Betting on esports may be a relatively new product in the online gambling world but when it comes to the data that the games generate it could be argued it is far ahead of traditional sports.
This data advantage arrives for obvious reasons. With football, tennis or any of the other popular means for in-play betting, the task of transmitting information from events taking place on the pitch to the screens of the global betting audience is an offline to online process.
It is also a major resource-consuming task.
Convey the action through a computer keyboard and mouse, however, and suddenly all the data you will ever need is at your disposal, immediately and at scale.
Esports is big data in action as we can judge when we see the news from OpenAI, a company backed by Elon Musk, which recently produced a bot that in the space of a matter of months went from being a complete novice at DOTA 2 to beating the world’s best in early August in a series of one-on-one match-ups.
To achieve this feat, the team at OpenAI downloaded the action from over 5.8 million ‘expert level’ games in the period since November last year in order to generate the information that would inform and teach the bot to learn the rules of the game. With massive online battle arena (MOBA) games such as DOTA 2, that is millions upon millions of data points.
“Esports is not the first sport to generate data that is so critical to those that offer and place bets,” says James Watson, head of esports at Sportradar. “But since it is played on computers, the scale of data that it generates per match is far greater than a normal sport.”
However, there is a counter-intuitive problem with esports data – the latency period involved in much of the data that comes from streams which can sometimes means delays of up to 30 to 40 seconds when it comes to what the viewers see.
In order to effectively cover the ground, Sportradar has partnered with both the largest esports tournament organiser ESL and specialist esports data provider Dojo Madness in order to answer the technological challenges in collection, processing and distribution.
Watson says the technical side of esports data can be an eye-opener in terms of complexity and it means the model from traditional sports isn’t fit for purpose. The amount of data that can be generated is simply too vast and building models than can both handle these data flows and then manipulate them in the context of in-play betting is a suitably gigantic task.
“We’ve deliberately taken this step-by-step,” says Watson. “When it comes to introducing new markets to bet on, you have to win the audience over – and that audience is very discerning. If you get it wrong with them, then the danger is they will turn off the product.”
He adds that both the potential and the appetite is there but he cautions that Sportradar doesn’t want to “simply throw open 100 new markets” on games such as CS:GO or DOTA 2 without having knowing for certain that the system works and that intrinsic systemic challenges have been ironed out.
“Our focus is on two main important technical areas of focus with live betting: that is the speed, quality and accuracy of the live data and then the underlying live-odds models that create the customer-facing product,” he says.
“Investment in both is critical to success,” he adds, saying that the company’s long-term cooperation agreement with DOJO Madness has already created a market-leading product with a focus around machine learning and analysis of extremely large data sets.
“Live data acquisition is a challenge due to the fragmentation of the industry, but due to the complexity and amount of said data, is important to gather via technical means.”
Watson then adds an interesting coda – that this same technical bridge is set to occur on traditional sports as well. “Today’s rugby players are now wearing GPS chips on their backs, generating statistics that will be able to take followers to a whole new level of analysis and engagement,” he says.
“Data is valuable to all users, no matter who they are. Any sport that can find ways to arm its fans and followers with more data and more data points will definitely find interest levels rise.”
Global positioning of rugby players – it truly is a brave new world.