The United States' soccer team isn't the best in the world, and--despite it being 2010 next year--not a championship threat. Given our position, a World Cup campaign is judged to be successful is we make it from the group state (of 32 teams in 8 groups) to the knockout state (of 16 teams). To enter the knockout stage, you have to be one of the top two teams in your group. Tomorrow, the groups will be determined by a lottery and partial-seeding system. Succinctly put, this system screws the United States.
According to Nate Silver's new soccer rankings (called SPI), the United State is the 16th best team in the world, and the 15th best team to make it to the World Cup finals (Russia--ranked 14th--failed to qualify). So, if the groups were drawn randomly, about half the time would we be paired with two teams ranked below us and one team ranked above us. In these cases, we would expect to qualify. The other half of the time, we would be paired with two teams ranked about us and would expect not to qualify. For those doing the math at home, the actual probability of being in an easier group is 57.5%. (Also, if the official, but less-rigorous FIFA rankings were used, the US would be in an even better spot, as the 13th most-highly ranked team in the tournament.)
Often, to prevent the uncertainty of match-ups described above, tournaments will seed teams so that the best teams play the weaker teams in early matches. In soccer tournaments (e.g., Euro 2008) often the top quartile of teams will be seeded in "Pot A", the next fourth of teams in "Pot B", and so on for pots "C" and "D". Each group comprises one team from each Pot. Under this scheme, the US would be in Pot B and be guaranteed to have two weaker teams (from Pots C and D) in its group.
A potentially even smarter systems would used the specific ranking of each team to assign groups. The highest ranked team would be paired with the lowest ranked team as well as the two middle rankings (16 and 17). Sports fans would recognize this type of ranking from the NCAA basketball tournaments' regional groups. In that case 16 teams are split into four groups of: (1,8,9,16),(4,5,12,13),(3,6,11,14), and (2,7,10,15). Again, under this formula, the US would be guaranteed to have two weaker team in its group.
But does the World Cup use either of these systems? No. Alas, the Euro-centric FIFA powers that be have conspired to stack the deck against the U.S. and other teams in our position. The World Cup Seeding system uses Pots (as described in the Euro 2008-scheme) but only sorts Pot A by ranking (these eight teams are called "seeded"). The other Pots are determined by geography (i.e., continent). Here's the breakdown:
- 8 seeded teams (5 from Europe, 2 from South America, and the African host nation)
- 8 from Europe
- 3 from North America
- 3 from South America
- 5 from Africa
- 4 from Asia
- 1 from Oceania
FIFA must takes the non-seeded teams and puts them into three pots of 8. This year, the math worked out well: there are two scenarios that do not split continents across pots. Scenario 1: Seeded / Europe / N.A+Africa / S.A+Asia+Oceania. Scenario 2: Seeded / Europe / N.A.+Asia+Oceania / S.A.+Africa.
Unfortunately, the U.S. is screwed in either scenario. Why? The Oceania, Asian, African, and non-U.S. North American teams are on average worse than the U.S. And those are the teams that the U.S. is guaranteed
not to play in the group stage under one of the two scenarios (or both). Let's take a closer look by examining the actually SPI values (larger is better, U.S.=78.6) of teams who would be in the U.S.'s pot in either scenario:
| |
Scen. 1: N.A.+Afirca
| Team | Cont. | SPI | < U.S.? |
|---|
| United States | N.A. | 78.6 | -- |
| Mexico | N.A. | 77 | Y |
| Honduras | N.A. | 75.1 | Y |
| Ivory Coast | Afr. | 80.6 | N |
| Cameroon | Afr. | 78.9 | N |
| Nigeria | Afr. | 73.3 | Y |
| Ghana | Afr. | 72.9 | Y |
| Algeria | Afr. | 69.4 | Y |
| Median In-Pot SPI: | 75.1 | Y |
| Median Out-Pot SPI: | 78.1 | Y |
|
|
Scen. 2: N.A.+Asia+Oceania
| Team | Cont. | SPI | < U.S.? |
|---|
| United States | N.A. | 78.6 | -- |
| Mexico | N.A. | 77 | Y |
| Honduras | N.A. | 75.1 | Y |
| Australia | Afr. | 75.4 | Y |
| Japan | Afr. | 73 | Y |
| South Korea | Afr. | 70.2 | Y |
| North Korea | Afr. | 56.9 | Y |
| New Zealand | Afr. | 56.9 | Y |
| Median In-Pot SPI: | 73 | Y |
| Median Out-Pot SPI: | 79.5 | N |
|
|
Check out how many worse-than-U.S. teams these scenarios take out of contention for being in the same group as the U.S. In scenario 2, every team in our Pot is worse than us, and the median SPI of the other teams actually leapfrogs the U.S.! What ridiculousness. And yet, because of the way soccer/football is spread across the continents, this will happen every World Cup until: (1) the U.S. becomes a seeded team or (2) the balance of soccer power is significantly altered across the globe. For the short- and medium-terms, we just have to get used to tough draws.
So who wins with this seeding system? European teams (e.g., Slovenia and Slovakia) who barely qualify. They get to rest easy know that they won't have both a seeded team and a very good unseeded European team (e.g., France) in their group.
As for this World Cup specifically, FIFA determined (randomly? by fiat? who knows...) to go with Scenario #2, which really dampens our hopes of a decent draw. (It also means there won't be a U.S. v North Korea match -- too bad.) So that 57.5% chance of being favored to advance that would result from a completely random system? It drops to a
31.5% chance under this absurd, geographically-based system.
So no matter which balls Clarlize Theron pulls out of the hat tomorrow, I have just one thing to say: Go to hell, FIFA, and take your WC seeding system with you!
(I just noticed that the NY Times has a similar article up, but of course it goes into much less depth.)
