How Do You Play a Perfect Game? Modern Esports Is Turning to Big Data to Find the Answers

It's not only traditional sport that is using big data analytics and hybrid cloud technologies to improve performance. Professional esports teams and broadcasters are using it in new and innovative ways.

Esports is big business. How big? The 2017 International Dota 2 Championships had a prize pool of $24 million (£19 million). Hosted in the 17,459-seat Seattle KeyArena, the final was viewed by over five million simultaneous viewers online.

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 totalling a cool $3.5 million (£2.5 million).

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.

Esports is big business, attracting huge prizes and vast audiences.

Games can be recorded, re-watched, analysed, and fed into databases so they can be referenced and compared. In years gone by, pro players would need to physically attend tournaments to watch their rivals play. Nowadays, creating pre-game strategies, analysing matches post-game and guiding training regimes are all informed by access to match footage and data analysis online.

Discover how Intel® technologies can unlock the potential of your business data ›

“Back when I started there weren’t really many resources [for data],” says esports broadcaster and analyst Harry Thomas (aka Lethal_HT). “Now we are spoilt for choice. With Elgato or AVerMedia we can record our scrims [practice games], our matches and tournaments. We can use it to correct mistakes, look back on how things have progressed and how we can move forward as a team. In ten years, you’ll probably see a lot more people creating APIs to record stats further, you might get heat maps, eye tracking, anything like that.”

It’s already possible to glean some statistical info on players and matches from dedicated websites like overwatchtracker.com and dotabuff.com. But the ability to dive deep and distribute that data in a digestible way isn’t always an easy process. It’s why specialised gaming analytics firms have sprung up, not only capturing the vast wealth of match information available, but using algorithms to automatically suggest improvements.

Mobalytics[1] pitches itself as the first ‘personal performance analytics platform that highlights your strengths and weaknesses to help you boost your game.’ Devoted to League of Legends, it measures in-game performance for different skill sets such as farming, teamplay and consistency and then generates a personal ‘GPI’ (Gamer Performance Index) score. This translates into ratings for a player’s strengths and weaknesses, numbers that can then be crunched to produce actionable advice on how to improve performance.

Analysing game data for Overwatch can reveal weaknesses that can be corrected by bespoke training regimes.

In God we trust, all others must bring data

“League of legends is an incredibly complex game,” says Amine Issa – co-founder and War Chief of Science at Mobalytics, and ex-Fnatic player. “You have 139 champions, all sorts of permutations that constantly occur, the game is constantly changing – it’s chess on steroids in terms of strategic branching paths.

“The tools [to measure it] get bigger and better – rather than me printing out a stack of papers and doing the analysis myself, we can train a computer to think like our team and to process a vast amount of data for different players, provide that information and do analytics for a huge amount of people all at once.”

Using data-based machine learning in this sense is not so much about replacing human intuition as augmenting it with a second opinion to guide practice.

“There’s this great saying that goes – In God we trust, all others must bring data,” Issa adds. “So, you could create a clone of yourself for your team… This clone could look at things in a way you’ve programmed it to, and then you’ll basically always have this powerful second opinion. And you understand the strengths and weaknesses of this second opinion, which you can cross check with your own biases and your decision-making.”

As well as helping players improve, data analytics is also heavily employed by the platforms broadcasting the action. Twitch and YouTube use analytics to recommend content based on previous viewing information. It’s also employed to automatically moderate chat systems to filter out inappropriate comments.

By using analytics, winning becomes more about insight rather than luck.

Many esports streamers use StreamHatchet, software that provides live dashboards of viewer counts, comparison between streaming sessions, information about the audience reached, and team analytics. YouTube offers services such as this under the analytics tab in the streamer’s channel, while Google Analytics can provide extra information, such as geolocation data.

At the heart of esports are the tournament operators. ESL[2] organises some of the largest events in the world – 173,000 people attended the 2017 IEM in Katowice, with many millions more watching the action live or later via VOD. The firm has dedicated technology and its data teams are looking at how to further use data analytics and hybrid cloud computing to enhance their events.

“Data analysis of the match data is certainly a growth area,” says James Dean, UK Managing Director of ESL. “It’s something we started doing several years ago… We’re taking the data from live games and tournaments in a log format and are now starting to process that. We can pull out a huge amount of information – and there are lots of different uses for that information.”

Like any form of high profile competition, the betting industry has made moves to capitalise on the huge viewership figures that esports generates. A subject of controversy in its own right, the pressure is on to make sure everything is within the rules. ESL employs data analytics to help regulate and highlight any potential wrongdoing, whilst also allowing above board betting to work.

“As the stakes get bigger in esports and the prize money is getting bigger, and the number of tournaments around the world grows, we need to make sure everything is safe and secure in terms of the integrity of the tournament itself. A huge amount of the data is used here –  not only are we looking for anomalies for play in tournaments, especially if they are online, but we are also using that data in the gambling industry. They can actually start creating markets for in play betting, and matchmaking odds. There’s a huge industry growing just purely from that fact that data analysis exists in the first place.”

Segments of this video are used by kind permission of ESL

One of the ways in which ESL stays at the cutting edge of data analytics is by working with universities to collaborate on research. One result of this collaboration is Echo[3], a data-driven production tool launched last year provides tournament organisers with the ability to automatically detect extraordinary plays and events in live matches and generate graphics designed to help an audience appreciate what is happening.

Leicester university is also researching the relationship between data analytics and esports. William Darler, Lecturer in Marketing at the University of Leicester School of Business believes that they are only going to get more entwined in the future, with potential for not only providing intelligence for marketing and professional performance, but also with regards to industry regulation.

“There will be a rise in advertising as companies realise the revenue potential of esports and analytics will be used to customise these adverts,” he explains. “Team scouts may use analytics software to sign on new players based on their performance in online matches, [while] software will be developed to more accurately detect cheating... All of these analytics techniques will help create new regulatory standards for future esports events.”

Cloud computing and data analytics is embedded into the core of esports. Where traditional sports have often struggled to capture performance data, the digital nature of gaming competition means that there’s no shortage of information. Increasingly sophisticated APIs and cutting-edge fields, such as machine learning, look set to open up even more avenues to help players win, boost the viewer experience, help firms understand how to best align their businesses, and regulate the competition itself.

Data has become, quite literally, a game changer.

Unleash the power of your business data with leading-edge Intel® technologies ›

The secret to Formula 1’s success? Big talent meets big data

With millions at stake, every second counts in Formula 1. It’s why teams are turning to big data analytics to overtake their rivals on (and off) the track.

Read more

Engineering football excellence with sensors, stats and data analytics

A football match might seem impossible to model, but with the right data and a powerful analytics platform, teams are beginning to develop algorithms for success.

Read more

How big data is becoming cricket’s 13th man

This is what happens when you take a million dollar Hawk-Eye camera and shrink it down to a $200 sensor in a bat.

Read more