SK Gaming's remarkable season, by the numbers
The best League of Legends players in the world will converge in Paris next week for Riot Games’ annual all-star tournament.
And as with any all-star event, fans are eager to see what happens when players who dominated on opposite sides of the table finally clash. Soren "Bjergsen" Bjerg, once one of the best mid laners in Europe and now North American MVP, will go toe-to-toe for the first time in months with Henrik "Froggen" Hansen, the current best at the position in Europe. And will any team be able to take down Korean side SK Telecom T1, who won the world championships last year and easily marched to the top of the Korean tables again this year?
But perhaps more interesting than who has been invited to the event is who hasn't.
Not a single player from SK Gaming, the best team in Europe during the regular season, will travel to Paris. SK's highest vote-getter, marksman Adrian “CandyPanda” Wübbelmann, only ranked 13th in Europe. Serious fans of the League Championship Series (LCS) will tell you why: SK Gaming simply don't have any stars, players who could make their mark on any given game through raw talent.
But then how did SK march so convincingly to the top of the table? The anecdotal answer is "teamwork." But that means very little without some quantifiable evidence to back it up.
So we took a deep dive into the stats to see what narrative the numbers might tell about SK's remarkable season. Here's what we found.
At the start of the Spring Split, no one expected SK Gaming to challenge Europe’s best.
Just before the season began, the team sacked their long-time captain and fan favorite Carlos “Ocelote” Rodriguez. The move was a long time coming. Rodriguez’s form was no longer what it had been when he joined the team three years ago. Still, no one expected the re-tooled roster, built with a mix of LCS veterans like Dennis “Svenskeren” Johnsen and newcomers like Jesse “Jesiz” Lee, to be much more than a middling team. They lacked the star power of squads like Gambit Gaming and defending champs Fnatic, with no one to carry in the late game.
When team captain Patrick “Nyph” Funke ended up leaving for Alliance—the super team hand-picked by superstar mid laner Hansen—it seemed SK's season was certain to be doomed.
SK was left without the bottom lane duo considered to be Europe’s best.
After four weeks of regular season play, all the predictions looked accurate. SK were sixth place out of eight, with a 4-6 record. Then something changed. Over the remaining seven weeks, SK dominated, posting a 14-4 record on their way to the top spot in the standings.
What propelled this transformation is key to understanding how a League of Legends team can succeed when all its players are, by appearance, playing average games.
To begin, we looked at generic team stats, such as Kill/Death/Assist (KDA) ratios, Gold Per Minute (GPM) numbers, Creep Score Per Minute (CSPM), and objective related stats.
Those basic numbers confirmed what we already knew—SK was good at securing objectives and in seeing through the late game. But that was about it. Not satisfied, we dug a little deeper, looking at granular statistics over the course of the season.
We recorded gold and creep score for each of SK’s lanes and lane opponents at the 5, 10, 15, and 20 minute marks. We examined the jungler, Johnsen, at a deeper level, analyzing where he spent his time, which lanes he pressured and when, and his success rates in those pressures. We also took a look at SK’s champion selections to find any significant trends.
Graphic by LoLStats.gg
What immediately jumped out was this: Over the course of the season, when SK Gaming had a lead, no matter how small, they won. And they usually gained that lead by securing objectives, even sacrificing kills to do so.
SK's snowball rate—the percentage of games win after securing certain objectives—were off the charts. They won 93.75 percent of games where they took the first dragon, the best in the league. In fact, from week five onward, SK won every single game in which they secured the first dragon. That’s compared to a league rate of 65.18 percent.
SK secured 16 first dragons, tied for the most in the EU LCS, and won 15 of those 16 games. They killed Baron Nashor first a whopping 17 times, two more than the second ranked team, Roccat, and won 16 of those matches.
Each dragon is worth at least 650 gold and Baron clocks in at a whopping 1500 gold. And that’s before mentioning the experience gain for each or the all-important Exalted with Baron Nashor buff, which provides a huge, four minute combat advantage. SK leveraged these objectives better than any team in the league.
When SK Gaming got a lead, they closed out the match, and did it efficiently, with a 38:25 average game time that made them second fastest in the league after Roccat.
Click for a larger image | Graphic by LoLStats.gg
It makes sense, then, that SK's early season stumbles all centered around getting that lead in the first place.
Overall, the team had difficulty in the laning phase during the first four weeks. Only SK's top laner, Simon “Fredy122” Payne, managed to outfarm opponents. In the jungle, Johnsen’s ganks—pressuring a lane in an attempt to secure a kill—were ineffectual. He attempted 3.4 ganks per game with only a 21 percent success rate.
Then SK made a couple of adjustments that turned their season around.
One key change was shifting Johnsen’s focus in the jungle, using him more on the bottom. Johnsen ganked less often after week four, 2.94 times per game compared to 3.4—but with a much higher success rate, 47 percent. He sacrificed gank attempts in the top lane to improve his efficiency on the rest of the map. The number of ganks in the middle and bottom lane remained the same, and were more successful. Johnsen favored champions with strong counter-gank ability, like Pantheon, Elise, and Evelyn. He was particularly effective with Pantheon, as SK went 6-0 in matches with him playing the spear-toting Spartan.
Graphic by LoLStats.gg
This paid dividends for bottom lane duo Adrian “CandyPanda” Wübbelmann and Christoph “nRated” Seitz. The pair saw a 5.56 percent increase in gold at five minutes compared to their level through four weeks, with a 12.35 percent drop in gold at five minutes for their lane opponents. And they carried that advantage into the late game.
A stronger bottom lane duo and more bottom lane jungle presence paid off in another way—dragon control. SK claimed 57.14 percent of dragons in weeks five through 11, a 33.32 percent increase over their week four mark.That 57.4 percent mark would have lead the league over a full season..
The shift in jungle focus wouldn’t have worked without Payne’s top lane prowess. SK left him on the top lane island, and he thrived. While his opponents saw a 12.71 percent increase in gold earned at the five minute mark, probably due to less pressure from SK’s jungler, Payne was up 24.7 percent while left to his own devices. SK went 7-4 in matches with him on Renekton and 5-0 in games with him on Trundle.
Graphic by LoLStats.gg
SK also figured out how to snowball towers in the second half of the season. Through week four, they only won 20 percent of games while taking the first tower, but their mark the rest of the way, 77.78 percent, obliterated the league average of 55.36 percent and would have topped the league for a full season.
Part of that is due to their champion selections. SK struggled with mid laner Lee on assassin champions, going 1-4, but loved the siege ability of Nidalee and Ziggs, combing for a 9-2 mark. By placing Lee, a newcomer to the LCS who probably doesn’t have the experience to compete with the star mid laners in Europe, into a siege role, SK was simply using their talent in the most efficient way.
While the jungle shift was an important part of SK’s success, it was also representative of the team's real strength: the ability to adapt, both in macro strategy and in-game tactics.
SK stayed ahead of metagame-changing patches throughout the season, maintaining their team style and success while adjusting their strategies. They did the same in the server, out-rotating opponents to turn a small lead into a victory, a lane advantage into an objective, an objective into a victory.
While our methodology revealed a portion of what made SK so dominant, it’s hardly the whole tale. A more in-depth analysis of the pick and ban phase may reveal more about how they held an advantage in so many matches. Better ways to analyze the late game, and especially success in team fights, be it team composition, positioning, tactics, or mechanics, may reveal more secrets yet
Statistics may never explain the beauty behind one of Seitz’s well-timed engages, or the artful way Wübbelmann cleans up during a hectic battle, while Payne and Johnsen keep the enemy team’s back line engaged. But they do tell an important part of that story.