Welcome to a statistical breakdown of the NA LCS!

What makes teams different from one another? TSM and CLG fans could probably give you thousands of differences, but how exactly do their playstyles vary? What are their strengths and weaknesses? Today, we’re going to look for trends utilizing a few different metrics and see if they tell consistent stories.

We will be using three different numbers, as explained below:

The Percentage Gold Lead is calculated relative to an average of both teams’ gold. It creates symmetrical percentages in both gold leads and deficits for the teams in-game.

We will discuss these statistics in both winning and losing contexts, but my overall focus is weighted towards the latter. It’s easier to make good decisions and “play well” when ahead, but when teams lose, they are pushed to make risky decisions and do a better job of determining a team’s strengths and weaknesses.

A team’s Gold per Minute (GPM) tells us how overbearing a team is in victory and how well they keep up when falling behind. This a simple chart, but a few pointers on how to read it wouldn’t be amiss:

• Scores are an average of GPM in all Summer Split games up to Week 8 (20 games).
• Blue represents all games for a team, Green represents victories, and Red represents defeats.
• Teams are sorted by current standing in the League.
• Small differences are key – This is a per-minute statistic; Small aspects of gameplay do add up over time.
• The purpose of this chart is for people to compare how different a team’s wins and losses are. To see Victory/Defeat versions of this chart, click here. Teams are ranked by numerical score.

Many teams appear close at first glance, but we can already determine a few points. These numbers will also help us frame later statistics:

• TSM and EG are the most efficient gold producers in victory. All of the top teams are fairly close, but CLG is half-a-step behind, and Cloud 9 below them, with compLexity getting noticeably much less team gold when they win. Both Cloud 9 and Complexity have had some dramatic comebacks this season, which partially explains this trend, however, this may also have something to with their overall play style.
• CLG appears to be the best team at ‘treading water’ while losing, followed closely by TSM. On the other hand, Cloud 9 presents some what of a contradiction as they tend to hold lower gold totals in their wins, though they are one of the best at staying even with their opposition throughout games. This unusual attribute of Cloud 9 will be an an important point of focus later.
• Dignitas is the team that seems to lose the hardest amongst the top teams – It has been said that Dignitas has not yet learned how to play from behind, and this statement is supported by the data. compLexity and Curse also struggle in this regard, but curiously, EG manages to stay afloat like a top team. The explanation for this probably has something to do with Altec’s ridiculous CS numbers, with EG commonly allowing Altec to freeze lanes and take all the farm, however, this may or may not be healthy for the teams overall win-ratio, despite their gold gains.

Minion farming is surprisingly divergent amongst teams and we can see here how teams balance farming and objectives. Having a high CS score can be viewed in two ways. On one hand, it can be seen as a team possessing better laning mechanics and making more efficient usage of the map. On the other hand, it can also be interpreted as an overemphasis on freezing and individual play,  which leads to rotational mistakes: For example, champions dying under tower in 3v1 or 4v2 dives.

• Once again, this is an average of a team’s score over 20 Season games.
• 1 MPM can also be seen as 10 Team CS in 10 minutes.
• Click here to see Victory/Defeat versions of this chart. Once again, teams are ranked by numerical score there, as opposed to present League standing here.

We have looked at the GPM in broader terms, and this will now allow us to investigate the MPM for individual teams.

TSM are the kings of farming. In fact, even when they lose, they still beat the winning average of half the League (CLG, C9, COL, and EG). For comparison, LMQ and CLG are the lowest-farming teams in defeat, and the difference between these two and TSM is about 1.5 minions a minute.

That might not sound too big, but let’s put that into perspective. Over the course of a 30-40 minute game, that’s 45-60 minions respectively, and at 20 gold a minion, TSM are gaining 900-1200 team gold to help catch up a gold deficit! That can be roughly described as TSM finding three minion waves per carry, while losing, that CLG or LMQ choose to give up (or have forcibly denied from them). From a player’s perspective, that’s a sizable chunk of gold and experience.

Dignitas is another a team that demonstrates strong farming skills, but they don’t quite compare to TSM in their losses. Shiphtur’s ability to keep on trucking regardless of how the game is going has been well-documented, so it’s most likely Zion and Crumbzz who are falling behind. ZionSpartan often struggles to find farm when behind (although the same could be said for many top laners), while Crumbzz wants to play a supportive vision-based jungle game, a la SKT T1 K’s Bengi.

