Boards›Gameplay

Placements are a possible contributing factor to the potential match making complaints

SilverEagleTwo·2/24/2020, 6:29:38 PM·1 votes·2,394 views

Based on observation of my games, below is just 1 instance.

An example of two lane matchups from my last ranked game

Concessions I realize this is just one game, but I'm monitoring I agree, I have no idea how Matchmaking works. Special Note: Darkdemon, Kai Guy, and Rujitra. - Let's agree to disagree. None of us are Riot employees. I acknowledge your disagreement with anything to do with how Placements and Matchmaking work. Please feel free to skip the reply.

Observation

  1. 52% Win rate, Current I3, Former B3 Top vs My top, a 40% win rate I1 player (possibly new or newer player).
  2. 52% win rate current B2 and Former S3 Enemy Mid vs our 16% win rate new Iron Mid

Conclusion: Placements are creating the situation of putting former players in the wrong division which creates a strain on Riot's matchmaking. Players end up with the wildly unfair match his which lead to theories on Smurfing and anger in general from the player base.

Suggestions: Since I'm complaining , here are my thoughts. But If Riot magically came to me and said pick 1? it's REMOVE placements at the start of the year and let us finish where we end.

  1. Remove placements. Let players start where they finish.

    1a) A lot of former Silver and Bronze players are spread all across Iron and Bronze. It appears (just a theory) that MMKing doesn't not account for prior season. Therefore, matchmaking gives the scenario above, I think we could all agree that MMR wasn't the same or close in either match, although, closer for the Top matchup (given 1 appears to be a new player, and the other appears to have 1 season under their belt).

  2. Address and tighten up your Match making.

    2a) Let's first accept this premise: for the scenario above, that both players were new players who hadn't placed in a prior season. Then why would that situation happen? A B2 52% win rate vs a I1 16% win rate and a I3 52% win rate vs a I1 40% Win rate? that seems off. I can't imagine their MMRs were even close.

    2b) Maybe consider including prior Season in Matchmaking and identify first time accounts (accounts with no prior ranked season, low level accounts) versus (accounts that have 1 or more seasons logged in Ranked on that account).

    2c) Win rate should be (if it isn't already) part of the equation. No where would any competitive situation would you place a player who is 2 -12 (16%) vs a 8 - 6 or 7 - 7 (between 57% and 50%) win rate in a "ranked" situation.

          2ic) I concede this sort of thing happens during the regular season in a lot of major sports, so it's a tough sell. However, I think it's pertinent for LoL, and should rarely if ever happen.  

     2iic) Think maybe Tennis players - does a 2-12 player ever face a top person in a Tournament? Heck, do they even Make the Tournament?

    2iiic) Last conession - This mismatch can sometime happen in Chess Trounaments. However, it happens because of a lack of players. If you have 15 players show up and 14 are rated 1600 and one is rated 1200, that 12 hundered is in for a    long tournament. I dont' believe (no proof, just feeling) that Lol has that problem I don't want to guess to concurrent players during ranked solo but I bet there are enough players to matche Bronze with Bronze, Iron with Iron and Silver with Silver.

  3. Maybe stop placing I4, I3, B2 and S4's in Ranked games together regardless of the "MMR".

    3a) I could see a B4 ranked match that maybe includes I1's who are playing above their I1 rank. And maybe even I2's. However, and I3 and I4 should never be in a B4 match. Nor should a B4 player be in a I3, I4 match. Isn't that what Promotions and Demotions are for?

         3ia) If an I3 player is winning enough to be matched in B4 games, after x amount of wins in I3. Assuming he's not in Promos - then skip him to mid I2. If he continues to win at the same rate, bump him to Mid I1 so on and so forth until he hits the wall.

      3ib) That formula might be different for higher ELOs, say above Gold. Yet if you did this in plat or diamond, I'd imagine the player might hit the wall faster.

