Analysis in Competitive League of Legends

Despite the tremendous strides competitive League of Legends has made recently, the industry is still fairly young with regards to how players interact with the support staff and management. While the responsibilities of the coaching role have reached a state of general definition from developing team's social and performance elements, the analyst role still remains relatively obscure with regards to scope and impact. An idea popularized in the past few years is that analysts most commonly watch a lot of competitive games and advise on pick/bans or strategy. While this is an important aspect this idea presents a fairly reductionist visualization of the role.

A comprehensive definition of an analyst is a person who maintains a high level understanding of the game by collecting as much relevant data as possible, constantly reviewing the data with experts in the community to interpret the details, and reporting on meaningful and actionable subset of information. The most challenging aspects of the role are remaining objective in the face of constant barrage of subjective commentary, cultivating and maintaining a network of relationships that can constantly help you critique your analysis, and most importantly, efficiently communicating all relevant information to the remaining support staff and team.

The figure above shows a sample subset of analyst directives, where the width represents relative importance to the team and the height represents the time invested or the analyst’s expertise in the topic. The figure will look different for teams depending on their priorities and the amount of support staff they have. For example, a team with a large infrastructure can generally pick up multiple members that excel in niche areas or devote time to specific sections, while it may make more sense for smaller teams to pick up a full time analyst who can focus on all general aspects, depending on their long-term goals.

 

 

Scouting


Different Regions

In line with the community perception of analysts, support staff often track games from different regions, notably both LCS regions along with Korean and Chinese leagues. A year or two ago, many teams didn't have the ability to invest in a proper coach or analyst, leading to the dominance of empirical analysis where the Korean competitive scene served as an adequate model to draw inspiration from. However, the top teams from other regions adapt the opportunity costs of their in-game decisions to maximize the efficiency of their picks and strategies, something that can't be determined by observation. Thus, if the assumptions are mistranslated, the utility that teams gain from using that innovative advantage decreases considerably.

In addition to watching the other regions, analysts are also responsible for translating the efficacy of particular strategies and picks within the context of their local region or team. Analysts help build an identity around the unique proclivities of their players and establish the team's original hierarchy of pre-game and in-game opportunity costs before using empirical data to bolster their strategy. Essentially, analysts are responsible for watching other regions, but the underlying purpose is to help the team and players make informed decisions about their own strategies instead of simply mimicking other teams. Recurring benefits of scouting different regions include new strategies, playstyles, or meta-shifts that have a seed in external regions. Short term benefits include having a more holistic perspective of opponents at international competitions like IEM Katowice or World Championships.

 


Upcoming Opponents
Scouting specific opponents has a more refined and deliberate approach. At the beginning of each week, a scouting report helps break down each opponent the team is preparing for and it includes everything from player's champion pools and play styles to general level 1s and lane swaps. Depending on the team, there's also a mention of damage output and gold allocation throughout the game and pick/ban patterns to consider what the team prioritizes before and in-game. After a discussion with other members of the coaching staff, it's decided how this impacts the week's scrim practice. For example, if the opponent is proficient at prioritizing a certain role or champion the team is not comfortable playing against, there's a discussion about whether it's worth altering play style to prepare counters or simply rely on specific champion bans. Finally, the day before each match there's another discussion finalizing pick/ban contingencies and specific in-game plans where the coach then gets the team's final feedback to make any changes.

 

Tools of the Trade

From post-game data to Riot's API, there's an abundance of readily available information for teams to consider. After extracting the data into a reasonable format, Tableau and Matlab allow analysts to dynamically consider data or perform in-depth analysis. For example, if a team wants to focus on the early game, the data helps find other teams that generate early game leads to answer questions like, is the gold concentrated in specific lanes or are the leads a result of overall pressure? How do these champions perform if they don't have early game leads? What are the opportunity costs of early dragons vs. towers? The answers to these questions help shape the direction of the conversation, and offer constructive precedence players can reference.

