9923170071 / 8108094992 info@dimensionless.in
Data Science in Esports

Data Science in Esports

Introduction

Electronic video gaming has extended from being a hobby into a serious sport and business. Earlier this year, eSports officially became a medal event in the 2022 Asian Games. According to data analytics expert Andrew Pearson, the rise of eSports presents exciting opportunities in data analytics and marketing.

There’s been an explosive growth in esports popularity over recent years, fuelled by games specifically designed with online competition in mind. Blizzard’s Overwatch is a case in point. When the Overwatch League debuted in January 2018, 415,000 viewers tuned in to watch.

The stakes are high. Each team in the Overwatch League stumped up $20 million (£14.4 million) for a city franchise. Participating gamers enjoy $50,000 (£36,000) salaries while competing for a prize pool totaling a cool $3.5 million (£2.5 million).

AI and ESPORTS

Just as data analytics is helping golfers, athletes, F1 teams, football clubs and cricketers improve their performance, esports is well-placed to follow suit. As with any sport, winning doesn’t just hinge on skill, dedication and luck. It’s often determined by strategy and the analysis of past performance. The secrets to success lie in data and esports is overflowing with it.

We can divide the idea of AI and Esports into 3 different aspects or perspectives. It can be as AI playing gaming sports themselves, game analytics platforms to provide the insights and details about the players and their gaming behavior and tactics and lastly, data science in the gaming industry to manage the business side of the games as products

AI & esports

Let us have a look at each of the aspects and discuss them in detail one by one!

ESPORTS by AI

Gamers have been pitting their wits and skill against computers since the earliest days of video games. The levels of difficulty were pre-programmed, and at a certain point in the game, the computer was simply unbeatable by all but the most gifted gamers.

Over time, the concept of difficulty levels evolved. For example, “Madden” NFL Football games have four different levels (ranging from Rookie to All-Madden) that make running plays more difficult, while first-person shooter (FPS) games like “Duke Nukem 3D” follow the same type of tiered difficulty (ranging from Piece of Cake to Damn, I’m Good) that makes it tougher to stay alive and kill enemies.

The rise of machine learning combined with the increasing popularity of esports (organized, multiplayer video game competitions that feature professional video gamers gaming against each other with millions of dollars of prize money on the line), may inextricably link AI to gaming and esports.

That said, the most common implementation of AI in esports is in the games themselves. Companies like AI Gaming want to develop smarter AI bots that would compete against each other in an effort to grow smarter and more competitive, while OpenAI, a research lab co-founded by Elon Musk, developed AI that can beat the top 1% of Dota 2 amateurs (though the AI lost a best-of-three match to some of the game’s best players in August 2018).

  • DeepMind’s AlphaGo used some surprising tactics while playing against Lee Sedol. People thought these wouldn’t actually work, but they did. A similar discovery of new tactics and strategies can happen in eSports. Players will have to re-think their every move while playing. Situations like these will give them insight that might not have been possible.
  • CSGO has a ‘6th man’ setup where an observer advises the players on their strategies. A bot can instead replace the ‘6th man’, a form of ‘Augmented advising’. Teams will have to augment the bot’s recommendations into their gameplay. Teams who do this well will be the winners. Since a lot of machine learning algorithms are democratized there won’t be a situation where teams are unfairly matched.
  • Like I mentioned earlier StarCraft II is a game with quite a bit of strategic depth. This also makes the game more difficult for new-comers. The presence of an in-game coach would be helpful. It would speed up the process of getting started on the game and decrease the learning time.

ESports at the end of the day is a form of sports. People tune in from all over the world to watch their favourite teams play and cheer for them. Only this time, they’ll be rooting for 5 players and a bot.

Game analytics platforms

game analytics platforms
Shadow GG

Shadow.gg — a Counter-Strike analytics platform that its creators claim will cause a significant leap in how esports professionals currently approach preparing for a match by giving fast and easy access to a large number of in-game statistics for any match in the platform’s library. Built primarily for teams looking for a competitive edge, the tool aims to help scout opponents, quickly view data, and visualize that data in meaningful ways. It lessens the burden on coaches and analysts to scout demos and lets your coaches, analysts, and players focus on only important rounds.

The core value proposition here is that coaches and players that use this tool will be able to arrive at conclusions about their opponents’ play, and their own play, that is either too time intensive to arrive at through basic demo review, or simply can’t be reasoned about by trying to estimate data from observing matchplay. We can begin identifying trends for teams and players with regards to their tendencies in relation to the economic context, or how they utilize grenades, or how they prefer to retake a particular site, to name a few.

AI in shadow GG

Obviously players still have to hit their shots in-game, and that’s on them; but going into a match armed with detailed information about which way your opponent leans in crucial situations could mean the difference between a comfortable win or a 16:14 loss; so we hope the value of the tool today and where we plan to take it over the next year will become rather self-evident.

