International Soccer Ratings

[Huckleberry]

Coming soon. An objective ratings set of international soccer squads. The ratings set will consist of matches from the last 6 years. Ratings will be based on the following:

1.) Win, loss, or draw
2.) Margin of win or loss
3.) Location of match
4.) Time since match was played
5.) Competition match belonged to

If you'll recall - or, rather, even though you don't recall - it was almost a year ago that I published a European Club Ratings set. Nobody volunteered to help me keep up with score gathering, so I said f it. This should be easier to stay on top of thanks to the fact that somebody else already does it and publishes the scores.

No whining when the USA mysteriously appears in the #1 position.

[Airhog]

The only problem with a rating like this is that in friendlies a squad may not play their best. Just an observation, I know that there it not really any way to control that though.

[Huckleberry]

Correct. Friendlies are weighted lower than all other matches. And, of course, only matches published as being between the "A" sides from both nations are used.

[ice4277]

One thing I do not notice is a 'strength of schedule' indicator, which I think would be extremely useful. The competition for the match is important, but a 2-1 win over England in WC qualifying should count more than a 3-0 win over Andorra IMO.

[Huckleberry]

Yeah, I didn't explicitly list that, but these ratings are based on an MLE algorithm. The link in my signature will show you other ratings sets. Hopefully that can convince you that who the result came against is clearly included.

[Huckleberry]

Here is the first release.

I just plugged some numbers in for competition weighting and also haven't really studied the effects of some of my game scoring choices. Here are some details on what I used to start out:

Competition Weights:

World Cup - 100
A Tournament - 95
B Tournament - 90
C Tournament - 80
Friendly - 70

Note that qualification stages for a competition are considered as fully part of the competition.

Game Scoring:

As with many ratings systems, there is basically one point up for grabs in each match. Here's how they break down right now:

Neutral site win ranges from .75 to 1.0 points based on the score.
Neutral site draw is .5 points for each team.
Neutral site loss ranges from 0 to .25 points.
Home win ranges from .67 to 1.0 points.
Home draw is worth .44 points.
Home loss ranges form 0 to .17 points.
Road win ranges from .83 to 1.0 points.
Road draw is worth .56 points.
Road loss ranges from 0 to .33 points.

The time weighting factor is based on the following graph. The x-axis is the week of the competition starting with week 1 through week 314 (6 years):



Remember that this whole system is one day's work. I decided to do it this morning. I only picked 6 years because that's what was readily available. Perhaps a different time frame would work better. I'll be taking a look at the weights I used and any ideas are welcome.

Also note that now non-existent nations are on there. It will be easier to just let them fade away once they don't have any matches left within the chosen time frame. Yugoslavia, for example. It's still interesting, IMO, to see where they are. What will happen, though, is that they will move toward the average team in the ratings set as their matches' time weights move to 0.

[ice4277]

Originally Posted by Huckleberry

Yeah, I didn't explicitly list that, but these ratings are based on an MLE algorithm. The link in my signature will show you other ratings sets. Hopefully that can convince you that who the result came against is clearly included.

Good deal. I hadn't checked your links and other polls out when I posted earlier

[MIJB#19]

Just my two cents: eliminate friendlies entirely. It's basically pointless to factor those games in. It's just as smart as using NFL pre-season games in a power ranking system.

[Huckleberry]

Originally Posted by MIJB#19

Just my two cents: eliminate friendlies entirely. It's basically pointless to factor those games in. It's just as smart as using NFL pre-season games in a power ranking system.