Generally, the data mentioned above isn't seen by coaches or players; analysts act as an intermediary source. But there is data that needs to be presented and manipulated by coaches and players. Excel and PowerPoint are traditional and powerful tools to help accomplish the task. Online excel documents shared within the team tracking scrim information let players go through and see where they did well, what they did poorly and help them present their arguments to the team regarding picks and strategic direction. PowerPoint or other presentation software helps summarize information for vision control, scouting reports, pick/bans and more. Finally, video editing software like Adobe Premiere allows analysts to put together presentations for reviews or introducing new concepts.

 

Introspective Analysis

BoxeR, the renowned Starcraft player and current coach of SK Telecom T1's Starcraft II team once described his strategy as one in which "even if the opponent had predicted it, he cannot stop me." Day[9] similarly advises his audience that "strategy and solid play doesn't revolve around tricks, surprises, or hidden information, but very solid planning and crisp execution."

An implicit and understated responsibility of analysts and support staff in general is to help team's development, rather than merely searching for new tricks or scouting other regional strategies. More often than not analytical resources are focused on external data, while team development follows a reactionary path rather than a proactive one. A methodology that reflects an infantile approach to delayed gratification, teams focus heavily on the short term in order to win games, rather than establishing a foundation for future success. New teams consistently fall prey to this trap, and while they may enjoy sporadic successes, they lack a cohesive identity and rarely achieve greatness.

In order to help direct team growth, support staff have to recognize the type of team approach that will best represent their players. Some teams are innovative and enjoy playing combinations of new or off-meta picks, while others are reflexive, those who have diverse pre-existing strategies and can adapt to various picks and game situations. Teams can be adaptive, those that heavily research and are the first to pick up and master the new meta styles, while other teams prefer a conservative approach, and shift slowly, letting their skill and teamwork carry them through the transition. Committing to an identity helps team create long term plans regarding how to approach patches, tough opponents, tournaments, roster changes and more.

Similarly players' development can take different routes. Mechanically adept players learn certain champions and playstyles at a different rate than tactical players. Other players are neither but are more open to the learning process, allowing the support staff to mold the player that the team needs. Analysts have to recognize the type of players they are working with in order to recommend plans for adapting during patches or preparing for upcoming tournaments and games. Investing the time to learn about the players and team before setting long term and short term goals is important for any team that wants to set themselves up for long term success.

 

 

Final Thoughts

After the coaching role is established, the analyst role becomes the next most important on the support staff. There is a considerable amount that analysts can be responsible for depending on the goals of the team. Some teams simply want to place well in their region, while others want to win worlds or create a lasting team legacy. As the expectations of the team increase, it becomes critical for the organization to invest in good analysts who not only knowledgeable about the game and various scenes, but can also communicate and present their ideas in a clear and persuasive manner. For a team expecting to maintain a top place in the regional standings and doing well at international competitions, the reasonable expectation is 60-80 hours of analysis work per week, from planning and creating content to scouting regions and analyzing scrims and more. We are slowly approaching an era in team e-sports where player and team development is going to become more important than raw talent, and surrounding the coach with a strong supporting infrastructure will ensure a team's long term success.

Not All Leaks are Created Equal

randall weem.png

Introduction

On December 29, 2014, the moderators of the League of Legends subreddit created a thread to discuss the topic of leaks and rumors. In response, DailyDot writer Richard Lewis posts a video on his YouTube channel to defend leaks on three premises: freedom of speech, public right to information, and objectifying the narrative. This piece will critique each argument in order to show that leaks can be good or bad, depending on the context in which they are presented.

The majority of this piece will revolve around quotes from each of these sources: Richard Lewis’s video discussing leaks, the preamble from the Society for Professional Journalists (SPJ) and the work of Kirk O. Hanson, the executive director of the Center for Applied Ethics at Santa Clara University. The tags below contain links to the relevant sources.

 

What is a leak?

A leak is defined as the intentional disclosure of secret information. 

This information can be personal. Consider a man who just proposed to his girlfriend and hasn't broken the news to his friends or family. A colleague who overhears the conversation and shares it publicly without the consent of the couple, despite the intention, has leaked the information. A couple who wanted to broach the sensitive topic with their parents, or have a quiet wedding with close friends, now have the awkward task of handling a narrative they didn't perpetuate.

