Data Science in eSports
How competitive gaming and data are intertwined
I remember the first time I played video games. I was 5 or 6 years old and I was at a friend’s house when I was living in Belgium. His family had a PlayStation and his brother was playing a motorcycle game. I was fascinated and instantly hooked. A few years later after this event, my parents bought a console for my sisters and me after enough begging.
We played amongst each other and with friends who would come over and due to the inherent nature of games, things got competitive. I remember getting particularly hooked on Super Smash Bros for the N64 and was excited when there was a small tournament at my school in La Paz, Bolivia. Over the years, the love I shared for competitive gaming with many others of my age group only deepened as technology advanced and companies began to capitalize on this passion for competitive gaming by setting up bigger and bigger tournaments for cash prizes. It is for this reason that it is no surprise that with the advance of technology and interconnectedness of the internet, eSports have risen into a billion industry in recent years.
Data and eSports
Before I begin talking about Data in eSports, for the uninitiated, eSports are competitive games with certain sets of rules much like football or chess (yes, chess is a recognized sport). The difference between other sports and eSports is that eSports are played virtually and the set of rules is confined to the design of the video game being played. The genres of video games that most naturally transition eSports are Multiplayer Online Battle Arenas (MOBAs), fighting games, Real Time Strategy games (RTS) and First Person Shooters (FPS) although many other genres not mentioned have big eSports tournaments. Some single player games are even made competitive by racing to see who can beat the game the fastest (speed runs).
Data for balancing games
Much like real life sports, eSports are analyzed and statistics are gathered about the game, strategies and players to make better strategical decisions, popularize players (and increase their sale price), how to profit more from events, and are used to hype up the crowd.
Unlike real life sports, because eSports are virtual, it is easier to automatically gather and compile statistics from matches. eSports also tend to have much more complex interactions in terms of game mechanics and balancing the game and use gameplay statistics to constantly update the game.
A simple example of this would be a fighting game. In fighting games there tend to be many different characters. If one character punches and it hurts the other player’s character twice as hard numerically as any other character can, everyone would pick that character and the game would be deemed imbalanced. Data is gathered in the matches and in the next update patch, that character’s punches will be tuned to hit about as hard as the other characters.
While a crude example, all competitive games are trying to balance their games so each character has an even playing field, and they do this by gathering many different data points from games being played.
Data for player decisions
When playing competitive games, players are usually made aware of certain statistics about the game itself or about their own gameplay that they can easily access to make decisions as to which character or map to pick or to help them strategize when playing the game.
Outside of the game, many players create strategy guides to improve gameplay which are usually founded on the data they have access to in the game as well as their own personal experiences.
Data for Player Statistics and Betting
Much like in real life sports, certain players are adept at using certain game mechanics and become known for their strengths and playstyles. The same thing goes for teams and statistics like win rates, which characters they play best and their best maps are used to measure the odds that one team will beat another. Much like real life sports, eSports betting is become popular and can now be found on sports betting websites. To showcase the rise of eSports betting further, Atari is building the first eSports centered hotel in Las Vegas so that players can play and bet on matches.
Artificial Intelligence and the Future of eSports
For many different competitive games, there has usually been the option to practice versus a “computer player” in which a computer plays against the player in the game and the computer is designed with a set basic behaviors that attempted to emulate a real player. Now with the advancements of AI and neural networks, designers are able to create computer players that are able to learn and adapt to the game like never before.
In 2017, at the biggest yearly eSports event, DotA 2’s The International, there was an event where an AI was uncovered and some of the top professional players were made to play one-on-one against the AI and were consistently losing badly which blew everyone away. Some of the players were determined to win and, after the event, learned how to beat the AI. Nevertheless, as the neural networks continue to improve, these computer players might prove to be unbeatable which raises a lot of questions for the industry. Will there be large viewership of game matches of AI bots trying to win against one another? How will players try to utilize these AI networks to try and cheat when competing?
Due to their virtual nature, eSports facilitate the gathering and spreading of gameplay statistics to inform designers and players on how to make better decisions when it comes to balancing and playing the game. The field of eSports is projected to grow and there should be more opportunities for data scientists and data engineers looking for employment in the field. It will be interesting to continue to follow eSports and see how and if AI has a notable impact on the games and their communities.
I hope you enjoyed this article and thanks for reading.