Finding a Niche: Evelynn's Subpar Summer Split
Introduction and Summary
Because of her invisibility, she can be one of the most terrifying spectres in all of League of Legends. Did she just walk over my ward? Is she following me through my jungle? Is she standing right behind me!?! Every moment becomes a nightmare (as CLG's Pobelter will tell you).
Evelynn was rarely seen in pro play in the 2015 Spring Split, but after some small changes in patch 5.6 (earlier stealth entry at later levels, and spell hits lower Dark Frenzy's cooldown instead of giving her move speed), then a 10-damage buff to her Hate Spikes in patch 5.7, she made an appearance in the hands of EDG’s jungler, Clearlove, in Game 5 of the MSI Finals. That led to an international trophy, and Evelynn has seen a surge in popularity since.
But popularity isn't the same as success. This split, Evelynn's bark has been worse than her Hate Spikes. Evelynn only has a 38% win rate in the Summer Split, thanks partly to poor use in some players' hands--we can't all be Clearlove--and partly to good countering by her opponents.
This article explores Evelynn's performances during the Summer Split so far. We'll look at how Evelynn is used, break down her results, and run some in-depth statistical analysis on her early-game win conditions, comparing her to some other common jungle picks to see if and how she is unique.
We'll see that Evelynn's biggest keys to success are ganking more, farming less, and valuing the first Tower above the first Dragon. This is not the case for every jungler: Sejuani, for example, benefits more from hard farming early, while Rek'Sai gets more benefit than others from taking the first Dragon. (More about that on page 3!)
The statistics in this article shouldn't change your fundamental idea of who Evelynn is and what she needs to do to win, but they might demonstrate just how binary Evelynn's play style is: either she ganks a lot, or she loses.
Read on if you're ready to tumble down the rabbit hole with me!
All statistics in this article are based on 2015 Summer Split regular season games from the LCK, LMS, NA LCS, and EU LCS, up until June 7, 2015.*
*The LPL does not make match statistics available, and cannot be included.
How is Evelynn Being Used?
Evelynn is picked as a jungler for her global early pressure. Since Eve is invisible, the enemy team always has to respect the idea that she could be standing nearby, ready to follow up any Crowd Control (CC) her laners land. As Froskurinn once put it, Evelynn "can apply pressure [on the entire map] without doing anything". That doesn't mean it's effective to just sit back and farm--we'll get into that later--but opponents have to play differently just because Eve is in the game.
This split, Evelynn is usually built with a Warrior's enchantment to go with either a Stalker's Blade or a Skirmisher's Sabre, which helps out her early gank power with extra burst damage. Afterwards, tank items come in to help her remain alive long enough in team fights to be relevant.
Evelynn is best paired with team-mates who bring their own CC. While Evelynn has an Area-of-Effect (AoE) slow on her ultimate, her kit lacks any other way to lock down her targets. This lack of CC, especially in the crucial early minutes before she reaches level 6, means that Evelynn is only an effective ganker if she has laners who are easy to gank for.
This split, the most common team-mates for Evelynn have been Alistar, Maokai, Sivir, and Azir. Evelynn's three most common team-mates in each role are listed below.
These are mostly champions who excel at receiving ganks, and Sivir has additional synergy when she gains access to On the Hunt at level 6 and is able to help Evelynn stick to targets at the start of a team fight. As an interesting side note, Hecarim is one of the less effective gank receivers on that list, and has also never won a game alongside Evelynn this split (0-7).
For an example of an excellent Evelynn-based team composition, look to the game that restarted the whole Evelynn trend, when Clearlove picked up Eve against SK Telecom T1 in Game 5 of the MSI Finals. Maokai, Morgana, and Alistar all provide a ton of CC, and we've already pointed out Sivir's synergy.
Based on Evelynn's skill kit and common team-mates, teams who pick her seem to have a common plan: gank lanes; secure kills; hit the mid game running. How well have Evelynn players been able to execute on that plan?
In short: very poorly. Eve's overall win rate in the five major leagues so far this split has been only 38%.
Europe has shown a little bit more success with the unseen jungle assassin, or perhaps less success with playing against her. One reason for that may be European teams' lower ward output vs. Evelynn: Evelynn's EU opponents only placed an average of 2.55 wards per minute (WPM), compared to the LCK's 3.18 WPM, the LMS's 3.13, and the NA LCS's only slightly higher 2.67.
Despite the poor record, Evelynn obviously is capable of winning now and then. To understand what an effective Evelynn looks like, and also an ineffective one, let's explore Evelynn's individual performances. First we'll review some end-game statistics to paint the big picture of how she has been contributing, then we'll dig into Evelynn's early game and late game contributions to see how she attempts to earn wins, and what it looks like when she fails.