Similarly, all organizations hold a wide array of information that’s secret due to its financial value, marketing potential, research and development specifications and more. Corporations seek to protect information due to the economic value from not having this information widely known. This concept is relevant in professional e-sports from team strategies to recruitment methods. For example, a player or organization with multiple options can leverage their options and dictate terms. But, a leak regarding their internal discussions mitigates this negotiating power.

 

Value and ownership of information

Similarly, marketing players and sponsors is a relevant part of the e-sports ecosystem. In the video, Richard Lewis goes into further detail:

In alignment with the discussion surrounding value in information, Lewis lists several sources of marketing revenue for organizations, and also points out that the original source of the news garners the most web traffic for that particular piece of information. Therefore, by leaking the information, Lewis diverts most of the web traffic and attention away from the organization and to his publishers like DailyDot. Naturally, there’s a long list of organizations frustrated with Lewis’s reports either because they undermine some strategic plan or diffuse the value of their information. In response to this Lewis comments:

Business regards information as a commodity and the possession of it as an asset. Economists would like to treat and account for information in the same way as physical assets; however, no discipline has provided an accepted model for such treatment although analogies abound.  

As inventory, information goes through the value-added stages of raw material (events or processes to be measured), work-in-progress (information in development), and finished goods (marketable information). Information gathering and presentation require capital investment and human labor. Besides being costly to acquire, information incurs management costs. Like physical assets, information faces quality control inspection before it can be distributed. (source)

Along the same lines, e-sports organizations have information that they want to protect from proliferation until they can redeem it for its value, such as roster changes, management decisions etc. for variety of reasons. But, Lewis claims that information within this context cannot be owned and implies that while he sympathizes with the frustrations of abused trust within their organization, he has the right to publish the information if he can obtain it and confirm its accuracy.

The problem with Lewis’s defense is that there is a lot of ethical and legal precedent to support that information of value can be owned and protected. From the personal privacy rights to corporate ethics surrounding proprietary information and more, the leaking of protected information is a serious issue. Even the SPJ declares as part of their core principles that:

This principle implies that just because you can obtain private information legally, you must also have an relevant reason for then sharing it with the public. Regarding the acquisition of information itself, Lewis himself quotes the SPJ to maintain the anonymity of his sources and goes on to posture his sources’ contractual obligation to conceal information does not concern him:

Lewis elevates his sources to some sort of undercover-hero status, with a tone implying that his sources innocently found themselves embroiled in a conspiracy or danger, and that they are now willing to risk their careers and livelihood to expose some deep-rooted corruption. But when we consider leaks in the e-sports community, only some of them actually expose malicious behavior or intent. Most of the leaks that have sources "close to the subject or team" should not be revealing the information they do. They’re breaching the trust of their friends and colleagues and should not be considered role models in the community.

Hanson elaborates on the ethics of confidential information and sources:

Hanson points out that any protected information that is obtained by a source that violates an obligation to keep it a secret is morally dubious to leak without a valid reason for doing so.

 

Public Interest vs. Public Right to Know

Lewis next weighs in on his reasons for revealing protected information, claiming that he weighs the discomfort to organizations against the benefit of the public and justifies it so:

While the argument may sound reasonable, there is a clear philosophical distinction between what the public is interested in knowing and what it has a right to know. People, even famous public figures and elected officials have the right to privacy despite the details of their private lives being the 'public interest'. Similarly, organizations retain the right to share or withhold information pertinent to their competitive advantage.

Referring back to the SPJ, the code of ethics state that:

Leaks, by definition, imply that the information should not have been accessible by proper channels, and preys on the curiosity of the community (excited to learn forbidden or secret information) to generate revenue for the website, at the expense of the involved parties.

For example, during the off-season, Martin ‘Rekkles’ Larsson, considered by some to be the best western AD carry, moved from the veteran Fnatic team to the new European powerhouse, Alliance. A fan favorite, the spark of a rumor in mid-October regarding this possibility resulted in many speculation threads and self-proclaimed Reddit detectives to search for evidence supporting this claim. 

