Thursday, November 02, 2006

CONCACAF grows up

This is really good news:
No guests in next year's Gold Cup, say CONCACAF

RIO DE JANEIRO, Oct 31 (Reuters) - No guest teams will be invited to take part in next year's CONCACAF Gold Cup, the organisers said on Tuesday.

CONCACAF -- the North, Central American and Caribbean Football Confederation -- said on their Web site that 12 teams would take part in next June's event, all of them from within the region itself.

"We have seen over the past decade a significant rise in the standards and successes of CONCACAF nations," said CONCACAF general secretary Chuck Blazer.
"So now it is time for two more of our countries to be able to compete in the finals of the Gold Cup, as one more step towards raising their game to a new level.

"As the Gold Cup was in its infancy and early growth, guest teams played a very welcome part in the success of the Gold Cup and its awareness throughout the region and, then, the rest of the world.

"Now, like all continental championships, except the Copa America, the Gold Cup is for us alone."

Brazil, Colombia, Ecuador, Peru, South Korea and South Africa have all taken part recently.

CONCACAF said that this time Mexico, Canada and United States would qualify automatically. They will be joined by five Central American teams and four from the Caribbean.

The tournament will again be held in the United States, which has hosted every Gold Cup since 1991 either in its own right or as co-hosts with Mexico.

The competition began in 1963 as the CONCACAF championship and was held on a two-yearly basis until 1971 when Mexico, the region's major power, began to lose interest.

The tournament ceased to be played and from 1973 until 1989, the winners of the region's World Cup qualifying tournaments were simply declared CONCACAF champions.

The tournament was re-launched as the Gold Cup in 1991.

Although it has been held on a regular basis since, it has still struggled to find its place in the international calendar with many countries, including the guests, fielding below-strength teams.
Next step is to host it somewhere other than the US, although I've heard rumours of one of the groups being hosted in Canada in the future.

Wednesday, October 18, 2006

FIFA rankings: Spotlight on Jamaica

The October edition of the FIFA men's rankings are out, and the top 20 looks like this:

Rank Team Pts:
Oct 06
+/-Rank
Sep 06
+/-Pts:
Sep 06
1 Brazil 1560 0 -14
2 Italy 1540 3 66
3 France 1483 -1 -51
4 Argentina 1446 -1 -46
5 England 1370 -1 -107
6 Germany 1339 2 48
7 Netherlands 1313 -1 -14
8 Czech Republic 1253 -1 -59
9 Portugal 1224 0 -48
10 Spain 1198 0 -57
11 Nigeria 1139 0 -27
12 Cameroon 1110 0 48
13 Ukraine 1054 0 25
14 Greece 979 8 146
15 Switzerland 978 -1 -38
16 Sweden 975 2 87
17 Uruguay 903 -2 -57
18 C├┤te d'Ivoire 893 1 14
19 Mexico 882 -2 -9
19 Croatia 882 5 57

As is often the case, the more interesting stories are found a little lower in the table. This month, I've directed my attention to Jamaica's surprising 20-place jump in the rankings, to 57th. With 3 wins and a loss during the most recent window, this seems to make sense. But look a little closer at their results:
Jamaica (JAM) 2:1 (2:1) Canada (CAN)
Friendly 2006 KINGSTON JAMAICA - 08.10.2006
Jamaica (JAM) 2:0 (2:0) Haiti (HAI)
Gold Cup 2007 > Qualifying CARIBBEAN Zone KINGSTON JAMAICA - 01.10.2006
Jamaica (JAM) 1:2 (0:0) St. Vincent and the Grenadines (VIN)
Gold Cup 2007 > Qualifying CARIBBEAN Zone KINGSTON JAMAICA - 29.09.2006
Jamaica (JAM) 4:0 (4:0) St. Lucia (LCA)
Gold Cup 2007 > Qualifying CARIBBEAN Zone KINGSTON JAMAICA - 27.09.2006

Going in reverse chronological order, we have a friendly win over Canada, a side ranked 6 spots higher than Jamaica in September (Jamaica was 77th, Canada 71st). A home friendly win over a similarly ranked team has a negligible effect on points accumulated, as seen by the opposite result, a 1:0 victory for Canada over Jamaica in Montreal in September, which didn't shoot the Canadians up the rankings.