Most of Evelynn's end-game statistics aren't too far off of the overall Jungler averages. Her deaths are a little high, her assists a little low; her kill participation is on the low side; her damage share is above average, her gold share below average; her ward clears are subpar.
The table below breaks out several stats by wins and losses, with overall averages for all Junglers for the sake of comparison.
The main stats that increase in Evelynn's wins are her assists, her kill participation, and her wards placed per minute. Interestingly, too, Evelynn's damage share tends to be a little bit lower in victories, probably because she helps her teammates get farther ahead so that they can carry harder.
(Note: We'll take a closer look on the next page at how different aspects of Evelynn's performances actually predict win rates, including tests of statistical significance, but we'll allow this section to stand on its own for now.)
Stats from the end of the game are certainly interesting, and relatively easy to talk about, but Evelynn is specifically picked for her early game. Let's take a look at how much Evelynn tends to contribute in the first 10, 15, and 20 minutes.
On average, Evelynn players are behind in CS and gold during the early stages of games. That's somewhat intentional: they spend more time looking for ganks and trying to get their carries ahead, which means less farming and smaller personal rewards.
Ideally, any deficits Evelynn is suffering should lead to bigger rewards for her laners. That comes via successful ganks, which we should see in Evelynn's early game kills and assists.
Evelynn players do tend to accelerate their kills and assists after the 10 minute mark, but despite what one might expect, Evelynn's average kill+assist numbers in the early game aren't better than other common jungle champions this split.
For whatever reason, Evelynn players haven't consistently shown that they can translate her theoretical ganking prowess into higher than average early-game kill and assist numbers.
That doesn't mean Evelynn is never successful in the early game, though. When Evelynn does achieve her early goals by securing successful ganks, she can still be very effective. On the next page we'll take a look at some of the biggest factors that lead to Evelynn wins.
It's time for the really fun stuff: what are the most important things Evelynn needs to do to ensure that she can win the game? We've discussed the theory behind playing Evelynn; now we're going to back up the theory statistically.
P-values are reported for the various statistics below. For readers who are not familiar with social research statistics, a p value tells us how much confidence we can have that the findings are not just random. In social science (which I'd argue LoL is most similar to), a p value of 0.05 is the most common threshold for statistical significance. Occasionally 0.10 is seen as an acceptable p, but this can be questionable.
First Blood Participation
When an Evelynn player gets involved in a First Blood kill, it has a massive impact on her win rate. Evelynn's overall win rate this split is 38%, but when she gets a First Blood kill or assist that jumps to a much more impressive 66.7%.
For an example of a game where Evelynn fails to get a First Blood kill or assist, watch through SK Gaming vs. Origen from Week 2 of the EU LCS. After an early skirmish gets turned around at 4:30, Svenskeren is set behind, and he's completely unable to find a footing afterwards, playing reactively instead of proactively and dying again in a skirmish at 13:20.
Meanwhile, watch EDG set up a Tower dive against SKT, taking First Blood and two kills for one.
Gragas, Rek'Sai, and Sejuani all benefit from First Blood as well, obviously, but the discrepancies aren't as large, to the point that Rek'Sai and Sejuani don't achieve statistical significance.
Gragas has a 35.8 percentage point difference in win rate, which isn't that much less than Evelynn's 40.6 percentage point gap, but Gragas's overall win rate is so much higher that he still wins 42.5% of his games even when he doesn't get a FB kill or assist.
In short, among these four commonly used junglers, Evelynn is the most reliant on First Blood participation to win games.
Kills + Assists and Creep Score at 15 Minutes
This part gets complicated: here's where we demonstrate how much more important kills and assists are to an Evelynn in the first 15 minutes, compared to hard farming.
The numbers below show how correlated kills+assists at 15 minutes and creep score at 15 minutes are with win rate. A large, positive correlation means that more K+As or more CS leads to more wins, while a large negative correlation shows that the champion wins more when their K+As or CS are lower at 15 minutes.
To help the interpretation, a correlation coefficient with an absolute value smaller than 0.3 is generally considered a "weak" relationship, 0.3 to 0.5 is a "moderate" relationship, and greater than 0.5 is a "strong" relationship.
First things first: Evelynn's kills + assists total at 15 minutes is important to her win rate, with a moderately strong 0.317 correlation coefficient. But her CS total at 15 minutes is not correlated with win rate. The stats make a pretty strong argument that ganking is more important for Evelynn than farming. Clearlove's approach in the MSI Finals agrees: at 15 minutes he was down by 16 CS, but had three assists under his belt, and that trend held all the way until 20 minutes, where he was down 15 CS but had snowballed his way to eight assists.