Lewis, aware of the public interest surrounding the issue, published a series of articles in the span of a few days like "Rekkles in talks with Alliance" to "Rekkles has had enough, buyout option is 15k", garnering front page status multiple times and yielding tens of thousands of hits. There’s however a difference between excited members of the community speculating about rumors and leaked articles posted by journalists claiming to have inside sources:

Lewis acknowledges that when he publishes a leak, the information transforms from a speculation into a state of quasi-confirmation. When two organizations are re-negotiating sponsorship details and contracts, transitioning players in and out of their teams, the process is time and information sensitive. Leaks force all involved parties to scramble and shift their plans to work around the loss of what they previously believed was privileged knowledge.
 
Speculation is an expected by-product of the community, to be taken somewhat lightly, but leaks that feed on the public interest by sacrificing the ability for teams, players, and organizations to conduct their private businesses are a serious matter, especially when their only purpose is to convert this community interest into clicks and revenue.

 

Controlling Spin

Lewis offers another reason to support leaks, by claiming that organizations tend to spin the information to suit their selfish narratives.

There are two sides to the spin issue, first of which is marketing. For the sake of self-promotion, you replace an individual, organization, or identity with a persona that offers a limited or incomplete picture of the real thing. Individuals spin themselves during interviews; advertising spins their companies and products. E-sports personalities and players are no different. One team may explain a roster move as a skill-based change or a personality-shift to support different narratives like competitive edge or dedication to their fans. However, there may be a series of complex underpinnings behind the decision such as salary, appearance, marketing, and more.

As Lewis rightly points out, the other side of the issue is perception. The audience needs additional information to assess if the conclusions offered or perpetrated by the source are valid or accurate. Independent reporters have the opportunity to provide a different perspective on the same issues. They can challenge assumptions, talk to experts, and present their own conclusions regarding the issues. However, Lewis then goes on to explain the mechanics of independent reporting:

Lewis implies that independent reporters are not motivated by the need or desire to spin the information…and then in the same thought, explains how readers that support them helps build reputation and generate revenue…the same reasons why most organizations or parties choose to control the narrative of information in the first place. In order to further contest this point, let’s consider the primary thrust of this video where Lewis leaks information regarding a secret Riot meeting regarding content creation:

In his diatribe, Lewis uses the knowledge of an internal Riot meeting a few weeks before the video to perpetuate the narrative of Riot as a greedy and manipulative organization trying to quash community content in order to have full marketing control over their product and infrastructure…while the evidence he offers to support his claims is tenuous at best.

But let’s delve into some of the issues: While Riot has assumed broadcasting control over the Chinese and Korean scenes, it is far more likely that the reason for doing so is to offer better produced free, content for the viewers. Also the content from in the new Riot talk show barely overlaps that of shows like Summoning Insight and First Blood. It even some promotes (albeit Riot-filtered) community content. From this perspective, these aren't particularly sinister in nature, but rather fairly logical and straightforward decisions from a company standpoint, and while their methods may be inexperienced or vary in efficacy, Riot’s dedication to their players and fans has never been in question.

This perfectly valid conjecture highlights some of the good resulting from Riot’s recent moves. Yes, a discussion perhaps needs to take place regarding the advantages of organic growth by promoting 3rd party content versus in-house investment in production value and marketing control. But clearly, there is more to the conversation than Lewis presents in his editorial.

It is common knowledge that Lewis and Riot have a poor working relationship, the Deman-IEM incident as the most recent conflict. Furthermore, he disagrees with a majority of Riot’s business policies regarding control of their product and surrounding infrastructure. According to many journalists in the community, if Riot begins to publish and promote their own content, it diverts business away from independent sources, Lewis being among them.

So, while Lewis offers a new perspective on Riot’s recent string of announcements, it is certainly not an objective one. He has both personal and professional reasons to perpetuate the negative narrative about Riot, spinning the information and evidence to support his claims, and disregarding the positive effects of Riot’s new initiatives.

In his assessment comparing good leaks to bad ones, Hanson points out that:

It is a misconception that journalists are unbiased or that they have no vested interest in how they present the news. Cable news networks like Fox News have been constantly accused of promoting conservative political positions and criticized for biased reporting. Web-based articles utilize click-bait headers and skew their reporting toward sensationalism to grab readers’ attention at the expense of honest, objective perspectives. 

To summarize, as Lewis rightly points out, independent journalism should exist to offer new and different views. But it does not justify leaking information to prevent organizations from establishing their own narratives. Journalists are still subject to personal biases and financial motivations in their reporting and players/organizations have the right to market themselves in the manner that they choose.

