Competitive esports organization Cloud9 fields 14 teams across 12 of the most popular esports. To give its teams an edge over opponents, Cloud9 has brought in data scientists and embraced the power of cloud computing with Microsoft Azure. With data tools like Microsoft Power BI, Cloud9 saves time and gives its players, coaches, and executives meaningful insights into individual and team performance. These insights help Cloud9 better the competition—simultaneously boosting its profile and the popularity of esports.
"By driving the league forward with data science, we increase visibility, boost revenue, and really take our teams and the greater esports world to the next level. " --- Jack Etienne: CEO and Cofounder
Over the past decade, Cloud9 CEO and Cofounder Jack Etienne has had a front-row seat to the explosive growth of esports. Since purchasing his first competitive, professional gaming team in 2013, he's seen the global esports industry approach USD1 billion in annual revenue and gain total viewership nearing 500 million in 2019, according to Business Insider.
"I paid USD10,000 for my first League of Legends team," recalls Etienne. "These days, an esports franchise can sell for USD30 or 40 million. In the 18-to-34 age demographic, League of Legends is the third most popular major professional league sport in the United States. And that's just one of the 12 games we play."
As the stakes have grown for esports, teams have embraced some of the best practices of traditional sports to help players succeed. From sports psychologists to nutritionists to strength trainers, organizations like Cloud9 bring in top talent to gain a competitive edge. Now, that talent has taken a new form: data scientists.
"Data gives us another way to take our teams to the next level," says Etienne. "Because these games are computer-based, we have a lot more data accessible to us than in traditional sports. The trick is finding innovative ways to gather that data and draw meaningful insights from it."
To help orchestrate its data collection and analysis, Cloud9 has partnered with Microsoft to harness the power of the cloud with Microsoft Azure and dig deep into data from its 14 teams with Microsoft Power BI. Cloud9 chose these tools to help overcome its big data challenges.
"We have an overwhelming amount of data to work with," says Etienne. "We feel there's no better company than Microsoft to help us understand all that data and crystallize it into information we can act on in positive ways for our teams. And it's not just the technology—the people at Microsoft have brought a lot of value and passion to this project, too."
Moving to the cloud for scalable data
Cloud9's initial data science efforts have focused on its highly successful League of Legends team. In League of Legends, five players work as a team to overcome an opposing team in battle. Players take on different roles and interact with each other, with in-game items, and against opponents. Competitive matches occur on weekends and Mondays, and the rest of the time players follow an intense regimen of practice games to hone their skills and scout out other players' strategies with coaches' help.
To do that scouting, players and coaches watch recorded video of opposing players in competitive matches and public games—the latter known as "solo queue" games—that take place throughout the week. They seek to identify behavioral tendencies of a team or individual player so that Cloud9 can avoid surprises and plan counter strategies in advance. In the past, this task took a lot of time.
"Human hours were our primary limitation," says Halee Mason, Lead Data Scientist at Cloud9. "An individual coach or player could only review a limited number of games to prepare for an official match. Our analysts collected data manually and tracked it on spreadsheets. This wasn't optimal, so we looked for ways to automate our processes."
Cloud9 sought to computerize the video analysis and combine it with data about solo queue games available from other sources, such as the public API provided by Riot Games, which makes League of Legends. This opened up possibilities for greater match insights, but it also brought on new challenges.
"We wanted to apply computer vision to game data and extract positional information about all 10 players throughout a game," explains Mason. "This gives us the potential for a much more granular view of game action and better understanding of opponents' tendencies and strategies. But when we're looking at many games a day, the data quickly grows into millions and millions of positional points. We needed a scalable platform to accommodate it."
Cloud9 chose Azure as the basis for its new Game Insights Platform due to the inherent scalability and diverse set of services and tools that come with Azure, such as Azure Container Instances and Azure Database for PostgreSQL, a fully managed service for the open-source PostgreSQL relational database management system. "With Azure, we have virtually unlimited computing power available on demand," says Danny Brady, Senior Data Engineer at Cloud9. "We can spin up resources as needed and turn them off when we're done so we aren't paying for idle compute resources when we don't need to. Also, by using Azure Database for PostgreSQL, we can focus on the health and delivery of our application, rather than database management, which is taken care of through this service."