It is the first 3 results, then, that did the job. Scorelines of 2:0 and 4:0 are impressive, but keep in mind that, in September's rankings Haiti and St. Lucia were ranked 118th and 115th respectively. A loss at home, in a regional qualifier, to St. Vincent and Grenadines (129th in September) is a terrible result, and yet it was not enough to prevent a jump nearly into the top 50.

That Jamaica failed to qualify for the Gold Cup from their group, which they hosted, is an enormous failure, and yet they are rewarded with one of the biggest jumps of the month.

The mind boggles.

Tuesday, October 17, 2006

Off topic

I have a few invites available for moola.com, an ad-supported site that gives you a penny and gives you chances to double your money repeatedly by watching an ad and playing a head-to-head game. Leave comment with your email address if you're interested.

Currently it is in Beta and registration by invitation only.

Update: I received 4 new invitations so I now have a total of 5 available. So just say the word and you're in.

Thursday, October 12, 2006

Just average

Since it is already covered, maybe even more extensively, over at Aggr Ranking, putting together an ordering of national sides based on averages of a set of alternative rankings myself is a bit of a vanity project.

But that's just what I've done. You can download the current edition (93kb, PDF) or find it on the sidebar. I intend to publish every month, within the first week or two of the month, based on the FIFA rankings and the following alternatives:
  • ELO: This is probably the most well-established and statistically rigorous of the alternatives.
  • Ziaian: Easily the most home-spun, but he publishes regularly, and the results fall well within the realm of believability.
  • Xocca: My favourite among my German discoveries.
  • Mondfoot: Kind of like the crazy French uncle in the rankings family. Still they publish consistently and have been doing so for a long time, without which the whole exercise is somewhat futile. Plus it brings a bit more ethnic diversity, and a certain joie de vivre to the table.
The method is astoundingly simple. Each of the 5 rankings systems has an ordering of countries. In my ranking, the countries are ordered in ascending order (ie: from lowest to highest) by an unweighted mean of the 5 rankings. A few notes:
  • The list of countries ranked is not exhaustive. Instead I rank only those 205 countries considered under the FIFA rankings. Other systems (ELO, Ziaian) consider more, others (Xocca) less.
  • For those countries which are not ranked under all systems, an average is taken of those in which they are ranked -- that is, the sample size is smaller.
That's all. The top 20 under the 12 October 2006 edition of the averaged ranking are:

Rk Country FIFA ELO Ziaian Mondfoot Xocca
Average









1 Brazil 1 1 2 1 1
1.2
2 France 2 2 3 2 2
2.2
3 Italy 5 3 1 9 5
4.6
4 Argentina 3 4 5 12 3
5.4
5 Netherlands 6 5 6 7 4
5.6
6 England 4 6 7 6 7
6
7 Germany 8 7 4 5 6
6
8 Czech Republic 7 10 9 4 9
7.8
9 Portugal 9 8 8 8 8
8.2
10 Spain 10 10 15 3 12
10
11 Denmark 16 9 12 11 14
12.4
12 Switzerland 14 13 13 23 17
16
13 Mexico 17 17 14 18 16
16.4
14 Sweden 18 12 11 34 11
17.2
15 Croatia 24 16 10 25 13
17.6
16 Romania 25 14 17 21 15
18.4
17 Turkey 28 17 19 16 18
19.6
18 Greece 22 19 22 19 23
21
19 Uruguay 15 15 18 50 10
21.6
20 Nigeria 11 21 48 14 25
23.8