 

Conclusion

Leaks can help provide alternative perspectives and reveal malfeasance, but not all leaks are created equal. Players, teams, and organizations have a right to privacy to conduct their business. If these rights have to be infringed upon, through furtive methods or asking confidential sources to break their NDAs and/or trust of their friends and colleagues, there should be a good reason for it. Leaks that are self-interested and financially motivated and use dubious methods to acquire the information, simply to reveal things before organizations can, should be seen as ethically immoral. The community approval of leaks will not wane. It is human nature to be curious and want access to private information on topics that interest them. The responsibility falls on the journalists to assess what good and harm leaks can do, and act accordingly.

Learning from SK Telecom

Following the mass Korean Exodus, SK Telecom T1 remained one of few teams to retain a full roster of top-level players, resulting in a crucial developmental advantage compared to the other teams that competed in the OGN pre-season. While teams like Samsung were still laying the foundation for their new (albeit extremely talented) roster, SKT forged ahead by testing the limits of relevant strategies within the context of their team, ultimately finishing the pre-season well ahead of the other teams. With the LCS about to start, it’s important that the coaches and analysts look towards SKT as an example; not merely to see what champions are strong in the current meta-game, but also their approach to pick/ban phase, their resource allocation, and their vision strategy.


The Pick-Ban Mind Games


Before delving into the pick ban phase, let’s first consider the landscape of relevant champions in this meta-game. There were 54 different champions picked over 36 games, 7 of which were used in multiple roles. Some of the more contested picks for each role are shown below, along with their pick, ban, and win rates.

 

It’s interesting to note that Gnar has been banned in all 36 pre-season games. Corki and Lissandra, despite being present in 83 and 81 % of the games only show a win rate of 35 and 33% respectively. With the exception of Ahri and Leblanc, the majority of mid champions are flex picks, a list that includes Jayce, Ezreal, Kassadin, Lissandra, Morgana etc. This flexibility also extends to item builds and strategy, allowing champions to overcome weaknesses at various points in the game, further opening up the possibilities for the pick ban phase. For example, Renekton normally struggles against Jayce; however, in their game against Samsung, SKT’s Marin opted into the matchup. With some pressure from Bengi and an unorthodox full damage build, Samsung’s Cuvee ended the game with a score of 1-8-2, 100 CS behind Marin.

These changes make pick ban phase much more interesting than ever before. Coaches, who are now allowed to participate in the process, not only have the responsibility of helping choose the champions but can also help outline strategies and objectives for that particular team combination. With that in mind, let’s consider SKT and their coach Kkoma’s brilliant pick ban phase against Najin em-Fire.

 

Each pick-ban phase is a separate game, played by exchanging information and bound by player skill and time. Recognizing that Najin has the first pick in Game 1, SKT choose to trade power picks for flexibility. Najin opt for Janna as their first pick, followed by SKT picking up Lissandra and Jarvan IV. Najin pick up Corki to complete their power bot lane set, after which SKT pick up Ezreal and Alistar. At this time, SKT does not know where Kassadin is playing or the last two champions, but between Lissandra, Jarvan IV, and Ezreal, SKT can use their last pick for top, jungle, mid, or AD Carry.


 
After Najin pick Zed and Elise for their final rotation, Kkoma picks Lee Sin for Bengi, locking in Jarvan IV for top lane. This forces Najin to lane swap out of the Kassadin-Jarvan IV matchup and allows SKT to avoid facing Corki and Janna in the bot lane. While this game exemplifies Kkoma’s emphasis on pick flexibility, the next game he manipulates power picks to focus on team synergy, putting Jarvan IV and Lissandra different positions and using Jayce and Graves-Janna to disrupt mid and bot respectively.

Four of the five games Faker played in the pre-season were on flex champions like Ezreal and Lissandra. It’s an approach to the game that uses the highly skilled players to adapt for the team, rather than having the team cater to the strengths of the star player. Professional players are an aggregate of many elements: mechanical skill, game sense, flexibility, etc. The pick-ban phase in Season 5 will test not only team’s knowledge of the current metagame, but also how they create an identity around the unique proclivities of their members.

