Sport simulations are wonderful. By compressing half a million individual World Cups into a few hours of computer time, we can see true percentages of expected success. Instead of reading anecdotal analysis on each group and team, we can look at numbers and learn Australia and Cameroon have less than 7% chance of making the elimination rounds. Why guess when we can know?
How does the simulation work
All teams have a strength rating (from http://eloratings.net), which is current as of May 18, 2014. Using Elo ratings (ER), an elegant metric for calculating current strength based on weighted recent relative performance, each match’s expected outcome was plugged into a Poisson random number generator. Using weights derived from historical regression analysis and the knowledge that soccer goals per game follow a Poisson distribution, the generator churns out game outcomes. Structuring the matches into the World Cup’s schedule and storing the results for a half million simulations gives us the analysis below. Note that percentages are rounded and so may not sum to 100% or 200% exactly.
Home field advantage
Studies have determined a home field ER has an advantage of approximately 100 points. This analysis is composed of two simulations, one without home field advantage (HFA) and one with it included. HFA is for Brazil only, not for the other South American teams. This article does not use the HFA enabled simulation for analysis. The next article, analyzing the elimination rounds, will use both sets of simulations.
The Luis Suarez Factor
What about Luis Suarez’s surgery? What about Landon Donovan being left out of the USA squad? Adjustments for events like these were not implemented because ER are an aggregate score. Nothing is summed from the player level. How much is Luis Suarez worth? Quite a bit, clearly, but how much? Anything would just be a guess so Uruguay remains at their current rating.
All tiebreakers are implemented in full, including the drawing of lots if all other options fail! Programming tiebreakers is not much fun so the less said the better. To the results!
Some group matches, like Spain versus Netherlands, appear to have a disproportionate effect on the overall results. Here are some key matches and how the odds of winning the World Cup change based on the match result.
Spain v Netherlands – 6/13/2014
The highest combined team rating for any game in the group stages has, as expected, a dramatic effect on the odds of winning the World Cup for both teams. Since Brazil awaits the second place team, winning this match helps avoid a difficult start to the elimination rounds. The simulation results look like this: