World Soccer XI: To South Africa, and Beyond!

[Badger Abroad]

Re: World Soccer XI: To South Africa, and Beyond!

USA got a very favorable draw today.

England
USA
Algeria
Slovenia

England is obviously very tough and I am a bit nervous about Slovenia, as the U.S. does not usually have much success against teams from central Europe (see the results against the Czech Republic and Poland in the last two World Cups). I have no idea about what kind of a side Algeria is.

 

[J.D.]

Re: World Soccer XI: To South Africa, and Beyond!

Looks like a pretty fair and balanced draw to me. It's never gonna be perfect. If you go by FIFA rankings, G is the toughest, F the easiest.

If Quarterfinals go to form:

Spain
Italy

Germany
Argentina

France
England

Netherlands
Brazil

 

[JohnN]

Re: World Soccer XI: To South Africa, and Beyond!

Nate Silver of 538 and Baseball Prospectus, with statistical analysis of the draw:

http://soccernet.espn.go.com/world-...r-considers-impact-draw?ver=us&ver=us&cc=5901

I enjoy any thoughtful soccer analysis, so I enjoyed the article, but it also got me a little fired up. I'm not a big SPI fan and I wasn't too impressed with the accuracy of Nate Silver's World Cup article. I did, however, enjoy reading it.

My two biggest problems with what he does is what I believe is an over reliance on friendly results and an over reliance on player's league performance.

Regarding friendly results, I'm glad he is trying to weigh games with a relevancy factor, but unfortunately the relevance of most international friendlies is almost zero. For one thing, I think the home field advantage for friendlies is probably four times what it is for other games. Also, accounting for missing players is almost impossible. Extracting anything from a Rooney-less England game or a Donovon-less American game is virtually pointless.

As for rating players on league performance and translating it into international performance, the problems are pretty obvious. Again, I applaud him for attempting the breakdown and I enjoy the analysis, but it's all pretty pointless. This is an obscure example, but it's a really good one just off the top of my head. Asamoah Gyan of Ghana/Rennes(France) is a decent striker. He doesn't get much support in the French league and puts up stats that are a little above average. Nothing that's going to register on SPI both because of the numbers and the quality of the French league. But, the guy can put the ball in the net big time. This summer with Essien, Appiah and Muntari feeding him from midfield he's probably going to be one of the top ten goal scorers. Ghana's expectations this summer don't depend on whether he plays in France or in England, but their SPI does. Ghana's SPI would be higher if he played at Arsenal, but whether he plays for Rennes or Arsenal doesn't matter, he'll be very effective for Ghana this Summer.

Sorry if this was all kind of rambling. I really do appreciate the effort and the methodology Silver employs. I just don't think it captures the essence of the game very well. Ninety percent of a players statistical performance depends on who the other ten guys on his team are. You'll never be able to put that in a computer. Joe Mauer would pretty much have the same batting average regardless where he played in the American League, but Wayne Rooney's number would fluctuate greatly depending on who he spent a Premier League season with. His talent level wouldn't change, but his stats sure would.

 

[Jimjamesak]

Re: World Soccer XI: To South Africa, and Beyond!

USA got a very favorable draw today.

England
USA
Algeria
Slovenia

Immediate thought after watching Algeria drawn into Group C:

"Well we might as well draw France now. I mean England, USA, Algeria, and France? I'd call it "Colony Wars: 2010!"

 

[realet]

Re: World Soccer XI: To South Africa, and Beyond!

England is obviously very tough

Which was going to be true of any Pot 1 team we got drawn against unless we lucked out and got Bafana Bafana.

and I am a bit nervous about Slovenia, as the U.S. does not usually have much success against teams from central Europe (see the results against the Czech Republic and Poland in the last two World Cups).

Looking at the rest of that pot, we could have done a lot worse than Slovenia. I'll take it.

I have no idea about what kind of a side Algeria is.

They were supposed to be the bonus of their pot. Again, we could have done much worse.

In fact, even with England being a tough team, we still could have done a lot worse there too. We'll probably still lose to them, but it feels like there's more of an inkling of possibility against them - not nearly much as the South Africans, but more than, say, Brazil.

By the way, Craig, I want my propers for getting the schedule right. Not much travel if you want to go to these games, although that first fixture is going to be a beast to get tickets to since it's between the two biggest ticket-buying nations in the smallest possible venue. Wonder if they might try to fudge that somehow (doubt they can). Still, Rustenburg, Johannesburg, and Pretoria are all right near each other. Most likely destination for the Round of 16 would be Bloemfontein and that's not terribly far away either.

