Danilo Borges / Portal ca Copa, Brazil, 06.14.14 Action on the field between Uruguay and Costa Rica in 2014 FIFA World Cup tournament gives UCI cognitive scientist Michael Lee a chance to test "crowd wisdom" algorithms for predicting winners.

UCI cognitive scientist scores big using crowdsourcing to predict World Cup outcome

Algorithms use Ranker data to predict host country Brazil as 2014 soccer champs

Millions of soccer fans around the globe watched in shock as World Cup defending champion Spain was knocked out of the running for a repeat title by losing to the Netherlands and Chile.

But UC Irvine cognitive scientist Michael Lee wasn’t surprised. He says the team to beat is host country Brazil and he bases this call on his “wisdom of the crowd” modeling skills. He has put those skills to work in partnership with Ranker – a website that solicits crowd opinions – to predict the outcome of the World Cup.

“The World Cup tournament structure places strong constraints on possible outcomes – something a good prediction should follow,” says Lee.

Together with Ravi Selker, a visiting graduate student from the University of Amsterdam, and Ravi Iyer, chief data scientist from Ranker, Lee created algorithms to aggregate and analyze different types of rankings provided by the site’s users, as well as the tournament structure, in order to arrive at their prediction.

Results, compiled before World Cup play began, can be found online. They boldly – and what appears to be quite accurately – predicted that England wouldn’t make it out of group play while huge underdog Costa Rica would make it through. Their final tally puts Brazil in the lead, followed by Germany, Argentina and Spain. With only one of these teams currently out of contention, their predictions may be right on goal.

Previously, Lee used the site’s data to predict  celebrity deaths, sales of pizza toppings and movie takings at the box office in 2013.

 

Also: