- England ranked 4th in the world after a less than inspiring performance at Euro 2012;
- Norway reaching the heights of second in the early 1990s, despite playing long ball under Egil Olsen and going out of USA 1994 at the group stage;
- Talking of the Americans, their national team was ranked as high as 4th during 2006, which even their goalkeeper Kasey Keller was surprised at!
Before we get on to the current rankings, let's start with a (short) history lesson. Until 1993 there was no official ranking system, although Scotland fans in particular had come up with unofficial alternatives. However, from August 1993, FIFA began to publish a regularly updated list that purported to rank all national teams recognised by them (167 at the time, 208 today) in order.
The original list was very basic - it was organised as a league, with 3 points for a win and 1 for a draw. So beating the Maldives (bottom of the original list) by the odd goal in a friendly counted the same as if your team had thrashed Germany (then leaders of the Ranking) 22-0 in the World Cup Final.
FIFA eventually responded to the increasing criticism of the Rankings by making changes in 1999. These meant that more points would be won for more important matches, and that the actual margin of victory would be taken into account. Yet this new system only lasted a further 7 years before another overhaul after the 2006 World Cup. On the plus side, the evaluation period was halved to four years, making the Rankings more current, and various other factors were balanced. However, some pretty important issues such as goals scored and the difference between home and away matches were dropped.
So, today plenty of people continue to criticise the rankings, and it's worth reading some of this here, here and most notably with the World Cup coming up, here. But rather than writing another "why oh why" kind of article on them, I think it's time to put both the FIFA rankings, and their main unofficial rival, the Elo ratings to the test. I've gone back and looked at the last 3 World Cups. For each, I've compared 5 different predictions:
1. The FIFA rankings, if 1st won, 2nd was the beaten finalist etc.
2. The FIFA rankings applied to the actual draw. So if team A played team B in the group stage, the higher ranking team is awarded the win, and so forth through the tournament.
3. The Elo ratings, if 1st won etc.
4. The Elo ratings, applied to the actual draw
5. The bookmakers favourites, as calculated by the pre-tournament odds.
I've taken the rankings and odds from the end of May, e.g. a week or two before the start of the tournament.
So, what happened in 2002?
Starting with the FIFA Rankings, we have a problem straight away. Colombia were ranked 4th, yet had not even qualified for the tournament in Japan and South Korea. So in come number 5 team Portugal. The Elo ratings also have the same problem, with their fourth-ranked team being the Dutch who also didn't qualify. They too have Portugal in 5th.
The obvious starting point is all were very wrong in predicting the winner - although few would have expected France to implode quite as dramatically as they did in real life. FIFA's high ranking of Brazil means there is at least one correct finalist predicted on the simple rankings, although if everyone played to their rank, France would have knocked them out in the quarter-finals. The Elo ratings fail to predict any of the final four - in fact of the 4 teams the two methods throw up, only Italy got out of the group stage!
Looking at the worked-through tournaments in particular, the FIFA rankings did correctly predict 12 of the last 16, while Elo only managed 10.
What of England? They should have qualified second from their group - which they did, although both ratings and the bookies thought Argentina would top that group rather than Sweden. Both sets of rankings then had Sven's boys going out versus France in the second round, although in actual fact they beat Denmark only to lose to Brazil after that Ronaldinho goal.
So, perhaps the FIFA rankings are slightly in the lead as we move on to the next tournament.
This time all 4 top-ranked FIFA sides qualified. However, the official rankings top 4 performed poorly, with Brazil getting furthest by reaching the quarters. FIFA ranked Italy at just 13th going into the tournament, and so fails to predict the winner reaching even the semi finals. In fact, the worked tournament sees Italy, Germany, Portugal all eliminated in the 2nd round and France going out in the quarters. It does correctly predict England's defeat at the same stage, although to Mexico rather than Portugal.
Have the Elo ratings done any better? Not really. Again, the top 4 didn't do the business, although at least in this worked tournament England beat Argentina! The bookmakers are again closest, this time predicting two of the final four (and it should be noted France 5th or 6th favourites with most odds makers...)
At this stage, the two rankings systems look almost entirely useless when it comes to the main international tournament. But they've got one last chance to redeem themselves, with South Africa 2010. After all, the method of calculation was changed for the FIFA ranks, so maybe this will have an impact?
It would certainly seem that the new FIFA rankings are an improvement. Of course, top-rated Brazil actually went out in the quarter-finals despite an early goal against Holland. However, the ranking did correctly predict Spain to reach the final and in the worked tournament, Germany to finish 3rd - the first exact finish of this exercise! In fact, had Brazil been ranked lower, the FIFA rankings would have correctly got the 1-2-3 in the worked example.
For the Elo rankings, it's a similar situation. Again, Spain went one better than predicted, and either Germany or Holland were correctly assumed to be in the top four depending on how it's worked out. Just as with FIFA, the over-rating of England was one of the main mistakes. The bookmakers produce our first successful winner prediction, but get the other 3 places badly wrong.
Therefore, it might be best to actively bet against the rankings this time around. As it stands, our erstwhile friends at FIFA go 1-2-3-4 Spain, Germany, Portugal and Brazil, while Elo have Brazil, Spain, Germany, Argentina. So, England to win it then?
No comments:
Post a Comment