Thursday, September 28, 2006

CONCACAF

Here is an excellent post about CONCACAF's historically strongest teams in the 146 editions of the FIFA rankings since 1993. It's a bit sad to see how poorly Canada has compared to much poorer and smaller countries, but hopefully that will turn around in the next few years. In a subsequent post, he ranks the CONCACAF teams historically with the following list:

All Time CONCACAF Ranking (1993-current)

1 Mexico 5,525
2 United States 5,377
3 Costa Rica 4,953
4 Jamaica 4,927
5 Honduras 4,917
6 Trinidad 4,852
7 Canada 4,505
8 El Salvador 4,192
9 Guatemala 4,141
10 Cuba 3,899
11 Barbados 3,770
12 Haiti 3,674
13 Panama 3,544
14 St. Vincent 3,347
15 Grenada 3,186
16 St. Lucia 3,153
17 St. Kitts 2,972
18 Surinam 2,884
19 Antigua 2,717
20 Bermuda 2,374
21 Domincan Rep. 2,314
22 Net. Antilles 2,109
23 Cayman Is. 2,063
24 Dominica 1,990
25 Guyana 1,826
26 Puerto Rico 1,633
27 Nicaragua 1,535
28 British VI 1,327
29 Belize 1,214
30 Aruba 1,171
31 Bahamas 745
32 Anguilla 482
33 USVI 359
34 Turks & Caicos 192
35 Monserrat 123


The blog, Climbing the Ladder, is a must for anybody interested in USA domestic soccer (MLS), or the CONCACAF scene generally.

Tuesday, September 26, 2006

The Germans are coming

A bit of googling and wikipedia searching has pointed me in the direction of a few new ranking sources. The key piece of the puzzle was the Aggr Ranking, a German website that hosts an aggregate ranking of FIFA's rankings, and some of the familiar alternatives like Elo, Ziaian, and AQB, as well as three others that are completely new to me:

The Roon BA is in English and provides a ranking, although no date, nor does it specify its method. It seems to be fairly in line with most of the more reputable lists. A women's international ranking is also available at this site.

Xocca is a German ranking and, besides the difficulty of trying to decipher country's names in the Teutonic tongue, also seems to be a respectable list. The method is provided, or at least sketched, but since my skills in German consist of looking for words I recognize from Dutch and filling in the blanks with semi-intelligent guesswork, I won't take the time to try to explain it right now.

Lastly, there is Bull's Fussball Weltrangliste, again German, but perhaps not quite as sound as the other new kids on the block. The top of the list looks OK but, even though I maybe biased, no ranking should have Canada (or should I say Kanada) at 109. Again, the method is explained, in some detail, but for now I'll leave it to my readers to figure it out.

The existence of these previously-unknown-to-me rankings leads me to believe that there are several more out there that I haven't found simply because they are published in some other language. I can do some French and Dutch googling, but beyond that I don't have the linguistic skills to find more. Any help is much appreciated.

update
: Check out the sidebar for permanent links to some of the above-mentioned rankings.

Monday, September 25, 2006

Hitting the books

These days, I'm doing a little extra-curricular reading on statistical methods with regards to ranking. So if you know of any academic studies of sports ranking systems, or any good non-sports-specific academic papers such as this one titled "On Minimum Violations Ranking in Paired Comparisons" that seem like they might be fruitful, leave me a comment with a link. If it is from a journal that isn't publicy accessible online, leave me the reference and I can probably get access to it. I'm not a statistics expert, but I learn fast so don't worry if they seem overly technical.

The idea behind all this is I've been interested for a while in developing an alternative to the FIFA rankings that might do a slightly better job. Again, I'm not a statistician, so this might seem a little presumptuous (well, it is definitely presumptuous), but it seems like fun.

And, for the heck of it, some search terms so that statisticians of the right sort will stumble upon this post: ranking, paired comparison, incomplete tournament, rank correlation, probability.