For comparison, here are two rankings sets. The first is with the above weights, the second is with:

WC	100
A Tournament	80
B Tournament	60
C Tournament	50

and friendlies not used:

Rank	Team	Rating
1	Netherlands	90.60
2	France	90.28
3	Argentina	89.98
4	Brazil	88.64
5	Czech Republic	88.41
6	England	88.33
7	Ireland	88.08
8	Spain	88.00
9	Italy	87.74
10	Germany	87.25
11	Portugal	86.43
12	Denmark	86.20
13	Sweden	85.00
14	Turkey	83.74
15	Croatia	83.57
16	Belgium	83.44
17	United States	82.98
18	Colombia	82.93
19	Cameroon	82.85
20	Romania	82.13
21	Russia	81.71
22	Yugoslavia	81.66
23	Uruguay	81.66
24	Nigeria	81.40
25	Mexico	81.06

Rank	Team	Rating
1	Germany	93.75
2	Argentina	91.84
3	England	91.62
4	Italy	91.16
5	Spain	91.10
6	France	90.97
7	Netherlands	90.59
8	Brazil	90.53
9	Ireland	90.00
10	Portugal	89.37
11	Turkey	88.66
12	Sweden	87.31
13	Czech Republic	86.99
14	Yugoslavia	86.66
15	United States	86.02
16	Costa Rica	85.68
17	Denmark	85.60
18	Croatia	85.32
19	Mexico	84.83
20	Russia	83.82
21	Belgium	83.71
22	Colombia	83.49
23	Cameroon	83.47
24	Japan	82.80
25	Senegal	82.19

[MIJB#19]

Interesting change of rankings. Germany moving up to first is puzzling to me, but if your numbers support it, then it must be realistic.

[ice4277]

Perhaps there is such a big discrepancy with Germany's rankings since they did not have to go through the World Cup qualifiers.

[MIJB#19]

Maybe, but Germany really, really, sucked in Euro 2004 qualification and the final round.

[Huckleberry]

With no friendlies Germany obviously has a limited number of games included. The Confederations Cup was more recent than Euro 2004 and they performed better at that competition, of course. Even when home games are taken into account.

And here are the rankings I think I'll probably settle on. Weights are 100, 90, 75, 60, 30:

1	Netherlands	98.997
2	Argentina	98.982
3	France	98.233
4	England	98.074
5	Brazil	97.574
6	Spain	97.514
7	Italy	97.434
8	Ireland	96.810
9	Czech Republic	96.478
10	Germany	96.332
11	Portugal	95.219
12	Denmark	94.870
13	Sweden	94.848
14	Turkey	94.563
15	Croatia	93.199
16	Cameroon	92.936
17	United States	92.909
18	Belgium	92.868
19	Yugoslavia	92.464
20	Colombia	92.137
21	Russia	91.927
22	Mexico	91.340
23	Nigeria	91.068
24	Romania	91.061
25	Costa Rica	90.891

[AlexB]

This is an excellent example of how no-one knows how good the CONCACAF two and African countries actually are, as generally they don't play the strongest opposition. So while Americans see themselves beating the hghly ranked Mexicans (who themselves see Mexico ranked as one of the best teams in the world) and the Europeans look at USA/Mexico as decent but not great sides, the fact is there is little to compare to CONCACAF/Africa to Europe/S.America.

From a European point of view your rankings look generally better than FIFA's (Although I'm surprised Brazil are so low) but I imagine North Americans will see this as inaccurate. It'll be interesting to see in June where USA/Mexico fit into world football...

[Mr. Wednesday]

Originally Posted by Huckleberry

For comparison, here are two rankings sets. The first is with the above weights, the second is with:

Just what are "A", "B", and "C" tournaments? If, e.g., you classify the Gold Cup as a "C" tournament and the European Championships as an "A" tournament, that's a huge handicap to CONCACAF teams, without necessarily sufficient justification for it.

[Icy]

6 Spain
I hope the gods of football can hear you. My country is always high rated, the Spanish La liga is probably the best clubs competion in the world (along with Italy and UK), and then when we play in a big international teams competition... we suck. Probably our big stars care way more about their club who pays them that about the national team, but it's really frustrating.

