onGamers News

Analyzing how gold is distributed for each team

A look at how each team distributes their gold and how this illustrates each team's strategy and star players.

With the Spring split nearing a close and with the wide data set we possess, now is a good time to look at how each team distributes its gold between players. The way this was calculated was by taking the average gold distribution between roles (21.1% of gold to top laner, 18.4% to Jungler, 22.3% to Mid, 23.2% to AD Carry, and 15.1% to Support), then from here we looked at how each team distributed their gold, and looked at how that compared relative to the averages. The follow represents how each team compares to the average:

This illustrates an interesting contrasting in playstyles for teams, and reaffirms some of the theories flying around. An example is how Coast feeds into their solo lanes, as we can see they have the highest proportion to top and a high amount to mid (so much so that NintendudeX’s lack of gold is actually “off the charts”). In the opposite direction, Dignitas has their mid and top significantly below average while the other three roles are well above average. One important thing to note is that having a higher than average proportion of gold isn’t single-handedly a measurement of skill, it is more representative of a team playstyle.

-Spellsy

Thorin's Take on the Graphs

What's interesting about these graphs is that fans will have a perception of who the star players or focal points of the teams are, and some of the players within those teams might knowingly maintain that status quo, but when you see the numbers illustrated like they are here, then you see who the teams are actually actively considering their stars.

When I first looked at these graphs many of them made sense in terms of the players widely considered the stars, but a few were really off, so let's run through them one by one:

TSM: There are no surprises on this graph, TSM's carries are Bjergsen and WildTurtle and each gets a similar amount of gold. Dyrus is well known as the "go even" Top and TheOddOne has never been about amassing gold.

C9: This graph has one of the first big surprises. Everyone could already have guessed that Meteos would have the most gold, that's been the formula for C9's success since they burst into the LCS. What would not have been anticipated is the degree to which Hai is a non-factor as a focal point and Balls is so heavily invested in. There has been a lot of hype around Hai, thanks in part to his play against fnatic at BotA, with some claiming he would be in the mix for best NA mid laner when Bjergsen arrived.

Spellsy's average gold distribution stat says that most teams give their Mid laner slightly more than the Top laner, yet here we see C9 massively prioritises Balls over Hai. Balls is often praised for consistency or his ability to play on a couple of champions, but this graph suggests that C9 themselves think of him as their sleeper carry, since most people are still going to think of Hai as the second star of the team after Meteos.

CLG: History would tell us that Doublelift should be the overwhelming gold leader, yet we actually see him behind Aphromoo in that regard. It's interesting that CLG are another team where the Mid laner is very hyped up, yet again we're seeing barely any gold going in that direction. Link gets even less than Nien, which is a surprise, again going against the stat for the average amongst teams. It is worth noting that Link’s numbers might be a little higher as the season goes on, as Chauster and Hotshot were playing mid for some of the games included in these stats.

Curse: Voyboy is the star name on Curse and yet many people in public will describe him as a legacy player and claim that Curse is keeping him around for his name and his streaming success. This graph seems to suggest there is some basis to that theory, as the majority of the gold is going to AD Carry Cop. What's surprising is that many people would have pointed to Quas as a potential carry on Curse, yet he is dwarfed by IWillDominate's gold, the latter often being the target of criticism this season. Public perception would tell us that Curse is a Top and Mid-centric team, yet this graph suggests they think of themselves as a Jungler and ADC focused side.

Dignitas: In line with what I said about Voyboy, the lack of gold going to "Star" Mid laner scarra in this graph is astounding. Scarra has always been the name you associate with Dignitas and they have always looked at their best when he was on his comfort champions and performing to a high level. Imaqtpie being the guy with the most gold should not be such a surprise, even with scarra's huge profile in the community, I think most people watching Dignitas since they entered LCS should confirm that qtpie has been the real carry for Dignitas in the last year or so.

A big surprise if how much gold is going the way of Crumbz, another jungler getting a lot of gold and yet also being criticised recently. What's interesting is that the graph for Curse and Dignitas is reasonably similar, yet few people would think of their teams the same way.