Upon further reading, the minimum violations ranking as described in the article cited above seems better suited to single-season competitions such as NCAA football. The international football calendar is ongoing and fluid and better suited to a system where points are accrued over time. It was a good read, though.

Monday, September 18, 2006

New rankings, same as the old rankings

It's been a long while since I posted on this blog, and there have already been three editions of FIFA's rankings for men's national sides, so it is probably time I share my two cents. Not only that, but it will help bump some of my ill-considered and incomplete World Cup projections off the top of this page.

I was optimistic when, earlier this year, press releases and articles began to appear indicating that FIFA recognized the flaws in its ranking procedures and had developed a new system. Before the first set of rankings were published following the World Cup finals in Germany, I posted indicating some of my reservations about the whole project.

Now it is time for me to take a swing at some of the more obvious faults from the first three editions of the rankings:
  1. They are extremely volatile. The inaugural July edition had Canada in 54th, probably a fair reflection of the country's performance and certainly a ranking which falls more in line with some of the other rating systems I refer to from time to time. Then in August, a 28 spot drop, having something to do with old results dropping out of the equation. And now, an 11-point jump on the strength(?) of a 1-0 friendly win over Jamaica in September. This was the only match for Canada since March (source).
  2. Expected victories -- that is, victories over relative minnows -- are given too much credit. One of the biggest movers and shakers in the September edition of the rankings is Cuba, who jumped 31 places from 100 to 69. The results that precipitated this jump were 6-0, 6-0 and 7-0 victories over the Turks and Caicois Islands, Bahamas, and Cayman Islands, ranked 170th, 138th and 177th respectively. Obviously the large margin of victory will have factored into the mix, but such scorelines are to be expected when such weak teams are involved.
That's all I have for now. Needless to say, in reforming their old ranking system, FIFA has not entirely solved the old problems and may have introduced some new ones.

Monday, June 12, 2006

DEF Jam

I'm a little behind on the previews, so I'll give it to you quick and dirty. I don't want to be dishonest and base my predictions on the matches already played, so I'll go with the picks I made in a pool with some friends.

Group D

1. Mexico
2. Portugal
3. Angola
4. Iran

Group E

1. Italy
2. Czech Republic
3. USA
4. Ghana

Group F

1. Brazil
2. Croatia
3. Japan
4. Australia

Japan beating Australia hurts, and I'm now inclined to believe that Iran will beat out Angola for 3rd, but that's how it is with prognostication.

An interesting, quasi-statistical note: Of 10 matches played so far, 23 goals have been scored, meaning the tournament is on pace for 147 goals, down about 15 from last time around.

Monday, June 05, 2006

Group B Preview

I don't have a snappy opening, so I'll get right into it.

England:
Rankings: 10th (FIFA), 4th (ELO), 5th (Ziaian), 4th (AQB)
Number of appearances (including 2006): 12
Best result: Champions (1966)

Paraguay:
Rankings: 33rd (FIFA), 29th (ELO), 21st (Ziaian), 22nd (AQB)
Number of appearances: 7
Best result: 2nd round (1986, 1998, 2002)

Sweden:
Rankings: 16th (FIFA), 12th (ELO), 11th (Ziaian), 27th (AQB)
Number of appearances: 11
Best result: Runners up (1958)

Trinidad and Tobago:
Rankings: 47th (FIFA), 67th (ELO), 56th (Ziaian), 67th (AQB)
Number of apperances: 1st
Best result: N/A

I don't think I am being unfair in expecting T&T's appearance at Germany 2006 to be a short one. I would not be surprised if they lose all three matches without scoring, like Canada did in 1986, although perhaps I am bitter because the Canucks failed to qualify. In any case, I'm looking for a three horse race. Paraguay is the lowest ranked of the three, while England might be suffering with the injury to Wayne Rooney, but I think it is Sweden that will falter.

Predicted order of finish:
1. England
2. Paraguay
3. Sweden
4. Trinidad & Tobago

Thursday, June 01, 2006

Group C Preview



Who said I had to go in order?

