If the old saying “the house always wins” is to stay a truism, then the house better start investing in artificial intelligence.
A team of researchers recently used AI to predict the outcome of 200 games in the National Hockey League (NHL) over a 20-week period.
The results, published in the paper “Artificial Swarm Intelligence vs Vegas Betting Markets”, were pretty staggering.
Specifically, the AI had 61 per cent accuracy – well above what their bookmaking counterparts could manage, with Vegas hitting the nail on the head 55 per cent of the time.
It may not sound like much, but over the course of 200 games it would be considered enough of an anomaly to see a person kicked out of a casino.
Really compounding the issue was that the AI also generated a pick of the week at a rate of 85 per cent accuracy over the 20-week period, yielding a massive 170 per cent return on investment.
The betting results came about by utilising a form of AI called “Swarm AI”.
Swarm AI is modelled after swarms in nature – such as fish, ants and birds – which are able to perform better as a group than as individuals.
Humans did not develop this ability to swarm through evolution, but we are now applying it to our technology.
Unanimous AI – the company behind the study – explain that Swarm AI “provides the interfaces and AI algorithms to enable ‘human swarms’ to converge online, combining the knowledge, wisdom, insights, and intuitions of diverse groups into a single emergent intelligence“.
“The results of this study are extremely promising,” said Gregg Willcox, director of research at Unanimous AI and a co-author of the paper.
“And while it’s fun to predict sports, we are currently applying the same techniques to a wide variety of other domains, including financial forecasting, business forecasting, and medical diagnosis, all with positive results.”