    
Game Economy and Resource Allocation


A game of League of Legends has limited resources on the map for its players in terms of experience and gold. Moscow 5 recognized that Darien’s repeated deaths allowed them to pick up dragons and buffs from the other side of the map, while their enemies lost wave after wave chasing Darien. CLG was criticized for having a singular strategy of feeding Doublelift and letting him carry the team. These are but a few examples of how some teams use resources differently.

The question then becomes, how can teams use resources efficiently? In terms of gameplay, the jungle and support roles are examples of how a team concentrates resources onto the top, mid, and ad carry roles during the laning phase. But once the laning phase ends, each team has a different philosophy on which members receive farm, push out lanes, and control vision. 

A top laner who is split pushing by himself will have more farm than one who is responsible for defending a turret with his team. A mid laner who has to retreat due to jungle pressure gives up a wave of creeps. While these seem like disparate events, teams influence resource distribution within the team based on their objective focus and overall strategy. The figure below shows the distribution of CS within each team during the preseason.

 

Teams like Samsung Galaxy and Najin prioritize farm onto their AD carries at the expense of their top laners, while KT Rolster does the opposite for its top laner. Teams like CJ Entus have a tight distribution, while other focus on their mid laners, notably IM and SKT. This data gives valuable insight into a team’s strategy and how teams will react. Let’s now consider the new SKT, a team composed of members from both K and S teams, specifically the top lane.

Another impressive facet of SKT’s strategy is their deliberate allocation of gold to their players. Unlike Impact, Marin prefers a carry-oriented approach to the game. His aggressive style relieves pressure from the map and has a strong presence in teamfights alongside Faker. However, this requires additional gold for him to buy items and remain relevant against his opposing laner.


 
The figure above shows the average CS difference for each player compared to their lane opponent at different phases of the game. During the preseason, when Impact was in the top lane, Faker and Bang would end on average with 121 and 39 CS above their lane opponent while Impact would be even or just below his lane opponent. In the games with Marin however, Faker and Bang’s CS leads are halved, CS that’s donated to Marin. This conscious effort has been rewarded by Marin’s impressive 10.5 KDA in their wins, compared to Impact’s 4.3. Marin and Bang have helped create a much stronger and relevant identity for SKT that wouldn’t have been possible with Impact and Piglet. If Marin was forced to follow Impact’s established utility role, SKT would not have been nearly as successful as it has been so far.

Western teams have segmented the game into phases of objective control, but resulting in haphazard distribution of resources. Sometimes farm is deliberately funneled to members in ways that is counter-productive to team’s goals. It’ll be interesting to see whether teams like CLG and Curse accommodate their new solo laners, or force them into established roles set by their predecessors.


Final Thoughts


SKT has started the preseason in dominating fashion, dropping only 1 game out of 10. Is this a foreshadowing for things to come? It’s hard to tell. SKT certainly looks like the most polished team; Bengi’s mechanics have drastically improved and the team has found a modicum of synergy with Marin’s carry-oriented style and Bang’s calculated approach. But, veteran teams like Najin and KT aren’t to be dismissed. Najin didn’t play Ohq in their set against SKT, a player who has ensured victory in every game he’s played in the preseason. Additionally, teams like Samsung and HUYA have shown more than glimpses of potential. Samsung especially who have shown time and again that their mechanical skill is top notch, will only get stronger under their veteran coaching staff as the year progresses.

Shifting to the west, the examples above show how SKT have formed an identity, displayed in their pick-ban phase and resource allocation. But if western teams are to compete against other regions, they must innovate themselves, discover their own opportunity costs and fortify weaknesses instead of mirroring other teams. Since Season 2, western teams have played behind the curve, relying on empirical analysis from regions with a more established infrastructure.
 
With Riot’s acknowledgement and financial support for coaches and analysts, it’s time for teams to learn from SKT, not only superficial elements of meta champions and vision strategy, but the deeper insights into how they approach the game to accentuate their players’ strengths within the context of the team. Will CLG and Alliance give the proper resources to their new members? How will Team8 and Roccat use the unique styles of their top laners? It’ll be interesting to see if the west adapts in the new season.