International Rankings Royale

I can't say I'm great at keeping up with projects but with the World Cup draw on the horizon I thought it'd be fun to dust this one off, especially as it takes very little time on my behalf. (This is key.) So if you're not a returning guests (most likely) I'll give you a quick run through.

  • In November 2012 I came up with my own rankings because I knew FIFA's were not good enough.
  • That next month I heckled FIFA's absurdity with bumping Haiti 18 spots to 39th (and Costa Rica was 66th).
  • In February of that next year I created the International Rankings Royale to compete against other rankings.
  • After five months (and over 200 games) I showed that a random guy could make better rankings than FIFA (and ELO and SPI).

My method of ranking teams was solely dependent on YouTube videos and then giving them a 0-100 score based off of FIFA the video game's scale. And it turned out to be more accurate. I wasn't trying to prove I was the best at analyzing teams. Instead, I succeeded in proving that that anyone with some knowledge of soccer could make a better ranking system. Unfortunately, it does take some time to do this. There have been 902 games (FIFA and non-sanctioned games). At 5 minutes a video it really starts to rack up. So I retire as the unofficial champion.

Moving forward, ELO typically is exalted as the best ranking system, far above FIFA. And no one really knows what to do with SPI. To compare them all, I took the rankings (frozen at the start of the month) from each system and said that for every 20 spots a team is, on average, 1 goal better. So a 20th ranked team and a 50th ranked team are split by 30 spots, or 1.5 goals (30 divided by 20). I take the real goal difference from this and tally it up over time.

Screen Shot 2013-12-02 at 1.00.37 AM

I didn't copy and paste the numbers because Wordpress (or at least my site) doesn't handle spreadsheet numbers that well. You can check out all the leg work here.

You have two sections, the second part split into two columns because it would have been too long. On the left, with ELO 1422, it shows the total goal differential from all the games. There were 87 games in January and ELO was off by a total of 142.1 goals while FIFA was only 134.4. Add it all up and ELO is in first for 2013 (so far) by 52 goals while SPI is trailing FIFA by 17 goals. On average, ELO is off 1.576 goals per game, only (however much less that is than 1.634 is) less than 1.634.

The two columns on the right track the three's right-wrong (win-loss) record with predicting games. This is even closer! Crazy, huh? Soooo does this mean that they're all the same? Well let's take a look at the draw scenario

Screen Shot 2013-12-02 at 1.56.59 AM

France arguably gets a pretty big shaft. The purple CONCACAF-AFC region stays the same. Some others can make some gripes and hallelujahs.

What does this mean? FIFA rankings could definitely be better but ELO isn't the end-all solution. And the World Cup seeding is open to some bickering. Outside of that we have to make do with what we have.