World Cup odds: what the models say
In case you missed it, the biggest event in the history of team sports begins again today. At 4 p.m., two hours after opening ceremonies, Croatia and host Brazil will kick off the 2014 World Cup.
Several people have built models to generate probabilities for winning the tournament. Here’s a rundown of a few prominent ones.
Three of Nate Silver’s writers worked together to make predictions and some neat interactive graphics. Built on the back of the Soccer Power Index, a predictive metric which Silver co-developed with ESPN in 2010, Brazil is the clear favorite to start, nearly 50-50 to win the whole damn thing. Argentina stands second at 15 percent, followed by Germany at 11 percent and Spain at eight percent.
For more on SPI, see here and here.
The finance giant got similar results with its model. It wasn’t built on as much data as SPI, though—Goldman just used Elo, team goals for over the last 10 games, upcoming opponents’ goals against over their last five games, and two binary variables—whether the match was at a World Cup, and whether it was on a country’s home soil.
Even though these cutoffs are arbitrary and Goldman didn’t consider international friendlies—or the performances of individual players playing for club teams around the world—the important thing is that this model also pegs Brazil as a huge, huge favorite.
Michael Caley, SB Nation and the Washington Post
In contrast to FiveThirtyEight and Goldman’s models, Caley’s is built on shots and scoring chances (when possible) rather than goals, because turning chances into goals is more random than creating chances in the first place. His model still identifies Brazil as the favorite, but Spain is a close second—both around 25 percent. Portugal, Germany, and France are the next three at around 5 percent. For the former two, that may partly be a function of both being in Group G—the “Group of Death,” along with Ghana and the United States. All four teams made it to the knockout stage in South Africa in 2010, but only two can advance this time.
Bilal Hafeez, Deutsche Bank
Apparently, the German bank picked England. England is often overrated, but…really?
“Our models point to Brazil having most chance of winning followed by Germany, Spain, and France. But to leave it at that would be boring and too conventional. So we introduce what some would call our bias but we like to think of as a discretionary overlay.”
Let’s just say the bank doesn’t have the greatest recent history of investing in western Europe.
Marten van Garderen and Ian Bright, ING
These two economists tried to use player value as a proxy for team quality. Their method pegged Spain and Germany as the top two teams, followed by Brazil, France, and Argentina. Unsurprisingly, though, it seems to overrate European teams (compared to South American ones).
It’s not unheard of for animals to perform better than “experts” at predicting the outcomes of sports games. During Euro 2008 and the 2010 World Cup, the prognosticator of the hour was Paul the Octopus. There are several copycat acts this time around. They haven’t predicted World Cup winners yet, though—just winners of some of the early matches.
Nelly the elephant—Germany draws with Portugal and defeats the United States and Ghana.
Big Head the turtle—Brazil defeats Croatia.
Shaheen the camel—Brazil defeats Croatia.
Others, including Chinese pandas and Gentoo penguins, will also be making picks.
Leave a Comment
Be nice. Don't use HTML tags. And consider reading our full comment policy.