Boards›Gameplay

So, I require proof

IP Masquerena·2/16/2020, 1:36:24 PM·6 votes·2,272 views

So, morning ~Nyaa

I wake up, check my other thread, and decide to browse around and get a laugh at the threads around when I stumble across this gem of a thread.

And one of them comments made me laugh a lot, but then I was like "but where did he get it anyway? I never found any actual testing done to prove it".

The comment is as follows:

Yes, they do it on purpose to make the climb longer.

No its not a theory, it's been proven.

yes you have a right to complain.

All the "proof" I can find around isn't really proof, it's just "it happened to me as well" type of thing.

So, I now ask the great minds of the Ego of Legends boar...League of Legends boards, to show some actual proof.

Let's see it, as currently there is nothing tangible. Everything points to the MM working as intended, BUT, if it is truly rigged (AHAHAHAHAHAHAHAHAHAHAHAHAHA) then there have to be experiments that were made to prove it is truly rigged, right? Collected data from them, statistics, etc.

The biggest "proof" people seem to state is this comment that can be found in that thread:

Their system attempt to force everyone to a 50% winrate within 1%

But, isn't that "49%-51%" the expect win rate you should have IF you are at the skill level you're supposed to be? Also, 51% is more than enough to allow you to climb.

Let's use the cutest and most modest person on the boards as example, myself ^-^

Ain't I cute?

Sure, Gold 4 in 58 games starting from B3 ain't "a big accomplishment", I'm still debating if I should care about ranked or not this season, so I kinda tryharded it early to get to Gold 4, so if I decide to not care about it, I'm already set to get the rewards, while if I decide to see how far I can climb, I have ample time to do so.

But notice the win rate so far: 67.2%.

Now, what I'm expecting to be told is: "58 games isn't enough for the MM to screw you, you need to play more until your win rate gets forced to 50% by the system".

But.....if I play more games with this win rate, I'll keep climbing, which will make my win rate lower over time as I keep facing better and better players until I reach my plateau where my "true skill level" is. When I reach that point, of course I won't be able to climb as fast as now anymore, because that's where I belong, only way to climb further would be to get better at the game.

Then there's the other possible argument: "You play Akali, Ahri and Neeko, of course you climb with OMFGBBQOPMETABROKEN champs", so hey, I decided to go take a look at the OP of that thread, most played champion: Cait, 35% wr over 31 games, has a general winrate of 51% mostly being carried by his games on Jhin, Sivir and Viktor who have quite high win rates (why he keeps refusing to play those more is beyond me)

Then I decided to look at the other two guys I mentioned, sadly the guy with "we have proof" doesn't play ranked, so can't say for him, but for the guy with "they balance it to force it to be a 50% withing 1% difference", most played champs? MF and Veigar, but despite those two, he may soon drop from S4 to B1 because of his 47% general winrate as he decided to start feeding on Senna.

So playing "OP" champs is clearly not the reason why my win rate is as high as it is, because other guys play OP champs as well, heck, those champs they play are statistically OP in the ranks those two guys are, but they still ain't climbing in double to even quadruple my game count.

So, I inquire the people here who believe that there is "proof" of Riot rigging the match-making, to please provide it, as I'm so heavily curious right now.

Until proof arrives, I bead you all farewell.

37 Comments

Paroe2/16/2020, 3:40:03 PM3 votes

Preface: The system will attempt to keep you at a soft 50% win rate, not a forced (hard) 50% by weighting your score on the team. This means that theres a lot of wiggle room, but over your last twenty games you should have either a positive or negative 50% win rate, where negative 50% means youre losing more than youre winning, and positive means youre winning more than youre losing. (Also known as climbing and struggling to climb) Edit: Also, in a better system you would be closer to a true 50% - 9 to 11 for example - when you reach "the rank you should be at", rather than the biased 13:7 or 7:13.

So unless im looking at this wrong... Youre currently at 39 wins, 19 losses. Over your last 20 games, you won 13 and lost 7 (counting a remake as a loss).

Do you see a pattern there? Youre at roughly a 50% total win rate, and over your last 20 games.

Lets take a look at ya boi Sneaky, too;
107 wins to 91 losses. 13 wins to 7 losses in the last twenty games.

And another streamer nightblue3; 220 wins to 231 losses 7 wins to 13 losses,

And hashinshin... 152 wins to 153 losses 7 wins to 13 losses

Solo renekton only; 53 wins to 80 losses 6 wins to 14 losses

Aphromoo: 91 wins to 104 losses 7 wins to 13 losses

Are you starting to see a pattern emerge, my EUNE contemporary? When i look at your games, i see a fairly regular pattern in your last ten games: Win 2, lose 1. Also, i see you have a favorite duo. Are you actually solo, or are you duoing this season?

Hazardus ducees2/16/2020, 2:50:59 PM2 votes

The problem here is that people are often paired with other players who don't correctly range in other's elo. Everyone has a … "numeric value" if you will, and each of these numbers are identified by your W/L Ratio. Your elo is determined and averaged by what you did within each and every game and sadly, this applies to every single rolee and special games where your team steamrolled and also got steamrolled. The value isn't determined by what you did, but what the team did under the time limit. With this ratio being effected by your elo, you are often paired in a very bad way. What I mean by this is that, if you have a high win ratio, then the enemies you're paired up with/against will also have these same ratio- but not the same ELO. Unfortunately, this ratio is actually stacked against you with your team though. Why? Because if some low ELO fellow has the same win ratio as you (who is in silver 4), they will be paired within your game, on your team. While this works with the enemy team too, it is less likely for everyone to have the same W/L ratio. Most of the times, these low ELO players had their "assumed elo" heightened because their win streak and steamrolled about 37.655% of them.

As of this point, there are youtube videos to prove that it doesn't matter how good you are if you have a win streak. My elo is often identified with those of platinum, but I'm always placed low ELO due to the teammates I must depend on to end with me. For some reason, low ELO enjoy drawing out games as long as possible, which in turn makes their lead die.

IVSakenz2/16/2020, 2:51:34 PM1 votes

The only thing wrong with the match making system(in ranked) that I can tell is it lets to big of a difference in ElO in the same match. In a gold Elo game you can get anywhere from Plat to low silver team mates.

In normal draft it is even worse.. how are you going to put a Brnze1 jungler against a plat+ or diamond.. how would it even be possible for that to be a fair game for anyone involved.

Matchmaking needs to be fixed

2gudaiya2/16/2020, 4:13:54 PM1 votes

the thing people are complaining about is that the system places them one standard deviation lower than they finished last season. and now they have anxiety tihnking they won't complete the climb or something. it's generally a lot of noise not worth hearing.

Kai Guy2/19/2020, 8:04:24 PM1 votes

more details on the math and such you used (tho, if I may inquire, could you word that as if you were talking to an audience yourself in like a TEDTalk fashion? maybe start a new comment chain, just for the more space?

No idea what a Tedtalk fashion is.

The math I used was Binominal Distribution. You might see it called a Bernoulli process. Its a method of calculating probability for 2 outcomes, usually a Yes or no / Success or Failure. It needs a fixed amount for trials and for a result to not impact future results. The formula. b(x; n, P) = nCx * Px * (1 – P)n – x

Its filled in as b = binomial probability x = total number of “successes” (pass or fail, heads or tails etc.) P = probability of a success on an individual trial n = number of trials

Its a little ugly to see it expressed that way so here is the more common way to see it. https://imgur.com/a/YOKybzI Calculators are very easy to find with a google search if you hate dealing with factorials.

What about it do you want to know?