Detailed Analysis of LCS Superweek with Statistics!

In my last article titled “6 Things LCS can Improve on,” two of the biggest things I wanted to see from the LCS is more statistics and analysis of the teams. For the Superweek I parsed a ton of the data from the 20 games and they have lead to some interesting insight into the current meta of the NA scene. This article will break down the different team play styles, snowballing, and some bonus quick stats, EU vs. NA and champion analysis!

Teams

Out of all the teams performances in this weeks LCS, three teams, TSM, Dig, and C9, stood out the most in their performance and statistics. Breaking down things like FB (first blood), objective control, and total kills we get exciting insight into how their actions show their priorities as a team and create unique play styles for each team. Analyzing these statistics actually gave a lot of insight not just into the meta, but also into the teams specifically as we see different play styles and priorities coming from each team. I am going to highlight a few of their picks and play styles that really stood out to me.

TSM

The first interesting team to look at is the “aggressive” TSM. People always hype TSM’s aggressiveness and their playmaker Reginald, but ever since late last split TSM has been singing a different tune on their aggressive plays, focusing objectives rather than kills. Their matches had the slowest first bloods of any games, coming in at an average of 9:05 (compared to the week’s average of 5:40). They prioritize kills the least and they have even said in interviews that they don’t go after kills unless it will provide objectives. This is clearly seen as they have given up FB 4 out of their 5 games. Despite their games averaging the slowest FBs, they have the fastest objectives in the league. TSM took dragon all 5 games for themselves and averaged at the quickest pace (8 min and 25 seconds, compared to league average of 9:55). They also force slightly faster first towers at the pace of 6:34, compared to the average 6:54. This prioritization of objectives over kills shows in their picks as well as they often choose Shen or jungle Elise. Oddone’s Elise is a priority pick for TSM not just because he hit challenger with it, but because Elise is one of the best early dragon takers due to her spiders acting as the tank which is essential to mitigate the high amount of damage that Dragon does early game. Out of the 4 quickest dragons during the week, TSM got 3 of them - all with Elise jungle. TSM’s objective focused play style can also be seen in their kill/death stats, having the lowest team kills (54, 2nd lowest is VES at 57) in the league but also the lowest team deaths (45, 2nd lowest is VUL at 57)). While people have been saying “TSM is getting back to their roots from s2” I think this is untrue, as they played a tanky-dps team fight oriented style back then and their style now is more close to the Season 2 CLG style of “objectives over teamfights”.

Dignitas

On the other side of the coin, we look at Dignitas. Dignitas actually has the highest amount of kills per minute, at 2.43 (the average being 1.95). They also average the fastest FB time of any team, coming in at 4:23 compared to the league average at 5:40. However, at the same time they have by FAR the slowest tower and Dragon times out of any other team, averaging 11:53 Dragon time (vs. league average 9:55), and 8:53 average tower time (vs. league average 6:54). These slow times come from Dig’s tendency to 2v2 over 2v1. In the 20 games of Superweek there were only three 2v2 matchups in NA, every other game was 2v1 mid/top vs 2v1 bot. All three of these 2v2s were forced by Dignitas. Many may attribute this to “slow adoption of the meta” and think that this contributed to Dig’s poor performance this week yet on the contrary in all three of these 2v2 games Dignitas ‘won’ the early game and came out with a gold lead at 10 minutes. Both of Dig’s wins were in these 3 games and the 3rd was the CLG/Dig game which... we all knew what a mess that was. I believe that Dig’s 2v2 abilities will be very important going forward as I believe the 3.8 patch will bring forth more 2v2 style in the NA LCS.

#Cloud9hypetrain

The last team I am going to look at is the very hyped Cloud 9. Going into this super week I strongly felt that they were going to do well yet their play still surprised me significantly - not just due to the resulting 5-0, but in HOW they got that 5-0. Many people attribute their success to ‘replicating the korean scene’ but I believe that extremely cheapens their accomplishments. What Cloud 9 showed in this super week was not strategic brilliance, or extreme mechanical skill (like we often see from ‘new talent’ teams), but a very poised and flexible team who has great decision making abilities. Cloud 9 does not get the fastest towers or push the most out of any team nor do they get an early advantage every game. They prioritize objectives, but there is nothing exceptional about it. They have average FB timers (5:36), a bit slower than average dragon timer (10:57), a bit faster tower timer (6:20), and they only got out to an early lead 2 out of 5 games. This is what makes Cloud 9 the scariest team in the league. They play like an experienced team despite this being their first week in LCS, they don’t get flustered when behind, they don’t throw games that they have the advantage in - they just play solid with excellent teamfighting skills (as seen by them having the 2nd highest assist per kill average, only beaten by CLG’s inflated stats due to their long games and thus having more teamfights than the average team).

Snowballing

The next big topic is one that comes up a lot in interviews with players: the issue of snowballing. Often times a lane getting first blood is attributed to it’s success, or when they gave up that dragon it doomed them - but how strong is the correlation between these things? I broke down the “snowball” effect on 6 possible advantages and looked at their correlation with winning. These six factors were first blood, first tower, first dragon, first baron, gold advantage at 10 minutes, and gold advantage at 20 minutes.

• First blood was the most neutral statistic with the winning team scoring first blood in 50% of the games, therefore it seems to have no impact on the winner. Also, first blood didn’t have much of a pattern to it, with the average time being 5 min 40 sec and almost equally in all 3 lanes (6 top, 5 mid, 6 bot, 3 drag).
• The next statistic, first tower, was also fairly neutral with 55% of the teams scoring first tower going on to win the game. The first tower fell on average at 6:54, was way more often top and bot than mid (6 top, 3 mid, 11 bot), and was favored to the blue side (13 purple tower deaths, 7 blue deaths).  
• The most bizarre statistic goes to first dragon (average time 9:55), which had a 40% win rate. More teams who got the first dragon went on to lose the game than win the game. This statistic really confused me at first but I think this can be attributed to the successful counter-play that has come from dragon fights, as often times teams will trade towers or kills for this dragon, and it often puts the team that initially started dragon into a vulnerable position to be ganked.
• The last neutral objective statistic - first Baron - leads to the highest win rate as expected, yet still a bit lower than predicted, at 66.6%.  The average time for the first Baron is 28:32. Baron is becoming less of a snowball instrument for teams as pushing has gained increased priority. Baron has become more of a common comeback attempt, as 1/3rd of the successful barons were gotten by teams who at the time had a gold disadvantage. This is often due to the fact that after winning a teamfight as a losing team you don’t have the map pressure to take towers, and so instead they will take baron to give a few minute buffer to try and make a stronger comeback and deny objectives from the other team.
• The next snowball factor to look at is gold lead at 10 minutes. This was the most definitive early factor with 64.7% of teams with a gold lead at 10 minutes going on to win the game. Of these advantages, the leading team averaged 13.5k gold while the trailing team averaged 12.2k, which is about a 10% lead. The largest lead was by team CST over team VUL on the first day of Superweek, coming in at 14.5k gold vs 11.8k gold with CST winning the game. While many people would think that blue side would tend to lead early game due to the double golem advantages, it turns out purple actually gets out to an early lead more often at a rate of 76.47%!
• The last factor tracked was the gold advantage at 20 min. This is the biggest predictor of win likeliness, with 70% of teams at a 20 minute gold lead going on to win their games.

Champions

Here we’ll take a look at one of the most accessible statistics - champion picks. <Here> is a link to the basic stats of the champions picked in LCS. There are some crazy things emerging here. For example: Kha’Zix’s mega OP patch where he is banned or picked in every game and is undefeated (5-0) when picked (this mostly due to the fact that in patch 3.7 all of his competition like Zed and TF were nerfed yet Kha’Zix remained untouched). However, the biggest topic I am going to be analysing is what I see as the “break-out” picks of this Superweek - champions rising in popularity in their respective roles, and find what to look for in the future week's picks.  