[Huckleberry]

Originally Posted by Mr. Wednesday

Just what are "A", "B", and "C" tournaments? If, e.g., you classify the Gold Cup as a "C" tournament and the European Championships as an "A" tournament, that's a huge handicap to CONCACAF teams, without necessarily sufficient justification for it.

I think there's a misunderstanding regarding the weights. Even if I had done what you hypothesize above, it wouldn't have the effect you're thinking it would. The weights tell the system how certain it can be that the result indicates true strength.

For example, Brazil beating Germany 2-0 in the World Cup tells the system that it can be X amount certain that Brazil is better than Germany. If Brazil were to beat Germany 2-0 in a friendly, I have allowed the system to be .3X amount certain that Brazil is better than Germany.

The results of the CONCACAF Gold Cup have no bearing on the CONCACAF teams' placement in the rankings relative to UEFA teams. Unless a UEFA team participates.

Here are some A, B, and C tournaments:

A Tournaments:
European Championship
African Nations' Cup
Asian Cup
CONCACAF Gold Cup
Confederations Cup
Copa America
Oceania Cup

B Tournaments:
Arab Cup
Asian Games
CECAFA Cup
CEMAC Cup
COSAFA Cup
Tiger Cup
U.S. Cup

C Tournaments:
Amilcar Cabral Cup
Cyprus Tournament
Iran Tournament
Kirin Cup
Merdeka Tournament
Tournoi de France

[KeyserSoze]

Originally Posted by Icy

6 Spain
I hope the gods of football can hear you. My country is always high rated, the Spanish La liga is probably the best clubs competion in the world (along with Italy and UK), and then when we play in a big international teams competition... we suck. Probably our big stars care way more about their club who pays them that about the national team, but it's really frustrating.

The problem is that we have not really big stars. In the Real are Ronaldo, Zidane, Beckham. In the Barsa the stars are Ronaldinho, Etoo Messi. The best national players of these two teams (Raul and Xavi) have long term injuries, so they will suffer to be 100% at Germany.

We have problems creating good situations of goal (Xavi is great, but the others...) and we have problems with the Strikers (Morientes, Villa...)

I think we have a good team but no a great team.
Great teams can be German (just they play home), Netherlands has a good group too, France, England. For me Spain is a step behind. Of course in 1 game competition all can happens (ask Greece) but I don't bet in Spain.

And then it´s Brazil. Ronaldo, Ronaldinho, Adriano, Kaka, Emerson, Ze Roberto, Juninho Pernambucano...... Hell. They have 2 great teams

[Icy]

Originally Posted by KeyserSoze

The problem is that we have not really big stars. In the Real are Ronaldo, Zidane, Beckham. In the Barsa the stars are Ronaldinho, Etoo Messi. The best national players of these two teams (Raul and Xavi) have long term injuries, so they will suffer to be 100% at Germany.

We have problems creating good situations of goal (Xavi is great, but the others...) and we have problems with the Strikers (Morientes, Villa...)

I think we have a good team but no a great team.
Great teams can be German (just they play home), Netherlands has a good group too, France, England. For me Spain is a step behind. Of course in 1 game competition all can happens (ask Greece) but I don't bet in Spain.

And then it´s Brazil. Ronaldo, Ronaldinho, Adriano, Kaka, Emerson, Ze Roberto, Juninho Pernambucano...... Hell. They have 2 great teams

Agree, as the best international stars come to play in our league, that is closing the doors to the Spanish young players and damaging our future.
Btw, how is going life? long time we don't chat

[Desnudo]

Originally Posted by Mr. Wednesday

Just what are "A", "B", and "C" tournaments? If, e.g., you classify the Gold Cup as a "C" tournament and the European Championships as an "A" tournament, that's a huge handicap to CONCACAF teams, without necessarily sufficient justification for it.

You consider the Gold cup as an "A?" it's pretty much a tune-up for teams.

[Mr. Wednesday]

While Arena does use it for player evaluation, he's also on record as considering it to be an important tournament for the U.S. to win, being our continental championship.