NA LCS Team Stat Analysis: Gold Distribution and Objective Prioritization

This is the first release in what will hopefully be a weekly series of articles looking to get to the heart of what a winning team looks and plays like.

This is the first release in what will hopefully be a weekly series of articles looking to get to the heart of what a winning team looks and plays like. Future goals of this project include devising a single advanced statistic similar to the baseball statistic Wins Above Replacement (WAR) that accurately depicts a player’s total value to their team and developing a database of games, player performance and statistics with the intention of improving the overall quality and availability of in depth analysis. This includes tracking how new patches effect previously established trends. While the small sample size of games played make it a little too early to look at individual players, we’ve decided to kick off the project with a brief look at team by team, game by game gold distribution from week 1 of the NA LCS followed by a quick look at objectives.

Team Wide Gold Distribution

Click for Hi-Def Resolution

The above graph illustrates the gold distribution of each team, in each game, sorted from highest to lowest within each role. Green shading indicates the team won, while red shading indicates the team lost. The average gold distribution for each role is in the blue box at the bottom of the column. As we are only a week into the season, small sample size limitations apply so the number of inferences we’ll be able to make are limited. But a few do stick out.

The thing that stands out most in this chart is the extreme correlation between Support Gold Percentage and winning the game. When a team’s Support had a higher than average percentage of their team’s gold the team won 90% of their games. Even with the limited sample size this would seem to indicate that getting your Support involved in the game is key to winning. Supports are responsible for vision control, initiating team fights and peeling for the ADC, so it stands to reason that teams need their support to do well. I just didn’t expect such a high correlation. Going forward, we’ll be looking to establish whether Support success is a cause of team success or a result.

The Marksmen position is an interesting case. It would appear that over-investing and under-investing in the position leads to defeat. Between 23% and 25% appears to be the sweet spot. Investing more than that starves the rest of the team of gold and prevents them from fulfilling their responsibilities. A great example of this is CLG’s loss to Liquid where their front line lacked the tank stats to fulfill their responsibilities and collapsed. Investing less in the position results in a lack of late game damage. 

The gold distribution chart does a good job of pinpointing each team’s identity. Liquid places a heavy emphasis on their bot lane and Jungle, which makes sense as they build their compositions around their ADC, and Xpecial and IWD are the team’s two shot callers. What stands out is how well Quas performs despite having the lowest gold investment of any top laner. The impact of this cannot be understated as having one player be so effective with comparatively little gold means that Liquid can have average or better investment in every other role. This is something that a shallower roster simply couldn’t do, as the low investment would cause the player’s performance to eventually plummet (Hello, CLG Top Laners).

Per their reputation, Cloud 9 invests a lot in Sneaky and Meteos while starving Hai of gold. With the NA LCS looking more competitive than ever, it’s quite possible that Cloud 9 needs to reinvent themselves again. One of the easiest ways to win a game is to put the opposing shot caller off their game. Having Hai get so little gold may be one of the causes of Cloud 9’s rough start to the season. Clearly their heavy investment in getting Balls going was ineffective in both of their games. It’s also a departure from past Cloud 9 strategies that have taken a more balanced approach to the game.

That TSM and Dignitas invest heavily in their Mid laners surprises nobody. But such an investment means that if that player isn’t able to carry, the team will lose. When Bjergsen was unable get the kills he needed in team fights with Team 8, TSM had no fallback option. TSM is going to need to diversify their strategies if they want to surpass their past successes. For Dignitas, it all comes down to whether Shiphtur is willing to make the plays his team needs in order to win. Teams that Shiphtur has played on have consistently invested heavily in getting him going with mixed to poor results. While it might be a good idea for Dignitas to try prioritizing a different player, their alternatives aren’t great. The Top Lane is hard to carry from, their bot lane is new to playing together and Crumbz has been struggling since the middle of season 4. It will come as no surprise to Dig fans that they performed much better in the game where fewer resources were devoted to Crumbz.


Click Here for Hi-Def Resolution

First Blood Win Rate: 70% 

First Blood went to the Support and Mid Lane in 3 games each and the Jungle and ADC in 2 games each. The Top Laner did not get first blood in any game.

1st Dragon Win Rate: 80%

Earliest First Dragon: 9:14 (Team Impulse against Dignitas)

Latest First Dragon: 19:33 (Team 8 against TSM)

This is a stronger correlation than expected considering the low impact of the first dragon at early levels. My guess would be that it has more to do with teams that get the first dragon being better at map control than it has to do with 1st Dragon being worth a lot. In short, 1st dragon is not the reason the team’s won, but a side effect of the reason the team’s won.

First Baron Win Rate: 50%

The primary cause of this is probably that teams haven’t figured out the best way to use the new baron yet. Cloud 9 got first baron in both of their games and failed to win either of them. Gravity and Impulse won in the games they didn’t get first baron, and lost in the games that they did. Team Liquid seemed to have the best grasp of how to split push correctly to maximize the number of baron-enhanced waves attacking the enemy base. I expect this to improve over the course of the split.

First Tower Win Rate: 60%

First Inhibitor Win Rate: 100%

First tower has historically only been a lukewarm indicator of team success. On the other hand, first inhibitor has always had a very strong correlation with team success, so it’s unsurprising that every team that got first inhibitor went on to win the game. Not a single losing team managed to even get an inhibitor, which indicates how much harder it is to turtle and come back in season 5.

Going Forward

We are currently tracking a number of individual and team statistics that we will be rolling out in the coming weeks once we have a larger sample size to work with. On an individual level, we’re tracking each player’s ability to gain an advantage over their lane opponent, if their team made an investment in gaining that advantage, and whether that investment was worthwhile. There is limited gold available to each team, and we want to know who they’re giving it to, and whether that decision was effective. Eventually we’ll arrive at the ideal gold distribution for each team to be successful, and the variances between each team. On the team side, we’ll be trying to find a correlation between specific objectives and winning the game. As an example, getting an early mid or bottom tower typically leads to victory, while an early top tower has little correlation with victory. The key will be identifying why these things lead to victory. The primary purpose of taking down towers is to gain map control. Taking Top Tower provides additional control over the baron area, but as baron doesn’t spawn until 20 minutes, taking an early top tower isn’t as effective as taking an early mid or bottom tower since those towers increase map control over the dragon area. In addition to establishing a relationship between specific objectives and victory, we’ll also be looking for patterns in each team’s play to see whether we can predict what objectives they’ll go for at specific times. If you have any feedback, or suggestions for the project, feel free to comment in the section below.