Friday, July 07, 2006

World Cup Final

I've been tracking win probabilities for the World Cup since the quarterfinals, using ELO rankings, and so far, the ELO rankings have called 4 of the 6 games (failing to predict Italy's win over Germany in the semis and France's quarterfinal win over Brasil). So as we head into the final weekend, here are the updated ELO rankings, following the method outlined here to update after the semi-final games:
  • Italy (2044)
  • France (2042)
  • Portugal (1979)
  • Germany (1913)
Keeping in mind that Germany gets a 100-point home field bonus in calculating the win probabilities, this generates a win probability of 55% for Germany in the third-place match and a win probability of 50.3% for Italy in the final.

Italy's advantage in the ELO rankings stems wholly from their 2nd goal in the 120th minute after Germany went all-out offensive. That extra goal gave them a two-goal margin, which upgraded the "K-factor" (see the original post) for that match. Without that goal, their ELO ranking would be 2026, and France would have a 52% win probability. Either way, it looks like it should be a good match.

(FYI, has Italy at 56.5% and has Italy as a 5/4 favorite, or 55.6%)


Post a Comment

<< Home