How to Avoid the "Group of Death"

Yesterday, we saw that the U.S. was in a difficult group, but that it was not the "group of death" (i.e. hardest). Rather, the U.S. was in the 3rd-hardest group. (As we'll see later, one could still define it as one of a few "groups of death" in this tournament.) We also saw that some groups were really weak and others were really strong. To understand why the groups wind up so imbalanced so often, we'll look at the selection process FIFA uses to decide the groups. Then, we'll consider a possible way to select the groups to avoid such large variations in difficulty.

The FIFA selection process puts the host country and the 7 highest-ranked teams in one pot. Then, the remaining teams are placed in the three other pots based on geographical affiliation. FIFA does this to ensure that no group has more than two teams in it from the same region. The result, however, is the unbalanced groups, as we've seen. Three groups each have three teams in top 16 (the ones who should advance to the round of 16 based on their ranking), while two groups have only 1 team in the top 16, giving an easy path for the highest-rated team in the group and giving the other three teams a (perhaps undeserved) shot at advancing. For example, in group C, a team no higher-ranked than 20 will advance to the round of 16, while in group F, a team ranked no higher than 25 will advance to the round of 16. Meanwhile, in group G (the USA's group), a team ranked either 3, 7, or 13 will go home, while in group B, a team ranked 2, 5, or 9 will not make it through.

This inherent unfairness in the current draw system suggests that a better system should be developed to allow the higher-ranked teams to have no artificial barriers to the group of 16. This fairer system would place the teams into four pots, like the current system, but with one difference: the teams would be assigned to pots based not on geography but on ranking. The host and the top 7 highest teams are in pot A, then the next 8 highest-ranked in B, etc. Then, a team is selected at random from each of the pots to make up each group. The "no more than two teams in a group from the same region" rule is eliminated because there are numerous teams from Europe (9 in this World Cup), so preventing 3 European sides from being in the same group is next to impossible, particularly since the goal is not protecting continents but giving the best teams the best chance to advance. It is highly unlikely that more than two teams from another region (Africa or North/Central America, e.g.) will be in the same group.

To see how this process might work, I took the current World Cup teams and put them into pots based on the October 2013 FIFA ranking (this will allow for a comparison to the current draw process). Then, I randomly drew one team from each pot and put them into a group. I then calculated the average ranking for each group. I repeated this process four times to see if there was significant variation in the results. I found that under the revised selection process, each group is a bit easier than the current system with an average rank of 21 for the revised system compared to an average rank of 20 for the FIFA system. (Remember that lower numbers indicate higher rankings and thus harder competition.) That is not a terribly large difference in average difficulty. The variation in difficulty, however, was significant. In the revised system, the standard deviation was 3.4, the minimum (hardest) group was 15.25, and the maximum (easiest) group was 28.25. Under the FIFA process, the standard deviation was 5.8, the minimum 12.25, and the maximum 31. In other words, the revised selection process produced a tournament of comparable difficulty to the current one, but it made each group more "balanced" in favor of the higher-seed teams.

 Moreover, this revised process did a far better job preventing a "group of death." (Note that when a tournament is composed of 32 of the best teams in the world, some groups will invariably be more difficult than others.) For the purposes of this analysis, I will define a "group of death" to be one in which the average ranking of the teams involved is below 16. In the four different draws I did, only 1 group (out of 32) had an average ranking of below 16. That group consisted of Germany (2), England (10), Ivory Coast (17), and Algeria (32). A challenging group, certainly, but not horrible. Compared to the current process, which created 3 groups with an average ranking below 16, the new process is better at preventing a "group of death," thus ensuring a fairer tournament.

Of course, this whole analysis assumes that FIFA wishes to have a group stage that is "fair." They may be satisfied with the current set-up, which tends to favor elite, European soccer nations. If, however, I were a national association whose team was in the top 16 but not the top 8, I would be lobbying hard for FIFA to go a group draw system such as the one I have outlined. Certainly as an American soccer fan, I would love to see FIFA adopt a revised system. The USA will probably never make the top 8 in soccer, so unless the system is revised, America's path out of the group stage will often be more difficult than it will be the soccer superpowers. Such is life, and it is this greater difficulty that makes the USA's ability to get out of the group stage as often as it does a more impressive feat.