Home Finance Machine Learning World Cup Predictions by Goldman Sachs Weren't Even Close

Machine Learning World Cup Predictions by Goldman Sachs Weren’t Even Close

The World Cup tournament recently held in Russia will go down as one of the biggest events of 2018.

The football fever took up the attention not only of football enthusiast around the globe but also that of some of the world’s leading investment banking institutions.

In fact, these bankers have in the last three weeks engaged their cutting-edge and costly tools in forecasting the winners of the 2018 World Cup as opposed to making efforts to hone their artificial intelligence algorithms for better investment decision making. Unfortunately, none of them made accurate predictions.

Goldman Sachs crunched uncountable data points about individual players and teams into four machine learning models before picking the most likely outcomes out of a million simulations.

In the semi-finals, for instance, it had Germany, Portugal, Brazil, and France. What’s more, Brazil was supposed to emerge as winners against Germany by scoring 2-1 in the finals. According to a report by Goldman Sachs, German Chief Economist Jan Hatzius cross-checked the final result in exceptional detail.

UBS also made a similar forecast on the contenders in the final rounds of the tournament. The Swiss bank considered Germany as the most probable champion with a 24% likelihood of winning with Brazil and Spain following. The prediction was made based on a team’s scores in the pre-qualification, an objective skill-level measurement dubbed Elo rating among other factors.

READ MORE: Top 25 AI Software for the Banking Industry

ING Group, a Dutch bank, adopted a different method through measuring a team ’s market value, particularly when assuming that a team ’s market worth is closely related to its success. For this reason, the bank predicted that Spain, which is the most valuable team all over the world, would defeat France, the second most valuable team, during the final match. In reality, only France made it to the semifinals.

Well-aware of the inherent unpredictability of soccer, Goldman Sachs updated statistical models during the tournament. An updated bracket prediction that was published on July 9, which was a day before the start of the semis, showed that the bank had projected the defeat of England by Belgium in the final round. Based on the various predictions, it seems like the algorithms used got everything wrong since none of the teams entered the final round.

The most embarrassing bit about the predictions made is that Goldman Sachs’ forecast in 2018 was further off that its initial trail back in the 2014 World Cup. Four years ago, the prediction utilized a smaller data set that dealt with only team-level data including the number of goals scored in the past ten international matches and the teams’ rankings.

Although the recent predictions did not predict the World Cup champion, they managed to mention three out of four teams in the semifinals.

Sven Jari Stehn, Manay Chaudhary, and Nicholas Fawcett, the three economists behind the Goldman Sach’s initiative, wrote that it is challenging to evaluate how much faith one ought to put into such projections. They added that despite the fanciest statistical methods, football predictions are still highly inaccurate. They attributed this reason to the unpredictable nature of the game.

Source Observer

Subscribe to our newsletter

Signup today for free and be the first to get notified on the latest news and insights on artificial intelligence

KC Cheung
KC Cheung
KC Cheung has over 18 years experience in the technology industry including media, payments, and software and has a keen interest in artificial intelligence, machine learning, deep learning, neural networks and its applications in business. Over the years he has worked with some of the leading technology companies, building and growing dynamic teams in a fast moving international environment.
- Advertisment -

MOST POPULAR

AI Model Development isn’t the End; it’s the Beginning

AI model development isn’t the end; it’s the beginning. Like children, successful models need continuous nurturing and monitoring throughout their lifecycle. Parenting is exhilarating and, if...