I like to keep my Dutch reading skills somewhat sharp so from time to time I head over to the Dutch Google News site to practice. This morning, there was a story about a ban on leeuwenhose ("lion pants"), a promotion by Bavaria, a Dutch beer company. These had been banned from earlier Dutch national matches so as not to offend title sponsor Heineken, but the ban was lifted for the upcoming friendlies, today (against Mexico, a 2-1 victory for the oranje), and on June 4th. The best line was found in the last paragraph:
Fans zoals de 18-jarige Martijn Willeboer stonden dan ook letterlijk in hun onderbroek te supporteren voor Robben en co. De KNVB wordt gesponsord door concurrent Heineken en daarom wil de Voetbalbond geen reclame voor Bavaria in het stadion.

For some who had not come prepared, they were forced to sit through the match in their underwear.

Now, without further ado, the preview.

Argentina:
Rankings: 9th (FIFA), 8th (ELO), 7th (Ziaian), 8th (AQB)
Key player: Juan Roman Riquelme

Netherlands:
Rankings: 3rd (FIFA), 2nd (ELO), 3rd (Ziaian), 6th (AQB)
Key player: Ruud van Nistelrooy

Serbia and Montenegro:
Rankings: 44th (FIFA), 26th (ELO), 18th (Ziaian), 25th (AQB)
Key player: Dejan Stankovic

Cote d'Ivoire:
Rankings: 33rd (FIFA), 30th (ELO), 32nd (Ziaian), 18th (AQB)
Key player: Didier Drogba

A lot of people have called this group the "group of death" or at the very least, one of the toughest groups in the tournament. And my team (Netherlands) will have to work hard to advance. But at the same time, this is a two-tiered group, with Netherlands and Argentina occupying the upper and Serbia and Cote d'Ivoire the lower. Either of these teams will likely need to beat one of the more powerful sides, and I don't expect that to happen. Argentina is always (well, usually) strong, and Holland will wait until the stakes are higher before choking.

My order of finish goes:
1. Holland
2. Argentina
3. Serbia and Montenegro
4. Cote d'Ivoire

Wednesday, May 31, 2006

Group A Preview

Host countries have a tendency to do well at World Cups. You have to look no further than Japan and, especially, Korea in 2002. So it really doesn't seem fair that, on top of the built-in home turf advantage, Germany was given an easy group. Let's break it down.

Germany:
Rankings: 19th (FIFA), 11th (ELO), 10th (Ziaian), 13th (mondfoot)
Country facts: Population, 82,422,299. National holiday, Unity Day, 3 October (1990).

Poland:
Rankings: 29th (FIFA), 22nd (ELO), 15th (Ziaian), 19th (mondfoot)
Country facts: Slightly smaller than New Mexico. Gender ratio,
0.94 male(s)/female

Ecuador:
Rankings: 39th (FIFA), 45th (ELO), 30th (Ziaian), 73rd (mondfoot)
Country facts: Gained independence 24 May 1822 (from Spain). Territorial rights to the Galapagos islands.

Costa Rica:
Rankings: 26th (FIFA), 38th (ELO), 38th (Ziaian), 67th (mondfoot)
Country facts: Has land boundaries with Nicaragua (309 km), and Panama (330 km). Has over 100 volcanic cones, including several major volcanoes.

All that this shows, apart from the total out-to-lunch-ness of the mondfoot rankings, is that after Germany this group is fairly wide open. The hosts rank is lower than it ought actually to be, simply because they haven't played meaningful matches recently. This is due to the fact that, as hosts, they earn an automatic bid to the final, and haven't played any qualifying matches, which are weighted more heavily in most calculations. I see an opening match win against Costa Rica, and results against Poland and Ecuador , to take first in the group. The Costa Ricans have had some turmoil recently, with a shake-up in their selection, so I don't expect much from them. For 2nd, it comes down to Poland and Ecuador, and I will side with the more tested (and higher-ranked) Poles, who should also have excellent fan support.

