An In-depth Contemplation about the League of Legends and Matchmaking.

AngryBearNoises·2/24/2015, 11:34:21 PM·1 votes·231 views

Hey everyone. This is one of my first posts despite playing League since mid-ish Season 2, but. Basically, the purpose of this discussion is to share my thoughts and see what other people think about the matchmaking in league of legends. Some of you might be thinking, what brought this topic to mind? I just completed a ranked game earlier, in my provisionals, and it really got me thinking. When I got into champion select, I noted that I preferred not to adc as it was my worst role but that I would fill what was neccesary.

Naturally, this meant I was the adc. I got a Leona as my support. Now what caused me to contemplate this specific subject was the fact that this Leona started with a red crystal, rushed a ruby sightstone, did not use said sightstone often, and did not know when to engage and when to back off. For example, zoned to our tower, Rek'sai came to gank, pinged for us to engage. I, playing as Graves, Quickdraw (E - A short dash that increases aa speed for several seconds) my way into the fight as soon as Rek'Sai uses Burrow (I forget which hotkey it is, but it is a dash) into his Unborrow knock up.

Leona hugs the tower while we fight, then when we retreat low without a single kill because of no hard cc, Leona flashes past us and uses her skillshot stun two-on-one and gets destroyed before we can even turn around to try and go back in.

So not only did Leona not buy a gold income item leading to low income for more items, she also didn't know how to play.

So, with my situation understood, the question remains: Why did that get me thinking of this particular subject? And the answer, is because I couldn't understand why that Leona thought these things were a good idea, even when being given advice on how to play better. I concluded that it was just a sheer difference in mechanical understanding of the game, and that I was out of my tier.

Which is when I began to contemplate the matchmaking system. Now, from my understanding, League of Legends uses a 'Total MMR' Matchmaking System, in other words. In SoloQ, let's say you have 800 MMR (That is the bronze V standard ELO). The matchmaking system will attempt to find people within a range under or over that MMR for the others playing for you team in that game. Then, once four others are found, your total sum is added up and it looks for another team with similar Total MMR as yours.

So there are three people with 400 MMR, one with 1000 MMR, and yourself with 800 MMR. That's 3000 MMR. Your enemy team has one person with 1200 MMR, one person with 1000 MMR, and three people with 300 MMR. The matchup is heavily skewed, but in total MMR, there's very little MMR difference.

This can be a very, very unfair thing. Chances are, half of your team will do averagely against an opponent of similar skill, but the other half will be stomped on. This makes for a very..unfun, experience for everyone involved. So with that in mind, let's look at theoretical purposing behind this after we get another bit of matchmaking information out of the way.

A Winstreak is, essentially, just like a killing spree in game. The more MMR you get as you keep getting matched against consecutively tougher opponents, but in tandem if you lose, you forfeit equally scaling ELO (or MMR). This makes sense, on paper. The system seems to be designed to try and keep your Win/Loss Ratio at 5/5 (or 50% each) and awards you with ELO for getting over that hurdle and maintaining your winning streak. On the other hand, what about a Losing Streak?

This is where things get complicated. Remember that bit earlier about Total MMR matchmaking? Well, just like your winning streak keeps attempting to be broken, so does in theory, your losing streak. By earlier noting, this would be possible by matching bad players with good ones by lieu of Total MMR, allowing the bad ones to be carried out of their loss streak. But.

This Can Be Counter intuitive.

Because in the end, League of Legends is a cooperative game, and it then becomes imbalanced as if you are the only competent player on your team, you have one person doing their job. And one person, cannot perform five roles at once.

So. How would this go about being fixed? The answer, is as it is, it really can't be unless the system is converted from Total MMR to P2P MMR. In other words, every team would need to have exactly the same MMR Between its members. If one has 800, another 900, two with 300, and one with 1000, then the enemy team would need to have exactly that amount as well, and converting to this system would create massive wait times. Like, five to ten minute minimum wait times.

Total MMR unfortunately, is the most efficient matchmaking system right now.

So, with that concluded, what are your thoughts? How do you think matchmaking could be improved, or what do you think matchmaking does in terms of relevance to win/loss ratios?

--Sincerely, VVendettas Veigar

1 Comments

Matthias91192/25/2015, 2:12:06 AM1 votes

So there are three people with 400 MMR, one with 1000 MMR, and yourself with 800 MMR. That's 3000 MMR. Your enemy team has one person with 1200 MMR, one person with 1000 MMR, and three people with 300 MMR. The matchup is heavily skewed, but in total MMR, there's very little MMR difference.

It is veeeeeeeeeeery unlikely the matchmaker would put together a team with such a large difference in skill unless the widely-separated players queued up together. They do try to get (adjusted) "total MMR" close, but they also try to minimize the spread within each team. In Ranked matches it's rare to have a spread of more than ~2 divisions from the lowest to highest player in the match if everyone is solo queued, and that's normally only a few hundred MMR at most.

When group queues are involved they make some other adjustments (it's not just a straight numerical sum or mean of their MMRs, and there are bonus applied for group size), but unless they can find a similarly-spaced group to put on the other team it's always hard to fairly match up groups with huge skill disparities.