The Bang roster dilemma and the end of SKT’s second age

The first February of the SK Telecom organisation’s time in League of Legends was an auspicious one — it marked the completion of the formation of SK Telecom T1 2, later to become SKT T1 K. Since then, its Februaries have been almost markedly immemorable. In 2014, the team were basking in the glory of a 15–0 Champions Winter season; the rest of the year saw two quarter-final exits from Champions, and ultimately a failure to reach Worlds 2014.

2015 saw the team drop a regular season serieses that split to the GE Tigers, something that would turn out to be the difference between the 1st and 2nd seed. SKT still trounced GE 3–0. They also announced the addition of Im “T0M” Jae-hyeon to the active roster; T0M played those finals, but would only play six further games for SKT, and as it happened, never played a professional game after 2015.

For 2016, the team found themselves in a slump domestically that saw them dropping consecutive serieses to the firmly mid-table Longzhu and Afreeca lineups, and falling to 6th by the time they left for Katowice at the end of the month. Naturally, they went undefeated once again at the IEM World Championships in March. In minor roster move news, Lee “Scout” Ye-chan left the active roster, and left the organisation the next month; amusingly enough, he immediately joined EDG, and has become arguably the most successful SKT trainee ‘export’ of the post-sister team era.

2017 saw an absolutely dominant SKT take some rare, albeit barely perceptible, black eyes; they were 2–0ed by Afreeca, and then taken to three games by a collapsing Longzhu, before once again riding roughshod over everyone into playoffs and at MSI. In another amusing coincidence, the team’s most recent trainee, Lee “Effort” Sang-ho, neither joined nor left the active roster; he is now, of course, sharing time with Lee “Wolf” Jae-wan at support.

Then, we get to 2018, and for the fourth consecutive year, there was some movement on the roster. This time, two new players were brought into the active roster. One was Choi “Pirean” Jun-sik, a three-split NA LCS mid and the latest in a long line of training partners for Lee “Faker” Sang-hyeok. The other was an ADC: Han “Leo” Gyeo-re.

In sports terms, two and a half years is nothing. In esports terms, it seems eternal, and at the least, is impressive. To that end, when the SKT machine finally started spluttering badly towards the end of 2017, many were willing to write it off as a hiccup for far longer than a strict analysis of the struggles and shortcomings of the lineup would otherwise dictate.

This is entirely understandable. The second era of SKT, despite what we see now, was an incredible feat. Sure, the first era had seen higher highs. The Impact-Bengi-Faker-Piglet-PoohManDu lineup had emerged from nowhere to get 3rd in Spring, 1st in Summer, crushed Worlds, and crowned it all with the perfect season in Winter. Asides from maybe Worlds 2015, the second-era team was never as far ahead of its competition as the first-era team was in the latter half of 2013.

The sustained nature, and the longevity, of the success of second-era SKT, however? Consecutive Worlds titles, consecutive MSI titles (and a 2–3 loss before that), four out of five LCK titles (including a 12–1 game record in finals). The team had its stutters here and there, but there was never any sense that they were truly shaken by them at any point.

SKT is now not working. Faker is still there, Bae “Bang” Jun-sik is still there, Kim “kkOma” Jeong-gyun is still there, the supporting cast from 2017 is largely still there. On a lot of teams, you would just write this off as the meta shifting against how they play.

Yet, there’s something wrong with that. The quiet beauty of the second era of SKT was that while they would have weaker patches, there was never a meta where they weren’t able to compete. The quieter beauty, still, was that this was the case not because they were flexible to everything, but because they instead were a team that understood what they were, and played to enable their essential tendencies.

The first era of SKT had been able to smash all three lanes, and abuse the fact that, simply put, Bae “Bengi” Seong-woong and Jung “Impact” Eon-yeong were smarter than their opponents to accelerate that into a win. The meta itself was shifting away from that being enough on its own (it’s not entirely a coincidence that SKT T1 K’s fall coincided closely with the universal adoption of teleport on top laners in early 2014), but in any case, that was not a luxury that 2015 SKT were going to be able to have. Chae “Piglet” Gwang-jin had burned his bridges and left for North America, Kim “Deft” Hyuk-kyu and Gu “Imp” Seung-bin had gone to China, and both contemporaneously and in retrospect, the ADC talent pool in Korea was as poor as it had ever been.