EG: This graph is in line with current thinking, in that the two young guns of EG (Poebelter and InnoX) are thought of the talent of the team and the carries and the graph confirms this. From a historical standpoint, it's interesting that yellowpete gets the least gold, since during CLG.EU's heyday he was often built up for the late-game to be the second carry, after Froggen. Then again, that was the prime era for ADC strength, whereas the position is weaker than it has ever been right now.

XDG: It should come as no surprise at all that this team is all about focusing everything into mancloud, it was his play that carried them so far last split and if they are going to return to the upper echelon of LCS NA then it will be off him returning to similar levels of form. What is worth noting is that the lack of gold going Xmithie's way. He was known for being the best NA proponent of that S3 low-econ style of jungling that the Koreans popularised, but he is largely credited as the second star of the team. So, in this era, one would expect him to be getting more gold. Then again, he hasn't been playing to the same level, and only just came back to the Jungle, so that is a consideration.

In a world where IWillDominate and Crumbz get so much gold, it does surprise me a little that less focus is put on Xmithie in that sense, though. Xmithie has only just come back to the jungle, so some of this may be a holdover from the Zuna Jungle era.

Coast: This graph is as one would expect, with the Mid and Top as the heavy carry positions for Coast. When we look at the breakdown of their gold in wins and losses though, we get a very interesting distinction. When Coast win it's when Zion is the monster carry who gets all the gold and kills everyone. When Coast loses, that's because Zion gets next to nothing and they go with Shiphtur to try and carry. Some of this goes down to the streaky nature of Zion's play, but it is pretty interesting in light of people's perception that Shiphtur is the star carry who needs a better team.

-Thorin

Data collected with help by the onGamers Stats team: Derek 'Kathix' Adams, Steven 'whedgehead' Falgout, Kent 'Traepoint' Frasure, Jake Morales, and James 'PelkaSupaFresh' Pelkey. Design by Ben 'Sarcasmappreciated' Li.

16 Comments
Posted by Varega

I love these stats, I think it's super informative, and a great format to read it in. Kudos dudes.

Posted by TommyHoare

How I would sum up Coast: Shiphtur always does 1.5x what he's required. Zion either does 2x what he's required, or 0.5x. Then the rest of the team try not to feed as they get carried. I don't think its the case that Shiphtur "can't carry" rather it's just impossible for him to do so in the situation he's given.

Posted by LesVerite

Regarding Coast, I think the losing graph looks like that because Coast needs both Zion and Shiphtur to do well in order to win because their bot lane and jungler are so sub par. And because Shiphtur is the more consistent player, he will most often look like the best performing player in those losses.

Zion's inconsistency is in part because when he hard carries, he tends to play champions that are heavily reliant on snowball (excepting the Lee Sin game). As Snowball conditions are much harder to achieve, his performance looks much more inconsistent. Especially since he doesn't have a good jungler on his team to help him snowball. In short, he's inconsistent because he plays champions that are inconsistent by nature and/or plays champions that are reliant on team mates. This is the tanky top meta, and the tank needs someone to do damage. Now, obviously top laners should be playing the champions that are strong within the meta. But with how sub par Coast's jungler and bot lane are playing, that leaves him up a creek without a paddle.

Posted by GelatiSTA

I love Spellsy and I love Thorin. This is great shit :)

Posted by Snapp

Could these graphs reflect the teams records? Interesting to see how investing too much gold into one person or not enough are the three teams with the worst records.

Posted by NQbass7

I find these graphs to be very unintuitive to read. For example, TSM looks like they give equal gold to all the players. But if you look at the key, the change in color from dark to light signifies average gold *for that position*, not the same gold all the way around. So TSM actually distributes their gold according to the league average, not equally (as it probably first appears to most, for example in the reddit comments where this article was linked).

I think it would be much clearer if the radial axis wasn't "gold in comparison to the average for that role," but rather actual % of team gold (but all axes equally scaled), with a separate outline of the average for that role layered underneath for comparison.