Also: ironic, since I was hoping for A... I like our draw better than Mexico's.

 

[Craig P.]

Re: World Soccer XI: To South Africa, and Beyond!

By the way, Craig, I want my propers for getting the schedule right.

All I would have claimed is that we were getting drawn into a position, and it wouldn't necessarily be 2. You had a 1/3 chance of being right.

 

[AFHockeyFan]

Re: World Soccer XI: To South Africa, and Beyond!

I'm pleasantly surprised by the draw, both for the US and the rest of the field. It appears fairly even.

I predict we beat Algeria, tie Slovenia and lose to England. As long as England takes care of Slovenia, we should advance in a 9-4-4-0 group depending on goal differential. All that gets us in a loss to Germany in the round of 16, but that's about the best that can be expected given where the US team is right now. I would be happy with those results.

 

[blockski]

Re: World Soccer XI: To South Africa, and Beyond!

Apparently the US is leading in fans buying tickets to the World Cup followed by England.

http://sportsillustrated.cnn.com/2009/soccer/12/04/wcup.fans/index.html?eref=sihp

And coincidentally, England-USA is in the competition's smallest venue with only 42,000.

Interesting, considering ESPN had an article that the US delegation for the 2018/2022 bid is promising a new record in attendance thinking 5 million is likely (1994 had 3.9 million)

Remember that the World Cup was only 24 teams back in 1994 as well.

24 teams = 52 total matches. 32 teams = 64 total matches.

The US averaged almost 69,000 in attendance per game in 1994. If you apply that average to 64 total matches, you still only get to 4.4 million in total attendance, so to hit 5 million, you'd need to really cram some of the larger venues.

I enjoy any thoughtful soccer analysis, so I enjoyed the article, but it also got me a little fired up. I'm not a big SPI fan and I wasn't too impressed with the accuracy of Nate Silver's World Cup article. I did, however, enjoy reading it.

My two biggest problems with what he does is what I believe is an over reliance on friendly results and an over reliance on player's league performance.

Regarding friendly results, I'm glad he is trying to weigh games with a relevancy factor, but unfortunately the relevance of most international friendlies is almost zero. For one thing, I think the home field advantage for friendlies is probably four times what it is for other games. Also, accounting for missing players is almost impossible. Extracting anything from a Rooney-less England game or a Donovon-less American game is virtually pointless.

As for rating players on league performance and translating it into international performance, the problems are pretty obvious. Again, I applaud him for attempting the breakdown and I enjoy the analysis, but it's all pretty pointless. This is an obscure example, but it's a really good one just off the top of my head. Asamoah Gyan of Ghana/Rennes(France) is a decent striker. He doesn't get much support in the French league and puts up stats that are a little above average. Nothing that's going to register on SPI both because of the numbers and the quality of the French league. But, the guy can put the ball in the net big time. This summer with Essien, Appiah and Muntari feeding him from midfield he's probably going to be one of the top ten goal scorers. Ghana's expectations this summer don't depend on whether he plays in France or in England, but their SPI does. Ghana's SPI would be higher if he played at Arsenal, but whether he plays for Rennes or Arsenal doesn't matter, he'll be very effective for Ghana this Summer.

Sorry if this was all kind of rambling. I really do appreciate the effort and the methodology Silver employs. I just don't think it captures the essence of the game very well. Ninety percent of a players statistical performance depends on who the other ten guys on his team are. You'll never be able to put that in a computer. Joe Mauer would pretty much have the same batting average regardless where he played in the American League, but Wayne Rooney's number would fluctuate greatly depending on who he spent a Premier League season with. His talent level wouldn't change, but his stats sure would.

There are going to be issues with any quantitative methodology. If this doesn't capture the essence of the game very well, then every other methodology must just be throwing darts at a board, no?

What would be better, however? From the perspective of someone like Silver who does quantitative analysis, I'm sure he'd be the first to acknowledge the shortcomings of this kind of analysis - but that doesn't make it completely invalid - it just is a limitation.

Again, just saying it's not totally accurate isn't much of a criticism - what data driven, quantitative analysis would be better?

 

[JohnN]

Re: World Soccer XI: To South Africa, and Beyond!

There are going to be issues with any quantitative methodology. If this doesn't capture the essence of the game very well, then every other methodology must just be throwing darts at a board, no?