This is an instance where the GPM and MPM do align and tell roughly the same tale; for other teams the picture is more complex.

EG is a team with unusual numbers. They manage to, as an entire team, farm almost exactly the same regardless of whether they’re winning or losing. It’s hard to draw a conclusion here, because EG has only won five games – Not a large sample size! We can recall from recent memory that one of their few wins (against Curse) was a comeback, so at least in that particular case they were playing like the “losing EG” for the first half of the game. This picture will become clearer after we examine the gold leads below.

This attribute of EG is not seen when global gold is in the picture. Their team GPM is not noticeably closer between win/losses or somehow different from what you see from other teams. They actually have the highest team GPM when they do win, while their losing GPM is pretty solidly average.

Cloud 9 is even more bizarre – They actually farm quite a bit better when they lose! There are two possible explanations. One is that Cloud 9, as a team, likes to trade objectives and accelerate the pace of the game. They are confident that they can come out on the better end of these strategical battles.

However, it could also be that their execution is sloppy; while they are ultimately victorious in these games, their rotations are costing them a lot of farm. Perhaps in their heyday Cloud 9 did a better job of preserving their farm while still playing a strong objective game.

Another attribute of Cloud 9 is that Hai likes to roam for vision and ganks. The obvious price is that he gives up his personal farm to do so. The way we’re going to tie both objective trades and Hai’s map play into the stats is that Cloud 9 only goes for these things when they are winning. When they are not winning, they default to farming their lanes like a more “normal” team. As seen in the overall gold statistics, they do a pretty good job of keeping up when they do play like this.

LMQ is the lowest farming team in defeat and fourth when they win.  The overall difference between winning and losing is about 2 MPM, tying them with Curse for largest disparity. Like other Chinese teams, they don’t turtle when behind, and instead choose to scrap it out – Let the chips fall where they may. Outcome aside, LMQ often throws their top laner Ackerman into a more supportive and roaming role rather than having him farm, which likely explains their winning and overall average.

In overall GPM, LMQ is pretty average for both victory and defeat. Perhaps their rotational game is better than people give them credit for, because they certainly don’t get much farm when they’re having trouble.

CLG probably wouldn’t describe themselves as similar to LMQ, but they are in that they don’t farm well in losses. Link and Seraph are usually the ones to fall behind on CS. When CLG loses, they are often described as looking “invisible” or “shut down,” but on overall gold, we see that they actually do the best job at staying in the game. Is farming while behind such a priority after all?  These stats would indicate that perhaps it is not, at least for CLG and LMQ.

The gold leads will show if the GPM tells the truth – Do CLG and LMQ really keep up that well without farm?

Curse is pretty different between when they win and when they lose – Around 2 MPM, again, like LMQ. This can be attributed to some of their wins being with big split pushers, such as Yasuo and Jax (for Voyboy and Quas respectively), that spend all game farming the side lanes. Note that their team GPM while losing is the second lowest in the league, while being 5th in losing MPM. This suggests that their issues are slightly more weighted towards the objective game than the farm game.

Complexity is the team with the overall fewest minions and lowest average GPM, in spite of having a better record than EG. Along with Cloud 9, Complexity has a had a few intriguing games this season where they managed to come from behind without actually swinging the gold total (Hello Robert’s six-item Tristana). Do these teams somehow do more with less?  Both have been seen making some strong rotational moves, but maybe it’s possible to chalk up Complexity as mere turtlers – The truth probably has aspects of both.

What you are about to see is the % gold lead by the minute, for every game a team has played. This is what truly matters when you consider trying to use the GPM and MPM from above. It’s been 20 games per team or 80 games total so far, and this is a lot of visual information to take in. Here’s how to read them and what to pay attention to:

• Blue lines represent winning games, Red lines represent losing games.
• Steep lines characterize blowouts, while shallower ones show difficulty closing or a good defense.
• If a blue game is mostly below the median (0%) it’s a comeback, and if a red game is mostly above, it’s a throw.
• Fun Fact – 6.25 % of games (5/80) were won by the team with a gold deficit.
• A few of the longest games have been cropped for better visibility (Those games went to full builds anyway).

