India Markets close in 3 hrs 34 mins

FIFA World Cup 2018: Here's who will win as per machine learning techniques

Pankaj Thuain
1 / 1

Did you know the FIFA World Cup winner does not get the real trophy: Check out these amazing facts

Earlier a team could get the original trophy as a permanent possession only if the country won the tournament thrice.

Moneycontrol News

Andrew Groll and his colleagues at the Technical University of Dortmund, Germany have combined machine learning and statistical methods to arrive at the most likely winner of the 2018 FIFA World Cup.

As per a report by the MIT Technology Review, Groll and his colleagues used a method called ‘Random Forest Approach’ to model the 2018 FIFA World Cup. They included factors like country’s GDP and population and FIFA’s ranking of teams. They also factored in the average age of players, number of Champions League players in the team, etc. Further, they included the ranking used by bookmakers too.

By following this method, Groll and Colleagues found Spain to be the most likely winner with a probability of 17.8. A big factor of this prediction is the structure of the tournament.

Also, according to Groll, if Germany clears the Group phase, then it is likely to face tougher competition in the 16-team knockout phase as compared to Spain. If both make it to the quarter-finals then they have similar chances of winning the cup.

Groll and colleagues simulated the tournament around 100,000 times and said, “According to the most probable tournament course, instead of the Spanish the German team would win the World Cup.” So, Spain clearly begins the tournament as a favourite to win the World Cup. However, if Germany makes it to the quarter-finals then they have a good chance of winning too.

The FIFA World Cup kicks off in Russia today with the hosts playing Saudi Arabia in the opening clash.

In recent years, machine-learning techniques have gained momentum, as it has the potential to outperform the conventional statistical methods. However, given the unpredictability of the game, it is difficult to determine outcomes in advance using statistical methods. As the tournament unfolds, we will see if machine learning techniques can help us predict the future.