What would be better, however? From the perspective of someone like Silver who does quantitative analysis, I'm sure he'd be the first to acknowledge the shortcomings of this kind of analysis - but that doesn't make it completely invalid - it just is a limitation.

Again, just saying it's not totally accurate isn't much of a criticism - what data driven, quantitative analysis would be better?

Fair points. Statistical comparisons are always difficult, but with a sport like soccer that is so dependent on player relationships, they are even more difficult, probably impossible.

My point wasn't that the system has flaws. My point was that as an "improvement" over other systems, I think it fails.

 

[Patman]

Re: World Soccer XI: To South Africa, and Beyond!

From the perspective of someone like Silver who does quantitative analysis, I'm sure he'd be the first to acknowledge the shortcomings of this kind of analysis

I'm going to go giggle in the corner for a minute, I'll be right back.

 

[bottomdweller]

Re: World Soccer XI: To South Africa, and Beyond!

Remember that the World Cup was only 24 teams back in 1994 as well.

24 teams = 52 total matches. 32 teams = 64 total matches.

The US averaged almost 69,000 in attendance per game in 1994. If you apply that average to 64 total matches, you still only get to 4.4 million in total attendance, so to hit 5 million, you'd need to really cram some of the larger venues.

You also have to look at the new stadiums that have been built in the US over the past 15 years. Gillette Stadium holds an extra 8500 people than Foxboro Stadium did. If you use the Cowboys new stadium instead of the Cotton Bowl you're going from 60,000 to potentially 100K plus with standing room. I don't think 5 million would be unreasonable. Plus there will be many more stadiums built in the next 10 years or so (as long as the economy rebounds to where it was).

 

[LynahFan]

Re: World Soccer XI: To South Africa, and Beyond!

I just hope Altidore, DeMerit, Dempsy, Guzan, Hahnemann, Howard, Johnson, Simek, and Spector are watching their backs over the rest of the EPL/Championship seasons!

 

[mookie1995]

Re: World Soccer XI: To South Africa, and Beyond!

quite the entertaining game here 'tween chelsea and man city. diving all over the place. save on a penalty kick late.

 

[blockski]

Re: World Soccer XI: To South Africa, and Beyond!

I'm going to go giggle in the corner for a minute, I'll be right back.

So, you think someone would actually argue that a performance ranking index is the be all and end all of soccer analysis?

I don't care who's name you put out there - anyone has to acknowledge the limits of quantitative analysis.

 

[mookie1995]

Re: World Soccer XI: To South Africa, and Beyond!

Remember that the World Cup was only 24 teams back in 1994 as well.

24 teams = 52 total matches. 32 teams = 64 total matches.

The US averaged almost 69,000 in attendance per game in 1994. If you apply that average to 64 total matches, you still only get to 4.4 million in total attendance, so to hit 5 million, you'd need to really cram some of the larger venues.

detroit - AA 107k
dallas - new cowboys stadium 100k (with standing room tickets not grass)
los angeles - rose bowl 98k
san fran - stanford stadium 85k
columbus - ohio stadium 101k
wash dc - fedex 92k
nyc - new giants stadium 82k (grass or field turf?)
knoxville - neyland 104k (throw one to the south )

 

[Patman]

Re: World Soccer XI: To South Africa, and Beyond!

So, you think someone would actually argue that a performance ranking index is the be all and end all of soccer analysis?

I don't care who's name you put out there - anyone has to acknowledge the limits of quantitative analysis.

anybody can... but will they up front? Any world soccer ranking is going to have all kinds of problems... there are about 3 or 4 of them that aren't Silver's (Jones, World ELO to name a couple). The thing is any mathematical soccer ranking is highly speculative because more so than any sport that people take interest in the teams vary in strength as opposed to what i nation CAN bring... it has a problem of being a very misleading measure. The only time we can be assured on being able to measure the truth strength of a nation's power is during the World Cup, World Cup Qualifying, and most continental championships. That's not a lot of data especially when you want to try to measure strength averaged over some close interval (a year, a month, etc.... forsaking the idea that you can adequately measure current strength as of tday).

Is anybody going to come out and say "yeah, these numbers look interesting but keep in mind that they're total horse ****". Silver doesn't strike me as somebody is going to say "this is the X, Y, and Z of why our analysis and models are flawed". I haven't been following these things closely. In general I'm incredibly skeptical as a brand especially since he's a political partisan. More so I'm a little edgy in that he has become a brand from which comes analytical truth.

