News for hackers. Stuff that matters.
Re:England to win World Cup, says JPM quant model by nir on Tuesday May 18, 2010 @10:54PM (#1359939)
Readers from non-soccer nations: every 4 years there's a soccer world cup. A few months before every soccer world cup, UK media is abuzz with predictions that, this time, England will really go all the way.

In the world cup itself, the English barely manages to graduate the group stage when, bruised and battered, it meets a Latin American or stronger European team that promptly wipes the floor with them in the next round.

A band of heros only few weeks ago, they then return to England to have the UK tabloids wipe the floor with them once again.

That's a soccer tradition many of us look forward to seeing again next month, in South Africa.

Re:England to win World Cup, says JPM quant model by megablast on Tuesday May 18, 2010 @11:48PM (#1360032)
England has great players, possibly the strongest side altogether. However, they never manage to play together as a team, unlike a lot of other countries.

This result is good news, now I don't have to watch any of the world cup. Rather watch sports where cheating is not such an integral part of the game, and where the officiating body actually tried to discourage it rather than encourage. (from a bitter aussie)

Re:England to win World Cup, says JPM quant model by patrickk on Wednesday May 19, 2010 @08:22AM (#1360757)
Last World Cup England's "great players" like Frank Lampard and Stevie Gerrard never turned up. Their only consistently good players in that tournament were a guy who used to ply his trade in Germany (Owen Hargreaves) and Wayne Rooney of course (yes I'm a Man Utd. fan). Manchester United bought Hargreaves as a result of that showing.

The media has hyped a lot of England's "stars" into the stratosphere so that a lot of people believe the rubbish that's published about them. Every time the World Cup rolls around all England have to do is show up and they'll win it. And when they don't it's the end of the world. Just cos you invented the sport, doesn't mean you have a divine right to win it! In Europe, I would admire the likes of Spain, Portugal and the Netherlands a lot more than England as they have a lot of fine, natural footballers (not to mention they generally play better as a unit) and I'd suspect one or two of the should reach the semis at least.

ps. I hope France get annihilated in the group stages. Yes, I'm Irish. Those rotten cheaters should never have been permitted to get on a plane to South Africa. Even if you're French you shouldn't support your own country. C'mon Mexico!

Re:England to win World Cup, says JPM quant model by zppx on Wednesday May 19, 2010 @12:12AM (#1360081)
Spain had the biggest number of players on last year best player in the world election, I'm brazilian but I think Spain or the Netherlands will win im Africa
Re:England to win World Cup, says JPM quant model by Lewisham on Tuesday May 18, 2010 @11:18PM (#1359983)
You have to believe!

Like that sticker from an earlier World Cup says: "Pride. Passion. Belief."

Don't be so pessimistic!

Re:England to win World Cup, says JPM quant model by nir on Tuesday May 18, 2010 @11:51PM (#1360040)
Pessimistic? I'm looking forward to it :)
Re:England to win World Cup, says JPM quant model by joubert on Tuesday May 18, 2010 @10:13PM (#1359851)
My money is on Nate Silver's prediction of Brazil winning, with Spain in 2nd and England 3rd.

Full rankings (updated periodically): http://soccernet.espn.go.com/spi/rankings/_/view/worldcup?cc...

Methodology: http://soccernet.espn.go.com/world-cup/story/_/id/4447078/ce...

Re:England to win World Cup, says JPM quant model by zzygan on Wednesday May 19, 2010 @08:26AM (#1360765)
I think rankings in general are pretty useless overall.

Every single World Cup in the past has turned up surprises in the weirdest kind of way.

My opinion is Spain 1st, Brazil 2nd, Ivory Coast 3rd

Re:England to win World Cup, says JPM quant model by rglullis on Tuesday May 18, 2010 @10:31PM (#1359886)
Every world cup we see predictions like that. Every world cup these predictions are wrong.

The beautiful thing about football (and maybe the reason it is not that popular in the US) is that stats don't matter nearly as much as they do in baseball, american football and basketball.

Re:England to win World Cup, says JPM quant model by btmorex on Tuesday May 18, 2010 @10:45PM (#1359917)
Just curious: what evidence is there that the stats matter less?

There are certainly fewer stats altogether and the stats that exist are more team-oriented stats, but I still don't really buy your conclusion.

For a counterpoint, baseball is pretty much the king of sports in terms of statistics and it's still pretty damn hard to predict a world series winner at the end of the regular season, much less at the beginning of the season.

Re:England to win World Cup, says JPM quant model by jdrock on Tuesday May 18, 2010 @11:14PM (#1359980)
Soccer should be much harder to quantify with stats because the game is basically a set of two continuous events.

