lost 12 LP, down to 88LP so one more win would put me back in series. When I lost my Bronze 3 promo? I lost 33LP... Now some claims the system "shouldn't" dock you more than 25 LP for a lost series, and yet here I was tanked even more. So that's another indication the system is borked. So let's see my present play.

95LP -> 64LP - It was a heavy bruiser/heavy CC team comp with an ADC top. We were gonna lose that anyway. Their compe forgave a lot more mistakes in teamfights than ours did.

64LP -> 42LP - Again a loss, we could have won that but some "try hards" felt it better to split-push and be a hero than be there for teamfights. So we kept losing teamfights and ground while that try-hard didn't make any progress.

42LP -> 55LP - A win, and almost threw it because the carry wanted to farm nearly the entire time. Fortunately Jax jungle and Tynrd top combined with my Karma shields was enough.

So I lose 23LP, lost another another 22LP and only gained 13LP on a win. Despite a recent 70% winrate and a 48% winrate the system claims I'm "above my ranking". Even though I'm not boosted and duo'd with other Bronze who weren't more than one division above me the system still labels me a Bronze Five and is actively trying to pull me down with less LP wins and more LP losses. And if you think THIS is bad, then think about all the "boosters" who are getting help a tier higher than them, much less a division.

So technically the system doesn't cater to boosters. Anyone who boosts will later be dependent on their booster to keep their place, or they get lucky in solo and hope their W/L ratio can keep them afloat. It's a "smoke and mirrors" system, while they encourage you to have friends and queue dynamically the system will punish you if you leave your friends as well.

14 Comments

B00B002 votes

tldr? i mean i have read it but all i read was the point loss and gain not the actual reason as to why you feel this screws you over, have i missed something or glossed over something?

i will check lolking when the maintenance is over but i am assuming your MMR is really low and this is what determines how much you win/lose and who you are matched with.

Gabelous1 votes

I use na.gg. And to be fair the "average MMR" is dynamic as well. I have seen bronze 4 range from 999 to 1,099, and usually settles between 1,010 and 1,040. My point was, on the basis of the ELO system the MMR is based of off I SHOULD be climbing. The only dependency in the data as I can see if I duo with others, and who I duo with seems irrelevant. If you do more than solo then anything you do while duo technically doesn't "count" for calculating MMR/ELO.

B00B001 votes

can you link me your na.gg? please? i asked as i was going to see if i could find an answer to your woes.

CoolMike1271 votes

Obviously, when you play a lot together in ranked (or just in general), you grow dependability and know the strengths and weaknesses of them. You should also be performing better because of the extra communication and synergy you posses. When you play on your own, you will notice the difference in gameplay from your friend(s) versus some stranger, and lose that extra communication and synergy you had when you were together.

Realistically, it is easier to play with friends, but if you are adaptive and skilled enough, you should be able to win games on your own, without them.

The REAL problem with Dynamic Queue is the fact that the system calculation of mmr is averaged out of all members in the premade and paired into a game with the enemy being around the same mmr average.

Dynamic Queue atm is practically Elo Boosting that is legal.

PDE5 Inhib1 votes

I think your LP gains/losses are determined by your MMR. Your MMR is viewed as shit since you lost so much. It's the system's way of keeping baddies from climbing if they happen to go on a lucky streak.

Foxstep1 votes

dude you got a low winrate, this is basically the only reason why you lose more than you gain.

Gabelous1 votes

My winrate is 48%. Given the ELO/MMR system uses a 50% winrate as a baseline is isn't that low. Especially when I have seen people with 43% winrates in the same division as I with a better MMR somehow.

Foxstep1 votes

{quoted}

My winrate is 48%. Given the ELO/MMR system uses a 50% winrate as a baseline is isn't that low. Especially when I have seen people with 43% winrates in the same division as I with a better MMR somehow.

Everything under 50% is low. The other ones below you do have an even lower winrate. The system works like this (rough explanation inc):

got more than 50% winrate? get more lp per win and less per lose got exactly 50% winrate? get as much as you lose got less than 50% winrate? lose more than you gain

Conphucius1 votes

everything under 50% is low?

i have 54% win ratio

i lose 20 lp per loss i gain 4 lp per win

how?

(in case its needed, http://euw.op.gg/summoner/userName=conphucius)

Gabelous1 votes

{quoted}

Everything under 50% is low. The other ones below you do have an even lower winrate. The system works like this (rough explanation inc):

got more than 50% winrate? get more lp per win and less per lose got exactly 50% winrate? get as much as you lose got less than 50% winrate? lose more than you gain

Okay how about this.

http://na.op.gg/summoner/userName=cptnstarfish

This was someone I supported and got an easy win on. 44% winrate and yet MMR is "normal". Shouldn't it be worse than mine at the same level?

LiveYiOrDieHard1 votes

Uhh no, your MMR just sucks, that's why you lose more than you gain. Winrate means nothing by itself. What matters is who those wins are against.

Gabelous1 votes

{quoted}

Uhh no, your MMR just sucks, that's why you lose more than you gain. Winrate means nothing by itself. What matters is who those wins are against.