With the Game Insights Platform in place, Cloud9 will be able to closely study the paths particular players and teams take through a game—information that can have a big impact on the outcome. Cloud9 will also be able to tie in additional data sources and add advanced filtering capabilities to gain even more detailed insights. Meanwhile, Cloud9 is using Power BI to bring the data to life and make it understandable to players and staff through self-serve reporting and robust prebuilt analytical solutions.
For a deeper look at the development and architecture of the Game Insights Platform, read this technical case study.
Pursuing new insights from data analytics
A crucial phase in any League of Legends match happens at the beginning, when teams draft a selection of "champions" they'll play in that match. Champions are pregenerated characters with specific strengths and weaknesses, and teams can choose from among the 148 champions available at any given time. Assuring the proper distribution of champions on a team can mean defeat or victory. To make things more complicated, the abilities and statistics of available champions change every two weeks when Riot Games releases balance patches.
Just as it's important to understand opponents' strategies for moving through the game—the domain of the Game Insights Platform—it's equally important to understand their drafting strategies. To increase this understanding, Cloud9 data scientists are working with Power BI to bring the power of data to the players, coaches, and Cloud9 executives.
"We're focused on data visualization and bringing together multiple data sources to provide a holistic view that the team can use to understand which champions their opponents have been practicing with," explains Mason. "We use Power BI dashboards to draw data from our database and keep up-to-date information at our fingertips every time the team needs to make drafting decisions."
For Cloud9 data scientists, working with Power BI felt like a natural transition. "I really enjoy visualizing the data and drawing conclusions from it," says Mason. "With Power BI, I can get right to analyzing the meaningful details and organize a view of the data that informs my hypothesis."
The scientists use Power BI both as an in-depth analytical tool and as a solution for fast insights. "If a player comes to us with a question about a particular champion or an opponent's recent drafting choices, we use Power BI to pull in the data, analyze it, summarize it, and present it to the player right away," says Mason. "In the past we would have had to go to multiple online resources and do a lot of manual aggregations in Microsoft Excel or write one-off scripts using Python."
Cloud9 has also created custom Power BI dashboards for players, coaches, and managers. These present clear information about opponents' strategies plus individual and team performance and progress, which the players, coaches, and managers consult to prepare for specific matches.
"It's awesome to see the players using the dashboards, which give them information in an easy-to-digest, consumable format," says Mason. "The coaches use the dashboards during weekly draft and strategy meetings, and they've become a vital part of Cloud9 match preparation. And Jack, our CEO, is also our most active Power BI user. He keeps his dashboard open so he has a really good pulse on how the team is doing and whether they're meeting their goals."
League of Legends team member Bobby Huang, widely known by his summoner's name of Blaber, has made Cloud9's data resources a regular part of his practice routine. "I use the dashboard almost every day now," he says. "As a player, I don't have much free time, so I don't want to spend it searching for statistics. And if I need something that's not on my dashboard, I ask the data team and then see the numbers right away."
Jonathan Nguyen, one of the Cloud9 League of Legends coaches, who usually goes by the name Westrice, echoes that sentiment. "We've got all the information we need in one spot with just the click of a button. It saves us a lot of time, and it helps get the players engaged," says Westrice. "That makes our draft meetings very efficient."
Going global and leveling up
For Cloud9, replacing old manual processes with the Game Insights Platform and custom Power BI dashboards has been a tremendous time saver and opened up new possibilities. Right now, the data science team has focused its data gathering and analysis efforts on North America, but the scalability of Azure will make it easy to expand into new regions and take the solution global—a key to winning world championships.
The combination of elite players, talented coaches, and a dedicated data science team is already paying dividends for Cloud9. During the Spring 2020 competitive season, the Cloud9 League of Legends team racked up the best winning percentage in league history—92.9 percent—and all five team members were named to the league's All-Pro roster.
Ultimately, Cloud9 hopes that its data science efforts will do more than just help its players and teams succeed. "Esports has the potential to be bigger than traditional sports—particularly in key advertising demographics," says Etienne. "By driving the league forward with data science, we increase visibility, boost revenue, and really take our teams and the greater esports world to the next level."