• Top: Kennen. This was one of the biggest breakouts of this Superweek, being picked/banned 19 times. Previously, in the Summer Promotion series and the Spring Playoffs, he was only picked 5 times in 44 games (with no patch changing him since then). What sparked his popularity was the rise of Kennen in Korean matches. Kennen is often paired with an aggressive initiator like Zac or tanky AP counterpart like Ryze who both focus on boosting his teamfighting ability. Although he has a weak win rate at 44.4%, I expect to see more of him in the coming weeks especially from teams like C9, VES, and VUL.
• Jungle: Zac and Elise. These two junglers dominated the Superweek. With Elise and Zac being picked 12 times and 10 times, respectively, they were a very common sight (even going head to head 6 games!). Even though both of these champions can play multiple roles, they were chosen for jungle most of the time. Elise showed her excellence at jungle control and ganking while Zac excelled at diving and teamfight initiations. I think Zac really showed his strength this week and will be pick/banned even more next week as more teams adopt him.
• Middle: Kha’Zix.  While Kha’Zix was always a popular pick, he really showed his dominance this week by being the only champion picked or banned in every game and with 100% win rate with more than 3 games played. However, I think the most interesting thing in this week with mid lane is the variety shown. Mid lane is the thing left most up to the personality of the player, as we’ve seen things like Zed from C9 and CST, Kayle from VES, Ori from TSM and DIG, and even Xerath from VUL. Also, the rise of 2v1 mid lanes has often left “mid” players to the side lanes for 2v1ing.
• ADC: Draven. Draven has made a big appearance in this week’s LCS, spearheaded by the new blood of ADCs; TSM’s WildTurtle, C9’s Sneaky, and VES’s Maplestreet. Draven leads his team to an early gold lead of at least 500g at the 10 minute mark 75% of the time (6/8 games, 1 tie and 1 fail). With the 3.8 patch and the nerf of Spirit of the Elder Lizard I can see a fall of Ezreal and a rise in the League of Draven!
• Support: Thresh. More than just EDward aka the “Thresh Prince” has been choosing this champion lately. It seems that TSM’s Xpecial, C9’s Lemonnation, CST’s Daydreamin and VUL’s Bloodwater have had success with him leading to a 95% pick/ban rate, and being tied with the most picked champion at 12 games. His immense playmaking abilities as well as his proficiency at 2v1ing (which happened in 85% of the games) has really brought him front and center in the NA scene. A shoutout to Nami as well, who despite her 0-3 performance this opening week will get more and more play as the weeks progress.

EU vs. NA

While this is more focused on analyzing the NA games in depth, I think one of the most interesting aspects that are easy to highlight in statistics are the differences between the NA and EU scenes.

The most apparent differences are in champion picks. <Here> is the excel data for the EU picks. There are a bunch of differences between these EU picks and NA picks. In NA there are some very popular picks which EU teams don’t value very much. For example Kennen who had a 95% inclusion rate in NA only had 35% pick/ban rate in EU, or Elise who had a 95% rate in NA but only 35% in EU as well. Kha’Zix was a must-pick in NA having 100% pick/ban rate and 100% win rate, while in EU he was only banned or picked 75% of the time. There were a few other popular NA picks like Draven, Ryze, Zac, and Karthus that did not get near the same amount of attention in EU. Similarly there were several champions in EU that were not popular in NA. Nunu was one of the big ones, with 65% pick/ban rate in EU while only having a 10% pick/ban rate in NA. Also Shen, who was selected 18 times in EU but only 5 times in NA (mostly from TSM). The other EU picks that were popular are Nami, TF, Malphite, Lissandra, and Varus.

The next crazy difference in EU and NA play was how they set up their lanes. In NA they have a heavy 2v1 focus, with the lanes ending up in a 2v1 matchup 85% of the time. However in EU it’s a different story, with about 40% of the games being 2v2. While on blue side, NA sent their dual lane bot 17 times, mid once, and top twice. EU moved their dual lane more when on blue side, only going bot 14 times, mid 2 times, and top 4 times. The biggest difference was in how they chose their dual lane position as purple side. NA sent their dual lane top 7 times, mid 10 times, and bot 3 times. EU sent their dual lane top 9 times, mid only 2 times, and bot 9 times. NA really favors the 2v1 mid as they had 2v1s mid in 50% of the games, while EU really doesnt like mid 2v1, only sending them mid in a few games. I think this has to deal with the EU mid laners being the “star players” of most of their teams (xpeke, Alex Ich, Bjergsen, Ocelote,  Froggen, etc.) and they believe they can or should 1v1.

Statistic Quick Fires (Stats are from NA Superweek)

• Average time for an LCS game is 38:20, for a CLG game it is 52:46.
• If game time had a normal distribution , the probability of the 71 minute Dig vs. CLG game is 0.2% (1 in 335 games). [mean = 38.345, std = 11.95]
• Team Velocity has both the lowest kills per minute of any team (1.617), and the highest deaths per minute (2.838).
• At 10 minutes the gold lead went to the Purple team 76.4% of the time.
• The Blue team wins 60% of the games.
• The team with the lead in gold at 10 minutes continues to lead at 20 minutes 76.4% of the time.
• TSM & CLG gave up FB in 4 out of 5 of their games, while CST got FB in 4 out of 5 of their games.
• While NA 2v1’s in 85% of games, EU only 2v1’s in 60% of games.
• 47.7% of all champs were picked or banned in the first week

Thank you for reading. If you want to see the full statistics of all the data I used for these numbers, go HERE.
I also wanted to give a big thanks to @jjordizzle for his help on editing and feedback and some help from @zeroaurora_hf and @zerglinator for help on the EU stats.