That is my point. Under the normal ELO system (the system ALL competitive video games use in one way or another) my MMR should be "where I am at" or at least not as bad as it claims. And yet RIto's system says I'm worse than I should be statistically speaking.

LiveYiOrDieHard1 votes

{quoted}

That is my point. Under the normal ELO system (the system ALL competitive video games use in one way or another) my MMR should be "where I am at" or at least not as bad as it claims. And yet RIto's system says I'm worse than I should be statistically speaking.

???

Listen, the LP system is like wrapping paper that goes over your real Elo/MMR rating. That MMR rating is where you belong. The LP system tries to piggyback off it and show you sort-of where that rating is at while also giving you the illusion that if you "grind out" enough games you're going to go up (even if you don't improve). But just grinding out games won't increase your MMR. You have to actually get better.

Your MMR right now is basically bronze 5. That's where you should be. But since you're bronze 4 in the LP system, it's going to give you less points for a win than it takes away for a loss. That's its way of trying to catch up to your MMR, while still making you think "gee I just gotta grind out more games to go up!"

LP is basically just a psychological trick by Riot to get people addicted to grinding out those games, to make them think they're going somewhere even when they're not actually improving.

Once you get better, you will start bringing your MMR up again. And then your rank in the LP system will follow.

Gabelous1 votes

{quoted}

???

Listen, the LP system is like wrapping paper that goes over your real Elo/MMR rating. That MMR rating is where you belong. The LP system tries to piggyback off it and show you sort-of where that rating is at while also giving you the illusion that if you "grind out" enough games you're going to go up (even if you don't improve). But just grinding out games won't increase your MMR. You have to actually get better.

Your MMR right now is basically bronze 5. That's where you should be. But since you're bronze 4 in the LP system, it's going to give you less points for a win than it takes away for a loss. That's its way of trying to catch up to your MMR, while still making you think "gee I just gotta grind out more games to go up!"

LP is basically just a psychological trick by Riot to get people addicted to grinding out those games, to make them think they're going somewhere even when they're not actually improving.

Once you get better, you will start bringing your MMR up again. And then your rank in the LP system will follow.

And how do I get better? As a support main? I rely on my teammates to help and what if you get someone whose just bad? Jungler who ends up feeding? Tanks who don't want to initiate? Players who don't see objectives? That person who wants to split-push because he saw it done in LCS but he doesn't have the build nor the time to do so effectively? I can only do so much on my own as a support. Zone out bot lane, punish mistakes, help my carry farm. But at the end of the day as a support you are ultimately reliant on the rest of the team to help. And if they don't/can't/won't then your MMR tanks despite how good you are. And that's implying a pool of players who ACTUALLY take rankd seriously, there's a whole sea full of trolls/kids/etc. who are there just to either make your game miserable or do whatever the frack they want because the system can't punish them fast enough nor hard enough. And even THEN there's boosters, and even with the 1 tier limit it is evidently more of a problem than ever before.

And then we come to the other side of the coin, the "carry myself" mentality. And that mentality has actually BUILT the "try-hards" who generally are worse players. Those mids who take whichever champion is OP at the time, or the jungler who thinks he can 1v4, and even the top who thinks he can split-push when in reality they can't. They become so "try hard" at trying to carry themselves to better MMR they are actually WORSE. They blame teammates for their losses and never look at their own gameplay mistakes because, in their eye, they cannot make mistakes.

So what then? How "fair" is this system? A system which measures "individual skill" in a team-based game? How do you measure that with any sort of accuracy? If you factored in winrate for MMR, then I should only be a little below average for where I am (and improving) and yet the system has deemed I should be MORE below average despite the statistical signs of improvement. With the only anamoly being I was duo'ing with a bronze 3 friend. And if you looked at those games HE benefited more than I did because the system will inherently pair us with worse players in my division which would net him more "easy wins".

It is moot now, suffered a complete string of losses because for some reason the ranked bronze pool is filled with people who feel the need to break the meta. from facing veigar support trying to help an AP teemo to out ADC going AP Karthus and then calling ME a troll for going Caitlyn as a support. It's been like that pretty much all day.

My findings only validate if you really want to "climb" you have a choice few champions to pick as jungle or top and "carry yourself" out. Doesn't make you an actual "better" player, simply you know how to take advantage of loopholes within the system. Just like REAL America!

The lesson here is MMR and skill are not synonymous, and being diamond doesn't mean you are "good" and simply you are better that exploiting the system.

More from Gameplay

3 Refunds per season, upvote if you want 3 refunds per season.
Gameplay764 comments
7,167
votes
3 Refunds per season, upvote if you want 3 refunds per season.
Gameplay156 comments
3,510
votes
UPVOTE FOR URF
Gameplay606 comments
2,440
votes
Yes plz.
Gameplay456 comments
1,721
votes
3 Refunds per season, up vote if you want 3 refunds per season.
Gameplay114 comments
1,716
votes
When is ANYONE at riot going to address the awful new player experience?
Gameplay634 comments
1,383
votes

More by Gabelous

ELO/MMR/LP System Reality
Gameplay7 comments
6
votes
Dealing with Toxicity = Time
Gameplay3 comments
1
votes
Upcoming "Remake" Function
Gameplay3 comments
1
votes