Bang was coming off a decent, if unspectacular, 2014. He was a serviceable LCK-level ADC. He was never going to crush lane, but he was fine. Hence, we come to this very odd truth: from the very beginning, Bang’s role was written into SKT, simply because there was no-one else at the time. Ancient players like Seon “Space” Ho-san and Kang “Cpt Jack” Hyung-woo were still starting for top teams, and while PraY would go on to have a breakout split, nobody thought anything of him (or the rest of the GE Tigers) going into 2015.

As it happened, the first roster perfectly covered up for Bang. One of the truisms of League as a game is that if you have pressure in three lanes, you’ve already won; if you have pressure in two lanes, you’re in control. Faker was, as he would go on to be, still Faker; as derided as the Lee “Easyhoon” Ji-hoon era was, it would never have happened if not for the fact that Easyhoon could create even more pressure on certain champions.

The charge has been frequently levied against SKT that they would not be a fraction of the team that they are without Faker. While true, and particularly true in those cases where the system breaks down (the ability of Faker to single-handedly annul all early-game pressure across the map was one of the remarkable, and not-spoken-about, components of SKT’s playoff dead frog bounce last summer), it’s true more in the sense of his archetype as a player. Crown could not succeed where Faker has, but it’s possible that post-rookie Bdd could.

Even with that being the case, however, the second era of SKT crystallised in such a way as to make full use of Faker’s talents. Mid has pressure. Bot does not have pressure; where Bang has been seen as a relatively successful laner (such as in spring 2016), it’s been predicated on high-CC kill lanes rather than priority.

It needs hence be the case, to follow the truism, that top generally has to have pressure. SKT’s first top laner of the post-sister teams era? Jang “MaRin” Gyeong-hwan, at a time when he was either the best or the second-best (behind Lee “Duke” Ho-seong) top in the world by almost all estimations.

In their 2015 roster, SKT achieved what was — and arguably, still is — the modern manifestation of the “smash three lanes roster”. SKT can always play for two lanes. It may be harder to play for mid-bot than top-mid, but their players are always equipped on an intrinsic level to do it. It’s something that sounds very simple. It is very simple. Yet it’s been a huge part, on a very fundamental level, of the team’s success, and has enabled the team’s core to ride on for as long as it has.

The challenge since 2015 has been sustaining the situation in top lane. MaRin was a star talent, and perhaps the only SKT top laner of the era that was actually recognised for his positive impact while on the team. Duke was an equal to MaRin upon joining, but fell off a little over the course of summer. With Heo “Huni” Seung-hoon, and especially Park “Untara” Ui-jin and Park “Thal” Kwon-hyuk, the pressure stopped being a given, and started being something that shaped the early game towards its happening, rather than being something that could simply be played off.

And now, we’re here today. SKT finished Spring 2018 with a precisely .500 record — 23 game wins and losses, 9 match wins and losses. They scraped into playoffs, won a Bo3 against KSV, and lost summarily to KT in the next round. Nobody’s betting on SKT to bounce back anymore. So what now?

There are, in truth, two ways you can view the failure of SKT in Summer 2017 and Spring 2018, with regards to that roster structure that was working beforehand.

The first relates to Bang’s passivity himself. One of the more intriguing failure conditions for an athlete in any sport, any game, is when they take what’s always worked for them, and modify their behaviour and actions on that basis to the point where it hurts their game. To use a simple example from League, let’s say we’re teaching a generic player to play mid. We tell that player that their first objective as mid is to farm, and that he should prioritise farm above all else.

The player does that. Presumably, he wins more games as a result, so he’s able to draw a psychological connection at both a conscious and an unconscious level — if I farm out, I will win. We have given the player a principle by which he can improve his own behaviour in the game. Of course, there are exceptions to this logic, and there will come a point where as the player modifies his own behaviour, he’s modifying it in a way that will make him win less, not more. To use a vulgar but easily understandable example, if he’s absolutely only thinking of farm, he goes to pick up bot farm when Baron is up, keeps pushing from his own tier-2 tower, and as a result, his team are forced to attempt to defend Baron 4v5 with no teleport.