Now take a look at the other champions, and here's what I meant when I said this was complicated: Gragas and Rek'Sai actually have negative correlations between their CS@15 and their win rate! In other words, the lower their CS, the higher their win rate. This does not mean that having less CS is inherently better: instead, we can read between the lines and see that if they have less CS, it will be because they're doing something else on the map, like ganking or Tower diving.
Between the negative CS correlations (or in Eve's case, the lack of correlation) and the positive kills+assists correlations, then, we can say that Evelynn, Gragas, and Rek'Sai win more often when they farm less and (successfully) fight more.
Sejuani is the only oddball here who wins more when her CS@15 is higher. Naturally she benefits from early kills and assists as well, but she should feel free to break out the Ranger's Trailblazer, put her head down, and go for the Gromp (to borrow a phrase from DoA).
(As a sidenote, I also ran numbers on CS Differential at 15 minutes, and found no statistically significant correlations for any of these four junglers. This suggests that it's ineffective for these junglers to focus on out-farming their opponents, whether through counter-jungling or soaking up a share of lane minions.)
First Dragon and First Tower
The early game isn't only about achieving individual goals like kills and farm; it's also about accomplishing team goals by securing objectives like Towers and Dragons. To round out our picture of Evelynn's early-game win conditions, we can check how these objectives are connected to win rate.
The numbers below show that the first Dragon doesn't have a statistically significant effect on win rate. That doesn't mean that it isn't beneficial, only that its benefits aren't large enough to be measurable statistically. (Rek'Sai comes closest with p = .106.)
The first Tower kill, on the other hand, does have a measurable effect: Evelynn's win rate jumps up by 32.9 percentage points with a first Tower kill. This may sometimes be related to First Blood effects, if a successful gank or Tower dive leads to free time in the lane to finish the Tower off.
Evelynn just barely skirts the edge of statistical significance here, with a large increase in win rate based on the first Tower kill. Sejuani benefits the most from first Tower, and Gragas also has a statistically significant increase in his win rate. Rek'Sai benefits too, but not as much as the others.
Summary of Findings
We've shown statistically that Evelynn wins more often when:
- She is involved in a First Blood kill.
- She has more kills and assists at 15 minutes.
- Her team kills the first Tower of the game.
We've also shown that we can't confidently claim that Evelynn wins more when:
- She has higher CS at 15 minutes.
- Her team kills the first Dragon of the game.
The takeaway is simply a confirmation that Evelynn has to gank to win. Ganking produces pre-15-minute kills and assists, her most valuable asset. Ganking also leads to Tower kills, even when it doesn't result in a kill on a champion, because it can bully the opponent out of lane, allowing free time to shove the wave and hit the Tower.
Hard farming doesn't help Evelynn win; controlling Dragon doesn't cut it, either. Even counter-jungling is questionable if it doesn't result in skirmishes that lead to kills. Only frequent ganking does what Evelynn needs.
We also showed that the story is not quite the same when we're talking about other junglers. Sure, they all benefit from First Blood and from having more early kills and assists, but some benefit more than others. Some junglers benefit from hard farming (Sejuani) while others don't. Some junglers benefit more from the first Dragon (Rek'Sai) while others benefit more from the first Tower (everyone else).
Every jungler has a different set of paths to victory. They all have their own niche.
Evelynn can be an effective pick, but too often she is either played too cautiously, or isn't given any opportunities to pull off the ganks she desperately needs. Evelynn's invisibility can be extremely intimidating, but the way to counter her is relatively straightforward: position safely in lane, ward heavily (don't be like the Europeans!), and even feel free to give up the first Dragon if it allows you to keep farming your own lane in safety. There is little reason to fear Evelynn if she has to resort to hard farming her own jungle.
These ideas aren't new ones. In fact, teams have been using them for weeks, leading to Evelynn's low win rate. Despite that, she is still getting picked, and still being either played poorly with too few gank attempts, or being easily countered with conservative laning.
Maybe these statistics will lead some teams back onto the right path and get Evelynn ganking more often. That falls on the laners' shoulders as much as the jungler's: lanes need to be gankable, or Towers need to be diveable and cleared of enemy vision, to facilitate Evelynn's involvement.
Despite her low win rate, Evelynn hasn't faded into invisibility yet this split, though she seems to be getting picked less and less. But perhaps she is just luring her opponents into a false sense of security, awaiting her opportunity to emerge from the shadows and strike again.