NXTAKE — Advanced CSGO Analytics
NXtake

Built by former daily fantasy sports professionals, NXTAKE is a leader in esports analytics and broadcast augmentation. Our company specializes in advanced analytics, data feeds, and esports prediction models. NXTAKE combines big data and simulation to bring next level analytics to the world of esports. Together, we have wagered enormous amounts of capital over the past few years and are sharing our expertise in a new and exciting industry. It can provide real-time analysis, coupled with live streams

Data science in esports

data science in esports | Dimensionless

Targetting new gamers using data analytics

One of the best examples of data science in this area, customer “segmentation.”

This is a HIGHLY desired function within digital marketing because it’s the analysis of your existing and potential markets in an effort to better understand customers.

Doing this exercise, you can take in vast amounts of data from dozens of data sources (web, social, email, forums, media listening, etc) and feed it into statistical models to extract customer segments, like “your potential target market for your new game consists of those that classify themselves as hardcore gamers between the ages of 14 and 31 years old, that play RPGs like Skyrim, and that average a GTX 1070 GPU.”

What you can then do with this information, is to apply that segment to paid advertising strategies. So, when you start the pre-order push, you can make sure that your digital ads are targeted at the people that you were able to isolate to that segment, and not blasted to that CS: GO player that doesn’t like RPGs.

Competitive game pricing

The goal of an effective BI system in the gaming industry must be able gathering gamer data from several types of external sources, and comparing that data with data in internal systems to arrive at conclusive decisions about a customer’s spending pattern, tastes and levels of satisfaction. A large part of the data analyzed in this case may large volumes of unstructured, social-media data.

Improving gameplay experience

Insights from gaming analytics also enable companies to improve the gameplay itself. For example, millions of player records could be analyzed to pinpoint the most likely in-game moments when players quit the game entirely; perhaps a series of quests are too boring or the challenges are too hard/easy based on character level. Identifying these gaming “bottlenecks” is critical to understanding the reasoning & timing behind a game’s churn rate. Gaming Designers and Developers can then re-examine the game’s storylines, quests, and challenges in order to refine the gaming experience and, hopefully, reduce the number of lost subscribers.

Analyzing the devices used by players also helps developers to create gaming experiences that work effectively for their user base. Exploring a dungeon via an iPhone is quite different than doing it using a widescreen attached to a laptop, so developers need to address issues such as screen size, available functionality, navigation, and character interactions. Data analytics empowers companies to address this challenge by modeling and visualizing massive amounts of heterogeneous data.

Game analytics to improve gaming infrastructure

Today, games sometimes have global player bases… so the architecture supporting those users needs to be configured and implemented correctly. Online games are particularly prone to network-related metrics, such as ping and lag rates — these issues are exacerbated during peak gaming times. Again, Big Data analytics enables gaming companies to use server and network data to understand exactly when, and how, their infrastructure is being pushed to its limits. This knowledge enables companies to scale up or down according to player need; in today’s world of cloud-based PaaS/IaaS architectures (where cost is tied to usage), this information can have a dramatic impact on a company’s bottom line.

Analysing competitors

Make a list of games that are using the same theme and some (or all) of your core mechanics. Both released and upcoming. Especially upcoming, because chances are you’ll be judged against them.

Make a basic SWOT analysis for every one of them, but also add an additional field: “How is our game different?” The key word here is “different”, not “better”, so you won’t get caught in wishful thinking “we’ll have better graphics and better balance”. Why your target audience should consider your game instead of another one? People only have so much time to play.

You can also check geographical distribution and stats for released games on Steam Spy or AppAnnie, but, frankly, it’s not that useful at this stage. You’ll look into it later when deciding on focusing your marketing and localization efforts.

If you’ll decide to check geo distribution for similar games, don’t trust it too much — an audience research you did previously will be more helpful. Other games might’ve done something specific to become popular in some countries, like partnering with a local publisher or getting a good video from a local YouTube celebrity.

For example, there aren’t many owners of The Witcher 3 from Poland on Steam despite the game being immensely popular in that country. That’s because most Poles bought the game from CDP.pl or GOG.com instead of going for much more expensive Steam version.

Conclusion

The gaming industry has a long way to go when we talk about the application of full-fledged data science in its applications or AI bots beating world class players in the complex games like counter strike and DOTA. In this blog too, we looked at how different aspects of data sciences are applied in the gaming industry. But what is clear at this point is the power of AI and the myriad companies looking to harness the same. Gaming appears to be poised as a sector ripe for this type of disruption and companies are getting in early to explore the types of ways to profit off of connecting AI developments with esports.

Stay tuned for more blogs!