2014 FIFA World Cup update: The Concacaf edge?

🔖 Arturo Galletti on Jun 18th, 2014

There is some strangeness afoot in the World Cup.

That is the average goal differential versus the expected as calculated based on the world futbol elo ratings (http://www.eloratings.net/). Africa teams, European teams and South American teams are performing more or less as expected. Asian teams are performing like Japanese bussinessmen on a three day bender but that matches what we see in the NFL (great infograph if you follow that link). Asia teams are doing the West Coast NFL team in the early game in the East but squared, while in Brazil (where I'm sure they're sticking to an austerity regimen worthy of a samurai). The big surprise is that Concacaf teams are loving playing in Brazil. In fact, except for little Honduras getting pasted by France, Concacaf teams have been outperforming expectations in all their matches.

The other point of note is that the two worst performances of the tournament so far versus the teams ratings have been Spain (-4.5 goals from expectation) and Portugal (-3.5 goals from expectation)

Does this make any real sense? I'd say it does and it comes down to Homepitch Advantage. As noted , we see similar patterns for teams traveling east to explain the Asia federation delta. Africa and Europe (with the noted exceptions) are to form. Conmebol teams are not changing time zones significantly but have all rivalries (particularly Argentina and Uruguay) against Brazil which will offset any travel advantage. Latin American fans despise Spain and Portugal and so they can expect hostile crowds at any and all venues in Brazil. Concacaf teams are not changing time zones and will in general enjoy positive crowds. 

The upshot is that, while I'd love some more data, this does lead to some exciting possibilities. What happens if I adjust for the homepitch effects I am seeing? If I give Brazil and all the Concacafs team the full Homepitch adjustment while giving Portugal and Spain the full negative homecourt effect and all the Asia federations teams half that effect I get:

That looks right. Of course that leads to this bit of madness:

The US and Mexico have a real shot at winning the World Cup and breaking the collective Brazilian heart all over again. With that edge, they're better than 50% to reach semis, better than 25% to reach finals and better than 8% to win. 

I pictured this scenario on twitter yesterday:

USA! USA! USA!  I blame Men in Blazers for that particular bit of madness. US versus Russia then Argentina then Mexico then Brazil is like an ABC/ESPN wet dream.

Let's talk a bit of business. As you might know, I did a sim for every fixture before the World Cup (detailed and explained here). You can see the full fixture sim here:

Reader and blogger Seth Burn @SethBurn has taken it upon himself to run a kelly betting simulation using my model as well as FiveThirtyEight's . He's documenting how FiveThirtyEight's model is doing on his own blog. I'm going to do both here.

After an initial hot start for 538, you'll note that my model is starting to take a lead. It is picking all draws for today so fingers crossed. I'll keep posting daily picks on twitter as we go along the tourney.

Let's finish this one strong. Here's the remaining group stage fixtures simulated with the updated Homepitch edge:

And for completions sake, here's the updated tournament sim without the corrections to the Homepitch edge:

Good hunting. I'm off to watch Australia and Holland over lunch!

-Arturo

Loading...