This is a deliberately extreme example, of course. To use a basketball example, what we’re talking about there might be a rec league player going from 15 shots a game to 50; a NBA player who’s caught themselves in a similar trap could go from 15 to 18 or 20. Yet, when we’re talking about the level of competition being as finely balanced as it is in elite competition, it could be just as crippling.

The eye test and statistical indicators both support the idea that Bang has become markedly more passive in his play over the past year in particular. For the latter, in 2018 Spring, across 46 regular-season games, Bang died 35 times (one of only two ADCs in LCK with sub-1 death per game), and had a 9.6 KDA — with a 50% win rate. As favourable as the Spring meta was to ADCs to some degree at both ends (and particularly into playoffs), with players like Martin “Rekkles” Larsson and Yiliang “Doublelift” Peng putting up historic performances and numbers in the West, it’s impossible not to see that as aberrant.

In the Bang era, SKT has rarely been an early-game powerhouse. But even by those standards, they’ve looked sluggish and more reactive than ever throughout this entire split, and across all of their games, we can almost see an echoing effect emerging. Bang plays passively, and communicates that he’s doing as such. Untara, Thal, Blank, and Park “Blossom” Beom-chan are — directly or otherwise — pressured to become more active, both by the sheer reality of how the early game is going to play out if they don’t set the bot lane up to have the privilege of being passive in the first place. In turn, given more support, Bang either feels able to, or feels forced to, play even more passively, which leads to top and jungle feeling the need to be even more aggressive, and so on.

Of course, ‘aggression’ is arguably the better case here. It can also just mean ‘collapse’. Take the KT series in playoffs; while game 4 was marked by the SKT top-jungle-mid trio’s failures in multiple skirmishes, the previous games had owed much more to Go “Score” Dong-bin and KT using the knowledge of this inherent nature of SKT (assisted, if not purely enabled, by glaringly one-note drafts on SKT’s end). If you want an encapsulation of SKT’s season, watch the first 10 minutes of game 3 — with Jarvan IV and Tristana on SKT’s side, the plan is so obvious as to be transparent, and Score spends the entire early game anticipating or tracking every single thing that Blank could possibly do. By giving themselves one option, SKT end up giving themselves no options.

This should all raise two questions. The first is, to my mind, the more fascinating one, and also one that’s basically unanswerable: why now? Why has Bang’s game only lapsed — indeed, been allowed to lapse — now, after so many years within the organisation? The only contribution I could make here is to note that, in every elite sport but particularly in esports, a large part of the jobs of the coaching staff is an ongoing process of regularisation. Becoming great, and maintaining greatness, is a constant push-and-pull; you’re working to create and regulate habits, to identify problems and apply or devise methods to deal with them as they come.

With that in mind, the departure of Choi “cCarter” Byoung-hoon. While there was always relatively little written about or by him publically, particularly in the English language media, it does seem that cCarter played the role of the pragmatist and the generalist within SKT, as a multi-game coach with significant tenure and expertise within the space but not, and never, the game. He’s an archetypal Korean-style head coach in that regard.

(I’ll attempt an one-paragraph explanation of the Korean coaching system, for those unfamiliar. The Korean head coach is closer to a manager on most Western teams. Asides from basic logistical duties, he’s generally tasked with interpersonal management, and contributing to the overarching structure of the team outside of the game itself. The Korean strategic coach is what we usually call a head coach in the West — he’s hired for his expertise, and he’s mostly focused on strategic concerns and overarching structures within the game.)

SKT now have kkOma as head coach, and ex-SKT players Lee “PoohManDu” Jeong-hyeon and Bengi filling out the hierarchy. As old (32) and experienced as kkOma is, and as much as he was able to observe cCarter over the previous five years, his new role demands a much different skillset to his old one, and the struggles to keep Bang — and, arguably, Blank — in an appropriate equilibrium does make one wonder.

