[Dr. Gamera]

Every time I see FIFA rankings cited as an evaluation of a national team's strength, I cringe. FIFA rankings have several systemic problems, such as:

• All losses are equally bad. Lose 0-9 at home to Belize in the Gold Cup, or 2-3 to Brazil in Brazil in the World Cup? Both are worth 0 points. While you can argue that losing anywhere to anyone should never help your ranking, it certainly seems as if a ranking system should punish you more for a nine-goal loss at home to a weak team than a one-goal loss away to possibly the strongest team in the world.
• The average of results is unweighted, even though, for example, a team is limited to earning many fewer points in a friendly than a competitive game (against opponents of similar ranking).
• The unweighted average leads to situations in which a team's ranking points are guaranteed to drop in its next game even if that team wins. The USA experienced this guaranteed loss of ranking points in the Gold Cup: even a win against Haiti (which, as it turned out, did not occur) would have cost the USA FIFA ranking points.
• Applying FIFA's regional strength multiplier to an opponent makes some sense, but applying it to a team itself makes little sense. Why is it worth less for the USA to defeat Spain in the Confederations Cup than it would have been for, say, Brazil or Peru to defeat Spain in the Confederations Cup?

Fortunately, we have alternatives that suffer from none of these problems. The ELO ratings have the theoretical virtue of being based on a system originally used for chess ratings. On voros's web site, he shows the ratings from two of his systems, systems that he continually tunes and tests for optimal performance (although his latest rankings are a little out of date, as they don't yet incorporate the most recent World Cup qualifying matchday.) If you prefer more subjective human involvement, then for teams that have qualified to the World Cup, you can use betting odds such as these (although subjectivity brings its own problems, including shorter odds for teams that are favored by the betting public, such as, for oddsmakers in the UK, England and perhaps even the USA.)

Consider this a plea for citing some other ranking system other than, or in addition to, FIFA rankings in your posts.

[hairynippleman]

Originally Posted by Dr. Gamera

Every time I see FIFA rankings cited as an evaluation of a national team's strength, I cringe. FIFA rankings have several systemic problems, such as:

• All losses are equally bad. Lose 0-9 at home to Belize in the Gold Cup, or 2-3 to Brazil in Brazil in the World Cup? Both are worth 0 points. While you can argue that losing anywhere to anyone should never help your ranking, it certainly seems as if a ranking system should punish you more for a nine-goal loss at home to a weak team than a one-goal loss away to possibly the strongest team in the world.
• The average of results is unweighted, even though, for example, a team is limited to earning many fewer points in a friendly than a competitive game (against opponents of similar ranking).
• The unweighted average leads to situations in which a team's ranking points are guaranteed to drop in its next game even if that team wins. The USA experienced this guaranteed loss of ranking points in the Gold Cup: even a win against Haiti (which, as it turned out, did not occur) would have cost the USA FIFA ranking points.
• Applying FIFA's regional strength multiplier to an opponent makes some sense, but applying it to a team itself makes little sense. Why is it worth less for the USA to defeat Spain in the Confederations Cup than it would have been for, say, Brazil or Peru to defeat Spain in the Confederations Cup?

Fortunately, we have alternatives that suffer from none of these problems. The ELO ratings have the theoretical virtue of being based on a system originally used for chess ratings. On voros's web site, he shows the ratings from two of his systems, systems that he continually tunes and tests for optimal performance (although his latest rankings are a little out of date, as they don't yet incorporate the most recent World Cup qualifying matchday.) If you prefer more subjective human involvement, then for teams that have qualified to the World Cup, you can use betting odds such as these (although subjectivity brings its own problems, including shorter odds for teams that are favored by the betting public, such as, for oddsmakers in the UK, England and perhaps even the USA.)

Consider this a plea for citing some other ranking system other than, or in addition to, FIFA rankings in your posts.

very nice evaluation. good content. i too cringe. well done.