7 Comments

Kai Guy2/24/2020, 10:43:45 PM2 votes

If riots placements and soft rest are rank only then its basically a worthless cosmetic feature that adds grind for no reason and generates player frustration. Ranks are what we are given to view skill so its what's used to build perception of fairness.

Sidenote. My beef on MM QQ is about people who promote toxic mentality with no relation or comprehension of the basic theory so they make claims that lack any relevance to the topic. I frequent MM threads to reduce the # of Players who give up because of "Elo hell" or " Rito forcing a loss time to f2f." Ruins gameplay. Its not a competitive game if people abandon competitive behavior. Any one can play for a loss.

The only productive use of placements I can see is if riot uses them with a soft reset to maintain rank values. Try to build it so bronze in season 8 represents in game skill of bronze in season 10. With the introduction of iron last year... I am on the fence if their current dev does that.

My suspicion is its mostly for perception and to compensate for the massive demotion protection range and to keep folks playing a min # of games to check that the meta shifts did not have dramatic impact on an individual.

Subdue2/25/2020, 1:38:46 AM2 votes

Win rate is not and should not be explicitly used in the matchmaking formula as it is already taken into account based on your MMR. Take two people, we'll call them John and Bob. John and Bob are exactly equal in skill. We'll say they're both Diamond 1 players. However, Bob has been playing for 5 months longer than John this season, so he reached Diamond 1 3 months ago, and has since been winning about half his games there, as expected. Thus, while Bob climbed to Diamond 1 with a 65% win rate, by now it's closer to 53%, as he's played hundreds of games since reaching Diamond 1 and won about half of them. John on the other hand just reached Diamond 1 at about the same pace as Bob. However, he hasn't played that many games in Diamond 1, so he's still sporting a 65% win rate. These are two people who are exactly the same level of skilled, but because one has simply played more games, his win rate is over 10% lower.

SilverEagleTwo2/25/2020, 2:12:56 AM1 votes

  1. Are you Riot Employee? If no then prove your statement about Win rate. I suspect you can't and therefore you're just like me, basing your argument off your observation. I'm making mine based on my own game experience, where are you basing yours from?

  2. I never said it should be explicitly used in the MM formula.

  3. you're comparing like situations, which wasn't the case in the example I provided.

The closest match (just considering Win Rate alone) was 52% v 40%, which on it's face isn't so bad. However, If we trust that both players have 1 account and additionally we accept the only games they have played were on these accounts, then it's easy to conclude (based on that observation alone) that MM needs tweaking.

Because It matched a player winning at 52% and a full season of ranked under his belt. (S10 being his 2nd season). And in that prior Season he was ranked B3. Where as the guy on my team is a new player. No prior Season under his belt. That on it's face is a mismatch, and I don't think knowing the MMR at this point even matters.

It doesn't matter because in this game that Experience counts for a lot.

I know this because of how terribly clueless I was in S6 Ranked compared to S7 ranked.

But to dispel your point a bit further, the MID matchup? That was straight up criminal.

A 52% win rate player, who was in B2 this season and in he prior Season Silver 3! Vs a new player with a 16% win rate. 16%!

It's why I rail so much against Placements more than Match Making.

Before I can safely conclude that Matchmaking is an actual problem I need Riot to remove placements. Otherwise without knowing exactly how it all works it's all guess work.

it's the whole reason I play on this 2nd account this year. I figured I'd use this account to figure out my Best most consistent supports. Once I had enough games I'd take that into my Main account and start placements Fresh with the knowledge I gained from this account. I was hoping for 7-3 or 8-2, and while I did place better 6-4 (instead of 4-6 on this account) I didn't win enough and I stated a couple times now I think the avg W/L for Placements in my skill group is 4-6, 5-5, and 6-4. I wanted to prove that if I got High Bronze or Silver from a 7-3 or 8-2 W/L record I placements then I'd have another solid point to argue at Riot. But I failed.