(Large infographic-sized image; open on second screen if you have one)

So what does this all tell us?  How can we relate it to GPM and MPM?

CLG previously looked pretty good as the team with the best losing GPM. However, this is distorted by one prominent throw in their match history (against Cloud 9). In this game, their gold lead reached 20% (!), and in fact, even at the end, CLG still held a gold lead. This is worth talking about as a distortion because CLG is tied for first. Ergo, if what few losses they have include a sizable throw, then there is a large impact on the losing average. They also have two other noticeable throws (against Curse and LMQ).

That said, CLG is also a team that has mounted several successful comebacks. It’s one thing to look good while losing and another thing entirely to actually overcome a gold deficit. Two of these comebacks have been from as far as a 20% deficit. CLG has the second best compilation of comebacks overall (after Cloud 9).

It’s been noted before by others, but a glance at the statistics confirms that CLG has the fastest average game time amongst the NA LCS, win or lose.

LMQ also have throws in their season that confuse the statistics and make their losing GPM seem better than it is. However, their story is more the sum of multiple games. LMQ have come back before, but they also have a number of games that abruptly end when the gold deficit becomes too much. As previously mentioned, LMQ directly fights for their comebacks, instead of turtling or playing the map. If LMQ is too behind to scrap with their opponents, then they simply lose.

LMQ has a few blowout victories, but they also have a couple curious games where they do not close well. In those games, we see them above a 10% gold lead without actually closing the game for 10 or more minutes. The cliche Chinese-playstyle explanation would be that LMQ is too kill-hungry to finish games properly and over dives (occasionally leading to throws). It could also indicate strategical issues beyond greed.

TSM are now much more definitively shown as the team with the closest losses. They almost never have a gold deficit of more than 20%, which is far better than every other team. They do have one significant throw against LMQ, but even that doesn’t skew the data by too much – LMQ’s gold lead in that game was never far above 10%.

That said, they don’t have many actual comebacks. They have two close games against Curse where the gold lead traded hands a few times, but none where they clearly pulled themselves out of a hole. TSM’s games are very much decided in the early and mid game. If they are ahead they win, and if they are behind they lose – They are most consistent team in this regard, and consistently at the top of every per-minute statistic.

The fact that they are so good at stopping the bleeding means that TSM should, in spite of not having a real comeback, be ultimately viewed positively here. On paper, conceptual improvement could make them the team with the best comeback potential.

Cloud 9 is one of the best closing teams in the LCS. Cloud 9 wins if they have a gold lead by 15 minutes (8/9 games AKA only one throw). The anatomy of their losses confirms the GPM and MPM statistics. Cloud 9  is very good at staying even when they lose (almost as good as TSM), but that doesn’t mean that they are approaching those games correctly.

Cloud 9 has a lot of close losses in high GPM games (as shown earlier). High GPM games are a hallmark of a slow laning phase and emphasis on farm. The opposite end of the spectrum (early objectives) means early advantages at the expense of farm. This means that Cloud 9 is not playing to their strengths in these games. They are instead playing the game their opponents want. Cloud 9 farms well on paper, but usually loses when they focus on it.

This can also be seen from a picks angle, because Cloud 9’s wins are heavily linked to Meteos getting Lee Sin (the most powerful early-game jungler). When Meteos cannot play Lee, Cloud 9 does not have the leverage to create early advantages. Instead, Meteos somewhat defaults to his old farming-based style, which he has tried to move away from. The emphasis Cloud 9 places on Lee Sin (or at least has thus far) indicates that they are aware of their preferred play-style.

The numbers suggest that Cloud 9  needs to play the early objective game and accelerate the pace. Their losses stem from giving their opponents too much breathing room in the early and mid game.

Dignitas have the largest number of blowout victories. They have multiple games where they exceed a 30% gold lead before 30 minutes. Dignitas is an investment bank. When they have money, they make more money (a lot more in some cases). Their work on the offensive vision game has made them (in terms of gold) the best at exploiting their leads.

Their ahead game is not perfect though – We can still see some throws. They also have a couple of harsh losses – Some of which hit the rough 30%. Dignitas still loses hard, and does so on every metric. While a team like Evil Geniuses has many losses, they have a lower percentage of “hard” losses than Dignitas. Over half of Dignitas’ losses approach blowout status, which is only surpassed by early-season compLexity. Dignitas could perhaps improve their comeback game if they focus on controlling the pace and finding farm when they are behind.