As a contrast, baseball is easier to quantify with stats because it's a series of discrete events.

In most sports statistical analysis, you try and drill down to a discrete event. In baseball, it's something like at-bat or on-base. In basketball, it's per-possession (though this isn't enough). But in soccer, what can you do? Per-possession is difficult because it overlooks the 21 other players on the field. Everything is team-oriented.

Re:England to win World Cup, says JPM quant model by mechanical_fish on Tuesday May 18, 2010 @11:48PM (#1360030)
Baseball is even better than that. It's a bunch of discrete events that are built out of many, many nigh-identical repetitions of even smaller discrete events.

The other point to make is that baseball controls the combinatorial explosion by either having the players act in relative isolation (when fielders handle the ball) or by grouping them in tiny groups of two or three (pitcher, batter, and to some extent the catcher), one of whom is generally a constant in any given game (the catcher), and one of whom only changes a few times per game (the pitcher). This is a fairly restricted set of possible matchups, and these are replayed over and over to get better stats. That's why baseball is the statistician's dream.

Similarly, in basketball there are fairly few players on the floor, and they tend to match up with each other fairly predictably. And while American football puts 22 players on the field, most of them never touch the ball, and most of the remaining players do the same sort of job over and over in the same relatively small part of the field, interacting with the same small subset of players on the other team.

But soccer is all over the place. ;)

Re:England to win World Cup, says JPM quant model by patrickk on Wednesday May 19, 2010 @08:38AM (#1360785)
In a top-class soccer match, you can't take your eyes off the action or you will miss something crucial. You could probably doze off for 20 minutes and miss little in a baseball game. Each to their own.

Just because you don't understand soccer isn't reason alone to describe it as "all over the place". It's a very exciting team sport (when played properly) and full of formations and tactics that coaches constantly try to innovate with:

The classic 4-4-2 formation

Beckenbauer's/Dutch football's 4-3-3

Jose Murinho's 4-5-1...Google em!

Also:

http://en.wikipedia.org/wiki/Formation_(football)#Common_mod...

Re:England to win World Cup, says JPM quant model by mechanical_fish on Thursday May 20, 2010 @06:57PM (#1365768)
You appear to have misinterpreted my sentence as some sort of insult. There's nothing wrong with being "all over the place"! No two minutes of soccer are exactly alike, but that is scarcely a bad thing!
Re:England to win World Cup, says JPM quant model by jubbam on Tuesday May 18, 2010 @11:50PM (#1360036)
Could it be that just the current stats collected are not enough to matter as much as the stats in baseball? Granted, it may be easier to identify stats that matter in baseball due to the discrete vs. continuous events you mention.

Maybe new stats collected at a later date will matter as much as those collected in baseball... or maybe not.

Re:England to win World Cup, says JPM quant model by JangoSteve on Wednesday May 19, 2010 @12:25AM (#1360104)
Here is the full reasoning in a nice story-format: http://www.rps.psu.edu/jun95/odds.html
Re:England to win World Cup, says JPM quant model by megablast on Tuesday May 18, 2010 @11:52PM (#1360045)
I am not sure that it is stats that make most other sports more popular than Soccer, in countries like Canada, USA, Australia, New Zealand.

For some reason, soccer nations seem to enjoy the fact that any team can win any game, due to the nature of the game. It is such a low scoring game, that one goal can decide the game. So all it takes is an incredibly lucky kick, or cheat, for a side to win.

You don't see this in many other sports.

So while the Premier league the result is certain (either Chelsea or Man United), the World cup, where in the later stages one game decides whether you move on or not, anyone has a chance.

Re:England to win World Cup, says JPM quant model by _sh on Wednesday May 19, 2010 @12:48AM (#1360139)
But this also means, from a supporter's perspective, that football is unsatisfying because there is a dearth of scoring (one lucky kick undoes 90 minutes of effort from the opposition), and this does not reflect the nature of the match. Also, draws are common, even when there is a clear imbalance between the skill/ability of the teams. I've heard it concluded that the reason football fans are seen to be more 'thuggish' (media hype aside) is that the game often does not satisfy the primal urge to defeat someone/something.

True, the 'one lucky kick' possibility also means that your team wins when it really shouldn't have. Is this really satisfying to a supporter?

(I've also heard it concluded that the glut of scoring makes basketball less appealing for similar reasons, but I haven't followed it enough to determine whether this is true).

Re:England to win World Cup, says JPM quant model by zzygan on Wednesday May 19, 2010 @08:22AM (#1360758)
I'm a huge football fan (Barcelona FWIW), so take what you will from the following comment. I watch way too many European games on late night cable, and the low scoring complaint is kind of a fallacy when you watch most games in Europe.