In case you didn't catch that:
  1. Germany
  2. Poland
  3. Ecuador
  4. Costa Rica

Tuesday, May 30, 2006

Bracketology

Here is a handy World Cup bracket (Microsoft Excel format, 14KB). I played around with it in OpenOffice, so I hope it opens fine.

I ended up with a Brazil-Czech Republic final, but that seems all too predictable considering the rankings, so I'll try again.

Expect an extensive preview post sometime before the festivities begin June 9th.

Thursday, April 20, 2006

April fools

It seems I don't have much to post about these days other than the release each month of FIFA's rankings. The top of the April edition is stable, with Brazil, the Czech Republic and the Netherlands continuing to occupy the top spots; the "stans" are the most upwardly mobile countries this month with Tajikistan jumping 16 spots to number 125 and (can-I-buy-a-vowel) Kyrgyzstan moving up 13 to occupy the 146th spot.

The top 20 looks like this:

1 Brazil 0 830 -5
2 Czech Republic 0 779 -10
3 Netherlands 0 774 -14
4 USA 1 760 -4
5 Spain 1 759 -4
6 Mexico 1 758 -4
7 France 1 754 -6
8 Portugal 2 753 -2
8 Argentina -4 753 -12
10 England -1 744 -12
11 Denmark 3 738 2
12 Nigeria 0 737 -1
13 Turkey -2 733 -10
14 Italy -2 731 -7
15 Cameroon 0 723 -3
16 Sweden 0 717 -8
17 Japan 1 709 -3
18 Egypt -1 708 -6
19 Greece 2 699 -5
19 Germany 3 699 -1

Germany, understandably, gained 3 spots for a 4-1 thrashing of the U.S., and Denmark also jumped 3 spots among these teams. The Americans, inexplicably, were rewarded for their tails-between-their-legs performance in Dortmund, with a move into 4th place.

Here are the results the U.S. have accumulated since the World Cup in 2002 (rank of opponent is current rank):

Brazil (1): 2 losses
Netherlands (3): 1 loss
Mexico (6): 2 wins, 1 loss, 1 draw
Argentina (8): 1 loss
England (10): 1 loss
Denmark (11): 1 loss
Turkey (13): 1 loss
Cameroon (15): 1 loss
Japan (17): 1 loss
Germany (19): 1 loss
Costa Rica (26): 2 wins, 1 loss, 1 draw
Colombia (27): 1 win
Poland (28): 2 wins, 1 draw
Paraguay (33): 1 win
Norway (40): 1 win
Honduras (41): 3 wins
Jamaica (44): 2 wins, 3 draws
Trinidad (47): 2 wins
Guatemala (61): 2 wins, 1 draw
Scotland (62): 1 draw
Wales (74): 1 win
Cuba (79): 2 wins
Panama (81): 3 wins, 2 draws
Canada (84): 2 wins, 1 draw
Haiti (104): 1 draw
New Zealand (120): 1 win
El Salvador (128): 4 wins
Grenada (154): 2 wins

The Americans, as should be expected from the region's top team, have done well against CONCACAF opponents (26 wins, 2 losses, 10 draws), but apart from good results against another paper tiger, Mexico, haven't performed well against FIFA's top 20.

The Americans' continuing climb in the face of some mediocre results could be due to the following stipulation in FIFA's ranking procedures, which I had previously not noticed: "For each team only the seven best results per year are given full weighting." I can't begin to understand why this is the case. Knocking off of the worst (and best, perhaps) result in a given year might be a useful way to disregard statistical outliers, but to disregard several poor results (only for countries who play enough matches, mind you), is downright foolhardy. I would be interested to know whether the other results are ignored, or simply not given "full weighting". Assigning a weight to matches 8+ that is somewhere between 0 and "full" would make the function of the rankings even more inscrutable.