Curse has trouble closing games. This can be seen in both victory and defeat. Two of their eight total victories show relatively long plateaus, while many of their losses are very close games (within a 10% deficit). Curse doesn’t throw, which is a plus, but they seem to struggle at creating a lead.

Despite what is seen in GPM or MPM, Curse often stays very close to their opponents when they lose. This is similar to Cloud 9, but because Curse has low GPM in these games, they are actually opposites. It means that Curse is losing against opponents who are also earning low. This can be made to fit the same story that we saw in the GPM-MPM disparity (Curse does better at farming than in overall gold). The problem Curse has is against early game rotational pressure.

To re-iterate, when early objectives are taken in League, it’s often at the expense of someone’s farm. Overall team GPM is usually higher in slower-paced games because not as much time is required if objectives are taken later. Therefore, Curse’s low losing GPM is because they are playing the accelerated game, but they manage to keep the deficit itself relatively small. Is Curse initiating these early game objective trades? If so, it’s possible they would be more comfortable in slower games with an emphasis on farming – Again, the opposite of Cloud 9.

compLexity is interesting because they they can win without much of a gold lead. Earlier, we mentioned that much of compLxity’s recent success involved turtling, and here it’s blatantly on show. In fact, compLexity has twice won while still in a deficit. Even without those games, compLexity has much less domineering wins. They only have one game (a blowout of Dignitas) where their gold lead surpassed 30%, and only one more where they went above 20%! Most other teams have significantly larger gold leads for longer periods of time when they win.

They also have two distinct kinds of losses. Early in the season they got crushed several times – Here we see short and steep deficits. Later however, when Complexity adopted the way of the turtle, they became much better at controlling their ultimate defeats. This change coincides with Brokenshard’s visa woes, but it is also marked by a shift in the marksman meta, and I think the latter is responsible.

The patch cut the early power spike of these champions (BT) and moved players towards hyper-carries. Having a weaker mid-game ADC means that snowballing and taking turrets is more difficult. Additionally, Robert seems particularly comfortable on these hyper-carry champions, such as Kog’Maw or the aforementioned Tristana.

Evil Geniuses throw a lot. This can be seen in both a positive and negative light. The good side is that the “worst” team in the League is scrappy, and very much so in their games. On paper, EG is an LCS-level team and not outclassed by their opponents scoring highly on many of the metrics we look at above. They have many losses, but they fight on, and the shapes of their gold deficits are not too steep.

The downside of this, of course, is that throws are losses, and EG already has plenty of those. Losing a game with a gold lead is a dubious enough distinction, but EG is actually on the butt-end of two awkward games. We have our explanation for EG’s strong minion scores when losing, and it’s simple – They should have won some of those games! EG is a team where, like compLexity, winning and losing games are less distinct from each other. This is in contrast to a team like TSM, who wins and loses when the gold says they should.

Also like compLexity, EG has, as of late, become very turtle-y and oriented around their marksman Altec. The difference is that they have not had as much success in the games where they stall it out, and they do still require an actual gold lead to close.

So now you know that TSM out-farms CLG, and that Cloud 9 wins when they don’t farm.  Complexity doesn’t need a gold lead to win.  Cloud 9 and Curse are mirrors of each other; both lose close games, but one wants to accelerate the game while other prefers a slower pace. Dignitas both wins and loses hard.

This is a pioneering attempt at comparing teams through statistical analysis. It’s a chance to take a fresh look at the numbers behind League and make new discoveries about the players.

We want to create identities for the teams they play on and have a justification for the  stories that have been told previously by mere observation. However people spectating games are always going to have their perceptions of a team colored by bias, both conscious and subconscious. Here, we can see real and objective differences between how teams play the game. What aspects of a team are unique and special, and what deficits hold them back? I hope I have at least gotten the ball rolling towards the answers.

I’d like to thank the following people who helped make this article possible.

• Riot Quickshot (Idea)
• DiffTheEnder (Methods of Analysis)
• Spencer (Data Collection)
• SamwiseGamG (Art)
• Fridgecake + Valkryie (Editing)