The second question is more straightforward, and leads into the second way to think about the roster structure: “it’s worked for two and a half years, why isn’t it working now?”

While the pool of talent able to operate at an appropriate level for a given team is not quite as small as both fans and pundits tend to think, clearly, it is limited. Given an infinite amount of time, there will only be so many players who one could possibly see developing to the appropriate level for an elite-level team. Given an actually reasonable amount of time, and a reasonable amount of resources, it’s rare that you won’t be able to count those players on two hands, or even one hand.

The problem SKT have is not only that the pool is small anyway, but that the restrictions that the Faker-Bang system puts on them has in places seemed to dry that up completely. The easiest place to see that is in terms of top lane. SKT are on their fourth and fifth top laner in the Faker-Bang era — MaRin, Duke, Huni, Untara, Thal. After MaRin worked, they have consistently hired tops who have not only been near the top of their class as talents (yes, even Untara), but have fit the mould stylistically — their nature style, and natural tendencies, play into how a Faker-Bang SKT will need to use them.

At this point, however, therein lies the problem. We’ll be generous — maybe there are 20 Korean top laners on the planet that are good enough. You can probably strike down half of those straight away as not being able to consistently play that carry style to a high enough level. You’ve already tried another five. Maybe two more are in China, two more are on LCK teams and don’t want to leave. Your pool of 20 has gone down to 1 in an instant. The turnover of talent in jungle and support hasn’t been as significant, but you’re still talking about significant strictures in what playstyle you’re willing to accept into the system if you’re committed to maintaing both primary carries, and everything about how the team will tend to play that it entails.

In that lies the problem for the rebuilding SKT. At first, the structure all fit together with the players they had — at least, well enough to win. Through 2016 and 2017, they had to start changing parts; each time, maybe the fit’s a little worse, maybe you have to take a hammer to it here and there, but it kept chugging along. It’s 2018 now. Maybe all the parts are out there. Maybe they can get them all together before something else breaks. You absolutely would not bet on it being the case.

Who is Leo? The SKT announcement simply describes him as an “amateur”, and in a statement to an Inven reporter, kkOma only said: “As for Leo, he’s young and mechanically talented. However, he’s inexperienced — as he has never played as a professional before”. All we really have to go off is his op.gg accounts (main, alt), which show up something a little notable — for Leo, in part, but moreso for SKT.

Leo could be the next big thing, or he could be gone by the time I’ve finished writing this. It’s impossible to tell from the outside. He’s your classic SKT trainee in many ways — after intermittent play through to mid-2017, he made a run to 900LP, finished season 7 at rank 30 779LP (which appears to have made him the #1 unsigned ADC on the ladder), and peaked at 1149LP in March. Impressive, but not remarkable in the pantheon of SKT rookies.

This is what’s interesting about Leo:

Notice those death numbers — lots of 4s, lots of 5s, sometimes even the odd 6 — and KDAs. That’s incredibly unusual for an ADC at the top of the Korean ladder: look at Bang, Deft, and Uzi:

and at a couple of other top ladder ADCs:

Why does this matter? It suggests that, at the very least, Leo hasn’t been brought in to play the role of mini-Bang, unlike almost every other SKT signing (including Pirean, incidentally; when you start the description of your backup to Faker with the point that “[he’s] a very aggressive player”, and end it with “so I think he’ll be able to quickly adapt to our team”, it’s clear what the idea is there). Yes, Leo is a young player; yes, he presumably can be moulded. But SKT haven’t ever been in the business of completely melting down a player and starting him over. Their policy throughout the lifespan of League of Legends has been to play to strengths; if they’re picking up Leo, and he isn’t going to leave in May or similar, it hints that something might be changing.

Leo could still just leave tomorrow. But it does seem that SKT are making their first halting steps into realising that, whatever the new configuration of their roster might be, there has to be a new configuration. The wells have run dry on 2015–2017 SKT, and now there are two points of interest going forward: what that new formula is going to be, and whether they have the time and the means to make it work going into 2019.