Posted by Spellsy

we did this and the issue was that it was significantly harder to read and compare across teams. the idea here is the shape, when you look at 2 different team's raw %s it is super hard to see a difference and thus you couldnt contrast between teams (which is one of the goals of this representation):

here is an extreme example, a side-by-side of TSM and XDG. http://puu.sh/7JEGI.png .. this is why we didnt choose this approach, because while 1% doesn't look like a lot in this approach, 1% is a significant variance.

Posted by Otterus

I totally agree with you. Plus I think the explanation on gold distribution made in this article reinforce the unintuitive aspect of the graphs. I.E. the part when Thorin says that Aphromoo get more gold than Doublelift.

Still a good article nonetheless, I found those graphs really interesting.

Posted by NQbass7

Hm. Yeah, that is harder to read... is it still hard if you put, say, 13% (or whatever the min across all roles is) as the center of the graph, instead of 0%?

Not sure if I have a better solution, just seems like a lot of people are misinterpreting the data.

Posted by fdsaf

You guys really need to put some time into more rigorous statistical interpretation if you're going to put this up and have it taken seriously. With little to go on from the description, I was able to think of a number of problems and biases with the graphs in the article.

1. You absolutely need to take into consideration the overall performance of each team. Teams which don't do as well have less gold to go around (that's one problem with this analysis). Thus, more inconsistent sources of gold (kills and higher turret gold due to proximity) are OVER emphasized whereas consistent gold from farming is minimized. This is a big problem! A team like TSM who wins on a regular basis will have some variance in the amount of gold they receive from all sources; that variance, which is relatively constant in a win* (more on this in a second), will be less than teams who are not performing well (for reasons already stated).

2. Gold spread (for lack of a better term) among roles varies hugely with average length of game. Again, this confounding factor is serious and really needs to be taken into consideration.

3. How did you calculate the average percentages of a team's gold by role? If you added the total gold for each role for all teams and divided by the total gold for all teams, then your average is driven by teams which had the most gold (and goes back to the problem identified in the first point above). If, on the other hand, you took the average of average percentage of gold by role for all teams, then what you have is some sort of Frankenstein statistic (average of averages) which doesn't have much, if any, practical application.

I think that's enough. I appreciate that no one is trying to publish a dissertation in statistics here, but there are big problems here which need to be addressed.

Posted by fabischuessler

Great analysis. But NA ist just less interesting than OGN. Please compare those two, that would be interesting.

Edited by Mejor

Will they be one from Korea and EU aswell? I Think EU have more simple to each other, with Mid as Star in all of the Teams. with Sec in diffrent. SK: mid/Jungle Rocat: mid/Support Fnatic: ADC/top allience: cant tell after Froggen CW: ADC/jungle gambit: MID/support Mil: ADC/TOP Supa: ADC/Sup

Posted by imchantastic

SOOOOOOOOOO this analysis on Shiphtur is total bullshit.

See, when Zion has a bad game, he only has to carry 3 players because Shiphtur can hold his own.
HOWEVER, when Shiphtur gets most of the gold, he has to carry 4 PLAYERS. THAT'S ONE MORE THAN ZION. BECAUSE ZION IS THE FOURTH.

look at Nintendude's gold. He is below negative.

Shiphtur hasn't had a bad game, Zion has. So who is the true pro player???

Posted by stormshadow

@imchantastic: you're comletely wrong though. Shiphtur during Coasts wins performs below average. Zion Performs WELL above average. During their losses Zion is STILL performing above the average.

The big issue in their losses is their bot lane. Shiphtur isn't some god of the mid lane. He performs below the average more often than zion does.

Posted by imchantastic

@stormshadow that was the most bullshit comment ever.
are you kidding? show me a game where shiphtur performs below average.

zion is incredibly inconsistent. he has a couple of good games as jax and people think he's this amazing top laner.

so please, shut. the. fuck. up.

Posted by imchantastic

@stormshadow look at that. shiphtur goes off and zion does below average. and they win.

what now??