Semifinal LCS Elo Rankings
Before I begin mourning over the multiple upsets that butchered the quarterfinal Elo predictions, I’m going to give a brief overview of the rankings in order to clear up some reoccurring questions.
Elo ratings are a system used to measure the relative skill level of head-to-head competitors. In this case I’m using them to rank the teams in both the EU and NA LCS.
Why are they all in gold?
I’ve set the league average to be 1500 Elo, so teams will fall in the spectrum from the high 1300’s to the low 1600’s generally. I used 1500 because it gives an easily visible range in skill with a nice round average. This isn’t meant to be compared to the ranked Elo of way back in seasons 1 and 2.
The gains and losses seem arbitrary. Why is that?
This statement (that I conveniently turned into a question to keep this FAQ style) couldn’t be further from the truth. The gains and Losses all happen according to the Elo formula. Elo is self-correcting, which means that if a lower rated team upsets a higher rated team, then they will gain much more Elo that the higher rated team would have. To put that in context, if Team Liquid, currently at 1529, upsets Cloud 9 and their 1630 rating, Team Liquid will gain far more Elo than Cloud 9 would if Cloud 9 won. Keep in mind that in the regular season teams would play two games a weekend, so if a team went 1-1, depending on the caliber of teams they played they might still gain or drop a decent rating, even though they went 1-1.
Are playoff games worth more?
Yes. In the Elo equation there is the constant K. This can be manipulated to amplify the gains and losses of a match. It stays at the same value for the entirety of the regular season, and scales up through the playoffs, giving more weight and significance to the games the farther a team makes it in the postseason.
Do complete stomps count for more point?
This question wasn’t actually asked by anyone, but I’m just going to answer it anyway. No, they don’t. Unlike the World football rankings, or the NFL rankings, League doesn’t have scores that we can use to establish a clear universal margin of victory modifier. Some games are complete stomps with a bunch of kills on one side, but others are just quick, decisive 25 minute wins with very little in the way of a kill score, that are over so fast there isn’t a massive gold lead built up by one side. Without a consistent way to compare the level of domination I decided to just transfer this idea to the playoff series. Series count as one game, but depending on how many games a team wins the series by their end result gets amplified. Take the series by 2 games and the constant K gets multiplied by 1.5, take a series by 3 games and K gets multiplied by 1.75. This way teams that crush over a series get a nice reward, and a losing team that takes it to game five isn’t punished unfairly.
That's going to be it with the clarifications for now, but I'll probably publish a dedicated FAQ article later to clear up as many questions as possible.
Friendly heads up: If you didn't bother reading any of the clarifications this is where you should probably start reading again.
Last weekend was brutal. The Elo predictions went 2-2, CLG’s fall created a large possibility of a repeat of summer 2014, Copenhagen wolves finally fell in line with their Elo and UoL brought out Shaco as they smashed the higher rated Gambit squad. Not even one series managed to go the full five games.
Team Liquid was the big mover of the quarterfinals, jumping up two spots over both the early exiting teams. Team Impulse closed in on Team Solo Mid, and while they didn’t pull even, this bodes well for an exciting semifinal. Gravity and CLG both slid down in the rankings as a result of their early exits. There is still a slim chance that they move up, as Team Liquid only has a marginal lead over those teams, and would undergo a slide of their own if Team Liquid fulfills the curse and ends up in 4th place.
Team Solo Mid
Cloud 9 has the highest rating in the NA LCS, which doesn’t bode well for Team Liquid’s chances. Predicted to win 64 percent of the time, expect Cloud 9 to either sweep the series in a repeat of last summer’s semifinals, or win in a 3-1 decision. While the series might not be all that close, these games could be a good measuring stick for Team Liquid fans to see how well Piglet is meshing with the rest of the team.
Team Solo Mid vs. Team Impulse looks to be a much closer series, with TSM predicted to be victorious 54 percent of the time. Team Impulse has been on a consistent rise the latter half of this split, up to the point where they’re slim underdogs against the number 1 seed of the playoffs. This series has a good chance to go the full five games given how good both these teams are.
Team Solo Mid
The EU LCS saw a lot of movement in the rankings, so much so that I had to show all ten teams. Copenhagen Wolves sweep at the hands of H2k Gaming sent them down to the 10th spot, just 0.3 Elo behind Roccat. This boosted up Meet your makers, Giants Gaming and Roccat all up a spot. Unicorns of Love and Gambit Gaming swapped spots after Unicorns 3-1 series victory.
Unicorns of Love
Meet Your Makers
The EU semifinals are on opposite ends of the spectrum. On one hand you have the series between SK Gaming and Unicorns of Love, which is predicted to be the most lopsided of all the LCS semifinal series. On the other you have the Fnatic and H2k series, which is supposed to be the closest series in the playoffs so far.
While a 3-1 or a 3-0 decision in favor of SK Gaming is the most likely outcome, Fnatic and H2k gaming look to end their series 3-2, with Fnatic barely edging out H2k in the end. This looks like the series to watch of the weekend, with Fnatic’s hyper-aggressive strategies facing off against a red-hot Ryu and the rest of the surging H2k squad. Fnatic is favored to win in the end by a slim margin, but this should be the closest series of the spring playoffs to date.
Unicorns of Love