So You Think You Can Predict the World Cup: Part 1
Football is a religion. Football is an integral part of local culture for many Latin American, South European and some African countries. It most certainly is the most beloved sport in the entire world, with passionate players and fans in practically every country.
The potential of the FIFA World Cup final to attract a sizable international audience from all over the world makes it a sporting event unlike any other.
According to official data from FIFA, an estimated 3.575 billion people watched matches at the 2018 World Cup, with 1.12bn tuning in for France’s final win over Croatia in Moscow.
Image courtesy: Pinterest
One fine day in 2014, a certain investment guy says he has a model that can forecast the outcome of the biggest sporting event in the world. How this model came into being, how it works, how it fails and what makes it the best of its kind is a very interesting story.
Joachim Klement, a research analyst and former Chief Investment Officer with over 20 years’ experience in financial markets, came up with this model in 2014. For the 2022 World Cup, his model predicted that England and Argentina will go head to head in the finals with Argentina winning the championship and taking home the trophy for the first time in 36 years.
To quote Klement himself, he produced the model to “satirise the economists’ hubris in trying to predict everything”.
However, in this attempt at mockery, he ended up giving the world a model that forecasted 3 out of the last 3 world cups correctly, a feat that no other predictive model has achieved so far.
Who is this man talking about deriding his own professional community? A very smart economist, I must say. In classic economist fashion, he says:
“If I am right, it’s skill and if I am wrong, it’s somebody else’s fault.”
-Joachim Klement,
Strategy, Accounting, and Sustainability FIFA World Cup Predictions
Image courtesy: Joachim Clement, Twitter
The model has its roots in a study by Robert Hoffmann, Lee Chew Ging and Bala Ramasamy from the University of Nottingham. According to Klement’s model, there are some really important factors at play that make a team good that don’t necessarily have to do with the team’s individual players but with the infrastructure the country has in place.
These factors are:
- GDP/capita: Football can be played anywhere, which is a positive thing, but for a nation to have a successful national team at the World Cup, a country needs some crucial infrastructure. To tap into the talent pool already there, quality football fields and academies where young players can be nurtured are necessary. Therefore, a country’s football team will probably be better the richer that country is. On the other hand, look at Nigeria and other African countries. They might have a massive talent pool but not the money to put the infrastructure in place to actually develop these talents to become world-class players.
Economists always function at the margin. And since this is an “economic” model applied to football, one needs to consider the fact that the marginal success rate here, like every other marginal rate in economics, starts to drop and eventually turns negative. If a nation is overly wealthy, children may forgo playing football in favour of other sports or video games because even today, football remains a game mostly played by working- and middle-class people.
- Population size: A nation’s talent pool is larger and more likely to produce athletes of the highest calibre who excel on the global stage if it has a larger population. But why then do nations like China and India struggle in football? The talent pool only matters if football is a mainstream sport, which is the problem. (Klement says, “India chose to play cricket, God knows why”). Researchers discovered in the aforementioned study that population size mostly affects countries in Latin America, where football is a religion. Thus, India and China are outliers in this study.
- Climatic conditions: It becomes extremely difficult or even impossible to play football in a country that is too cold or too hot (Qatar?!). It appears that 14°C, or around the average yearly temperature encountered in parts of Southern Europe and South America, is the perfect temperature for football. Is it a surprise then, that, with the exception of England in 1966 and Germany in 1954, 1974, 1990, and 2014, nations in these regions have won every World Cup?
- Hosting: Although not as big a factor as others, being a host nation is helpful, too. A team has a modest advantage when the bulk of the stadium’s spectators are rooting for them, and this advantage might mean the difference between winning, drawing, and losing.
The model can explain about 45% of the variation in national success in a World Cup using these variables. However, this indicates that luck accounts for about 55% of the final result. Thus, when predicting the results of games involving any two teams, the model incorporates a component of chance. Yes, a team can have a good chance of defeating another team, but surprises do occur and these factors are accounted for in the model.
Most football fans can skip through this next segment that explains how the different stages of the world cup work.
Teams compete in eight groups of four teams each during the group stage. Each group competes in a round-robin tournament, where each team is required to play three times against other teams in the group. This indicates that each group has six matches overall. A group’s teams are ranked using points. Each group’s top two teams advance to the knockout round. Teams compete against one other in one-off matches during the knockout stage which is a single-elimination tournament. If necessary, overtime and penalty shootouts are utilised to determine the winner. The knockouts stage is also called the round of 16. This is followed by the quarter-finals, the semi-finals, the third-place match (contested by the losing semi-finalists), and the final.
What does the Klement model say about the 2022 World Cup? Read on in part 2 of this article!
–Wani Deoras
SY BSc