Preseason 5: More data from the jungle

Patch 4.20

Last week I looked at some super-early data from the new jungle in patch 4.20. The patch had only been out a few days, so there was not much to work with, but it was still interesting to see how things were shaping up.

Today I'm going to look at the same analysis with more data (5x as many matches) which will give a clearer picture of where we're at. I looked at every challenger and master elo (mmr) solo queue game that has been played in NA on patch 4.20 so far (thousands of matches with tens of thousands of players). The goal is to see which jungle champions and items are popular, and winning.

This table gives a nice summary of the state of the jungle on the new Summoner's Rift in patch 4.20 at a relatively high level of competition.

ily

ChampionWin %Items
Fiddlesticks
60%
Elise
56%
Pantheon
55%
Maokai
55%
Nunu
54%
Warwick
53%
Nidalee
53%
Olaf
53%
Rammus
53%
Vi
53%
Rengar
52%
Jarvan IV
51%
Amumu
51%
Shyvana
50%
Xin Zhao
50%
Shaco
50%
Riven
49%
Lee Sin
48%
Kha'Zix
48%
Wukong
48%
Fizz
47%
Hecarim
47%
Skarner
46%
Nocturne
44%
Master Yi
42%
Udyr
40%
Sion
36%
Diana
35%
Zac
34%

With more data, the picture is a bit more clear than last week. Fiddlesticks is at the top, and is unsurprisingly shading toward the magus enchantment for his jungle item. The magus enchantment is rediculously gold efficient (especially compared to the season four Spectral Wraith item) so it is going to be popular on ap junglers. Also, Fizz has dropped off since last week. It was interesting to look closer at the data; I found that a single player was doing extremely well with him last week. There was so little data that that individual's play moved fizz to the top of the charts. This week, with more data, these numbers are more reliable.

I think it is notable that Warwick here has a 53% win rate. Other sites are reporting higher numbers for him, but they aren't restricting their analysis to high-level players. I do think that Warwick is better against worse players, since better players are more likely to mitigate him by kiting properly. It would be super-interesting to do this kind of analysis for each league (bronze, silver, gold, platinum, etc...). Unfortunately, the organization of the public data api makes that type of analysis difficult.

The Ranger's Trailblazer definitely seems to be the most popular item. The AOE smite on a heavily reduced cooldown drastically reduces clear times, and combined with the new longer respawn times for jungle camps, adds up to more free time for ganks. This benefits strong ganking junglers like Fiddlesticks, Elise, Pantheon, Nunu, Warwick, Rammus, Vi, Rengar, etc... All of them rank high on this list.

The nerfs to Amumu seem to have been 'just-right', if these numbers are any indication. With a winrate of 51% he's right in the middle of the pack.

Zac's win rate here is still abysmal. Other sites report his win-rate in general in the high forty percents. His poor performance here is likely a combination of him not being a strong jungler, and high-level players knowing how to mitigate him. Your mileage may vary.


BONUS: More stats!

Since it was easy, I also captured a few simple stats from this dataset of high-level solo queue matches on patch 4.20.

Enjoy.

ily

Blue side win rate: 54.7%

This number isn't too surprising, and actually matches reasonably well with data I've seen from the season four world championships. I still think it would be neat to put the dragon and baron in the actual bottom-right and top-left of the map. This would make the map symmetrical and would make positioning for these objectives more dicey, which I think could add strategic interest. The real answer to this problem is to change the client so it always appears (from your perspective) that your team is in the bottom left moving to top-right. It is my personal opinion that this number stems from asymmetries in the view-point perspective of the game. But we'll likely never know.

First blood win rate: 59.9%

So, this number is simply calculated as matches where the team that got first blood won divided by total match count. This number seems high to me. Again, this dataset comes from top-level solo queue players, so they're more likely to know how to take best advantage of first blood. However, other sites are reporting similar numbers, so this may be legit. Don't feed first blood. Ha!


Ok, back to playing now. Enough #LeagueMath for one day. Tell your friends, and keep theory crafting that 5-smite comp, I know I am.

Keep it real, and don't touch that dial; I'll be back real soon with more league'n and math'n here at LeagueMath.com.


No Poacher's Knife was harmed in the making of this article. This dataset was still a bit small. Get to playing so I have plenty to work with.

Peace.