This season just gone, i think the average goals per game in England this year was about 2.5. Sure this average has an upward trend by games like a 7-0 thrashing that Chelsea dished out in the last game of the season. The German league has an even higher average goals per game (cant seem to find that stat at the moment..)

There are low scoring games, but that is part of the way the game works, but I definitely would not say a low score is a regular thing week in week out anymore.

As to the hooligans that truly are complete idiots, there is no other explanation for their behavior other than stupidity. Its a gang mentality that causes the problem, not the game itself. There are some epic rivalries in Europe where the fans do really hate each other, and it goes back to an identity/gang thought pattern.

Re:England to win World Cup, says JPM quant model by hitonagashi on Wednesday May 19, 2010 @06:50AM (#1360631)
It is attracting precisely because of the chance of a lucky kick. You can very rarely turn off when winning, especially when good teams play each other, because there can always be a moment of magic to tie the game.

It means as a supporter, you are always on the edge of your seat, cheering your team on, because you know what can happen if they falter.

It also means that the rare occasions where you are 3-4 0 up and the game is safe are just that much more satisfying ;).

Re:England to win World Cup, says JPM quant model by patrickk on Wednesday May 19, 2010 @08:26AM (#1360767)
It's true that the Premier League is very predictable, but the new influx of cash to the likes of Manchester City could shake things in up a few years. Look where Chelsea came from - mid-table mediocrity.

You should also point out that La Liga (the Spanish Premier League) while very easy on the eye, is possibly the most uncompetitive league in the world as only Real Madrid or Barcelona will ever win it.

Re:England to win World Cup, says JPM quant model by watty on Tuesday May 18, 2010 @10:41PM (#1359904)
I'm not disagreeing but I'm curious if you have any source that shows soccer results correlate less with "rank" than other sports. Do the low scores have something to do with it (similar to hockey maybe)?
Re:England to win World Cup, says JPM quant model by Lewisham on Tuesday May 18, 2010 @11:21PM (#1359985)
Yes. I remember reading an article on a paper that claimed football was the most "exciting" sport as it had the greatest opportunity to provide upset due to the low scoring.

It's more likely that the David's can beat Goliath. This is also compounded by having a knockout cup (where the Goliaths can and do go out) and the teams playing each other in a league far less than most American sports, at about 25 matches a year. That's roughly the same as the NFL, but nothing like baseball, which is insane.

Re:England to win World Cup, says JPM quant model by nir on Tuesday May 18, 2010 @11:58PM (#1360059)
Consider consider all the various tournaments (Cup, FIFA, UEFA etc) soccer teams play.

The NFL season feels like it's over before it began. I guess the nature of the game is such that very few players could last much longer anyway.

Re:England to win World Cup, says JPM quant model by javanix on Tuesday May 18, 2010 @10:16PM (#1359858)
Mortgage-backed securities to stay profitable, says Merill-Lynch quant model.
Re:England to win World Cup, says JPM quant model by johnrob on Tuesday May 18, 2010 @11:05PM (#1359958)
There's not nearly enough data to make decent conclusions about national teams. They don't play often enough, and when they do the rosters are never the same. To make a decent prediction you have to use the final roster (submitted to FIFA) and factor in each player's condition.
Re:England to win World Cup, says JPM quant model by JangoSteve on Wednesday May 19, 2010 @12:30AM (#1360111)
There was actually an analysis done a while back of the outcomes of every game in the last world-cup that showed statistically just how unpredictable soccer games are. The punchline is that soccer is very unpredictable, especially given the bracket setup of the World Cup.

I can't seem to track down the article now, but I sort of remember the logic...

If Team A beats Team B, then Team B beats Team C, then Team C beats Team A, then you've demonstrated at least a 33% probability of a fluke outcome in the soccer bracket. This is highly simplified of course, but repeat this analysis over a much larger sample size, and you get some pretty convincing results (and I really wish I could find that paper now).

Re:England to win World Cup, says JPM quant model by tokenadult on Tuesday May 18, 2010 @11:54PM (#1360051)
A fascinating book to read as the World Cup nears:

http://www.amazon.com/Soccernomics-Australia-Turkey-Iraq-Are...

Re:England to win World Cup, says JPM quant model by steveplace on Tuesday May 18, 2010 @10:30PM (#1359882)
"Black swans" occur quite often in sports, so statistical modeling only goes so far. One torn ACL is all it takes and it can change models around.
Re:England to win World Cup, says JPM quant model by SkyMarshal on Wednesday May 19, 2010 @12:55AM (#1360151)
These guys predict Brazil vs Serbia in the final, w/ Brazil winning. Based on GDP, population, & other demographics.

http://www.informationisbeautiful.net/2010/infographical-mor...