Predictive analytics is key to modern sports leagues. It could be about training, the draft, or any aspect of the game.
Esports is no exception.
A new extension to the agreement between SAP and Team Liquid, a relationship that began in 2018, will see SAP's AI machine learning become part of the team's League of Legends character selection process. It starts this week with Riot Games' major Midseason Invitational (MSI).
What does that mean?
A League of Legends match can be lost before the match even begins if your team's characters are poorly drafted. The draft process is similar to a high-stakes chess match, with each team having only 30 seconds on stage to make a pick, trying to maximize their strengths and outwit their opponents. Select and ban champions (characters) in a way that exploits their weaknesses. This process is critical and has traditionally been done manually. Esports teams also face the added challenge of frequent game updates, where strategies and champion effectiveness can change. This dynamic environment requires a level of agility and rapid adaptation.
Team Liquid uses this SAP technology (along with a huge amount of data) before matches at MSI to plan the right heroes/characters to play during the competition. This gives Team Liquid analysts, coaches, and players the opportunity to conduct “what if” analysis and simulate new strategies.
“There's no clear-cut solution in the draft,” said Jesse Hart, Team Liquid's senior director of sports science and analytics. “Some things fluctuate in importance based on a variety of factors.”
In the case of Team Liquid and Hart, those factors include who Liquid is playing with, what champions they have played with in the past, and adjustments made by Riot Games (developer of League of Legends) . “AI can now look at and understand what a team's profile is like. That profile is different for each team, so it becomes a matter of AI matching his pattern.” [solver]we're trying to match the identity of the team in the draft with what they actually do,” Hart said.
Approaching stick and ball sports
Traditional sports teams have been pioneers in integrating technologies such as video analytics and performance tracking to gain insight into player effectiveness and opponent trends (e.g., sabermetrics in baseball and video analytics in soccer and basketball). Please think about it).
Esports is beginning to embrace a similar revolution. Tools like SAP's AI Core allow esports teams to analyze thousands of professional and amateur matches to simulate various draft scenarios and predict outcomes with high accuracy. This not only streamlines preparation, but also enhances strategic depth by allowing teams to more effectively predict and counter opponent moves.
However, in the case of League of Legends esports, the number of games played each year is not enough to get the sample size that Team Liquid wants to teach the AI what it needs to do. To get to that level quickly, Liquid has been tracking games where players go online and compete against other pros and talented amateurs in a type of pickup game (called solo queue).
However, what these solo queue games don't account for is when Riot Games issues “patches” to the game. This often changes the character's (champion's) abilities and attributes, requiring the AI to relearn from scratch.
“Game mechanics limit the amount of games available in a given patch,” said Thomas Esser, Global Sponsorship Director at SAP. “Therefore, we will initially only deal with solo queue data, because there are a lot of games available from those types of games.Currently, the latest he uses 200,000 solo queue games ”
SAP's AI software helps Team Liquid better prepare for League of Legends character draft
