LoL: The reasons for performance-based matchmaking

Kayle running from Soraka and Nunu.I’ve been on a crazy string of losses recently and I’ve been trying to sort out why. I can point to all sorts of things, but a lot of my problem can be attributed to the new map. I’ve been trying out some weird comps and checking into alternative hero combinations to try to pull out some wins. There have been a few occasions, though, where my defeat was almost surely due to my teammates.

Take my daily lunchtime game today. I was playing Ezreal solo in the middle lane on Summoner’s Rift. Everything was going well. I was harassing Ashe down to gankable levels and, had I not forgotten that my Cleanse was on “D” and my Ignite on “F” and not the other way around, I would have easily taken First Blood. Sadly, the bonus gold went to the opposing Warwick for ganking our Ryze on bottom at level 4.

Immediately Ryze started calling out the team for his death, mostly blaming his lanemate, Heimer, for not using turrets as wards. Though I disagree that Heimer should waste turrets that way, especially when wards are only 90 gold, I kept my mouth shut so as not to start one of those inevitable team implosions that leads to a quick loss. It didn’t matter. Ryze was constantly in bad positions on the map despite our encouragement to back off and before 20 minutes he was 0-6 with less than 20 creep kills.

I try not to call people out too much because everyone can have a bad game, but this is something different entirely. This was a player who ended the game 0-8-0 and spent the entire match blaming his teammates for the loss. I decided to take a quick look at this previous 10 game stats and here’s what I found.

Defeat: 0-8-0
Defeat: 2-5-5
Defeat: 1-7-6
Victory: 6-7-6
Defeat: 1-16-8
Victory: 8-11-11
Victory: 4-12-8
Victory: 1-4-12
Victory: 6-10-13

As I said, I’ve been on a losing streak, so I tried to remain objective, but look at those numbers. 1-16? That’s abysmal even for a new player. His total stat count for the last 10 games is 29-80-69. Now I realize stats rarely show the whole picture, but 29-80 is a scary teammate to have, especially for someone who picks DPS toons a large majority of the time. I think his only non-assassin game was one as Morgana. The rest were Ryze, Anivia, Twitch, and a random Heimer. Those stats suggest a player that doesn’t at all know the limits of his health pool compared to his DPS or someone who is perpetually in bad map position with regard to the rest of his team. The assist count is nice, but again, most of his toons have some form of AOE, so it’s not that surprising.

In the spirit of fairness, here are my stats for my own previous 10 matches:
Defeat: 2-1-3
Victory: 3-2-7
Victory: 7-1-6
Victory: 6-3-8
Defeat: 6-5-2
Defeat: 0-3-0
Defeat: 3-5-3
Defeat: 3-7-0
Defeat: 5-5-3
Defeat: 3-3-6

For total stats we have 38-35-38. Obviously that’s much more balanced, and this is one of my worst losing streaks in months. That 3-7-0 game is ugly, and I admittedly was playing like a jackass that game, but in the context of my other stats you can easily see that it was rare misstep in a string of decent performances.

All of this is to say that the I continue to be unimpressed with the ELO system. Basing a player’s rank solely on whether he wins or loses leads to crazy matchmaking results. I want to post on the official forums and beg for a performance-based system, but the reality is its just too hard. There are too many factors to consider when looking at stats, and you can bet there would be an army of angry players that want more credit for a win that the stats say they barely contributed to. The more reliable solution is to find a group of two or four other players with whom you can regularly premade for some kind of reliable ELO. The rest of the time you’re going to catch a lot of players who have hit the right games and made it into their respective ELO brackets.


Related Posts

2 responses to “LoL: The reasons for performance-based matchmaking”

Leave a Reply

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>