So I, like many South Africans, tuned in to watch France take on Croatia for the FIFA World Cup 2018 Final last week Sunday, where Les Bleus lifted the trophy after winning 4-2.
While one of several favourites, France was not the main contender to win the tournament, with Germany leading 9-2 in the odds according to Superbook in Las Vegas.
One company that did, however, pick the French to win in Russia was EA Sports, running multiple simulations for the tournament using the information and statistics from its popular game FIFA 18.
The announcement followed the release of a World Cup update for FIFA 18, where gamers could compete in a single player tournament or played against one another with one of the teams heading to Russia.
Why does it matter?
So why am I writing about a game simulation that picked the correct winner for a real-life sporting event a few weeks ahead of time?
Well the short answer is that EA Sports has a fairly solid track record when it comes to predicting winners for such events. This is in fact the third time in a row it has chosen the correct World Cup winner, and it’s gotten the NFL Super Bowl champion right on nine of the past 13 consecutive occasions.
It’s also something we discussed during the most recent episode of htxt’s.africast.
My question then is this: how does EA Sports do it, and could their tech be used to improve yours odds when betting?
The future of betting?
Part of the reason why EA’s prediction came right was thanks to the clever crowdsourcing of simulations to its gaming community.
More specifically it collated data for one month’s worth of play from the aforementioned World Cup FIFA 18 update. With it, more than 406 million virtual games (on console) were played, with one million goals being scored and 1.7 billion keeper saves recorded.
The most telling stat though was France leading the wins per country, garnering 248 000 victories and just surpassing Brazil and Germany with 226 000 and 105 000 respectively.
As such, despite the punditry and the bookies, both of which tipped either Germany or Brazil to win, it was data collected from the gamers that play FIFA day in and day out that proved correct.
With the world of sport betting looking for any kind of technological advantage to get more accurate predictions, could the crowdsourcing of data gathering to gamers be an untapped market?
Companies farm out clicks for their products on occasion, why not do the same when a large-scale sports event rolls around?
The next logical use case for sports betting has to be AI.
With likes of DeepMind’s AlphaGo and the Elon Musk-backed OpenAI already putting their digital wits against the world’s best gamers, it’s only a matter of time before their processing capabilities are put towards determining who will win the next big sporting tournament.
Factor in that these AIs can do in a few days what took the FIFA community a month, and I wouldn’t be surprised to see someone put them to use in much the same way that AI is being developed for the stock markets.
The only question left to ask is: how much you’ll be betting when EA Sports picks the winner for the 2022 FIFA World Cup in Qatar?