 

[blockski]

Re: World Soccer XI: To South Africa, and Beyond!

anybody can... but will they up front? Any world soccer ranking is going to have all kinds of problems... there are about 3 or 4 of them that aren't Silver's (Jones, World ELO to name a couple). The thing is any mathematical soccer ranking is highly speculative because more so than any sport that people take interest in the teams vary in strength as opposed to what i nation CAN bring... it has a problem of being a very misleading measure. The only time we can be assured on being able to measure the truth strength of a nation's power is during the World Cup, World Cup Qualifying, and most continental championships. That's not a lot of data especially when you want to try to measure strength averaged over some close interval (a year, a month, etc.... forsaking the idea that you can adequately measure current strength as of tday).

Is anybody going to come out and say "yeah, these numbers look interesting but keep in mind that they're total horse ****". Silver doesn't strike me as somebody is going to say "this is the X, Y, and Z of why our analysis and models are flawed". I haven't been following these things closely. In general I'm incredibly skeptical as a brand especially since he's a political partisan. More so I'm a little edgy in that he has become a brand from which comes analytical truth.

Pardon me, but this is a load of crap.

Buyer beware. All quantitative metrics have their limits. I can guarantee that the creator of the index/ranking knows this, and the user should, too. But it's not the creator's responsibility to reiterate what the index does and what it does not every single time.

The logical conclusion of your point is that soccer is difficult to model, so we shouldn't even try. I know you deal with statistics all the time, and I can't believe you really think that.

I also don't think Silver's political analysis has any impact. It shouldn't, as Silver has been one thing consistently - open about his personal biases and opinions. He lays his methodology out there for all to see.

Quite frankly, for you to assert that he's trying to find the 'analytical truth' is just wrong - the whole point about these kinds of rankings is that there is no one truth. This isn't a mathematical proof. There is no QED here. He's ceratainly not selling his indexes as analytical truth, he's selling them as analytical tools. It's still up to the individual as to how they choose to use those tools, as each tool is different, and each has their own strengths and limitations.

 

[Craig P.]

Re: World Soccer XI: To South Africa, and Beyond!

detroit - AA 107k
dallas - new cowboys stadium 100k (with standing room tickets not grass)
los angeles - rose bowl 98k
san fran - stanford stadium 85k
columbus - ohio stadium 101k
wash dc - fedex 92k
nyc - new giants stadium 82k (grass or field turf?)
knoxville - neyland 104k (throw one to the south )

Most of the college football stadia didn't make the most recent cut. There's only one or two left. And I think even those are longshots due to field size issues (and how much it would cost to fix that).

 

[blockski]

Re: World Soccer XI: To South Africa, and Beyond!

Most of the college football stadia didn't make the most recent cut. There's only one or two left. And I think even those are longshots due to field size issues (and how much it would cost to fix that).

Plus, Stanford Stadium no longer holds 85k after the renovation. Now it's 50k.

 

[Patman]

Re: World Soccer XI: To South Africa, and Beyond!

Pardon me, but this is a load of crap.

Buyer beware. All quantitative metrics have their limits. I can guarantee that the creator of the index/ranking knows this, and the user should, too. But it's not the creator's responsibility to reiterate what the index does and what it does not every single time.

The logical conclusion of your point is that soccer is difficult to model, so we shouldn't even try. I know you deal with statistics all the time, and I can't believe you really think that.

I also don't think Silver's political analysis has any impact. It shouldn't, as Silver has been one thing consistently - open about his personal biases and opinions. He lays his methodology out there for all to see.

Quite frankly, for you to assert that he's trying to find the 'analytical truth' is just wrong - the whole point about these kinds of rankings is that there is no one truth. This isn't a mathematical proof. There is no QED here. He's ceratainly not selling his indexes as analytical truth, he's selling them as analytical tools. It's still up to the individual as to how they choose to use those tools, as each tool is different, and each has their own strengths and limitations.

most models try to seek out some form of analytic truth... otherwise what's the point? I never said you couldn't (and possibly shouldn't) try... but you have to realize the weight of these things. How much does it really indicate what's going on and is it useful? Soccer rankings, IMO, should be taken with a large grain of salt. People try to build them because like a lot of other people they want a way to objectively compare the teams... but in a lot of ways the irregular nature of international soccer means that a math ranking can mislead much more than inform on the subject.

Here's the thing. What's the value of a bad answer? Is it better than some other answer? Why do we give it weight and credence?

My thing is that many of the measures are going to be bad but we go with what we've got. The other thing to me is that there have been other methods floating around for awhile... why aren't they getting the attention?