Ok, I know this is off-topic, but I've been talking to a guy at work about Barcelona soccer and I decided to get back to running some soccer stats. Here's a link to the new stats site, at EPLtables.com. Let me know what you think.

In our NHL analysis, we spend a lot of time looking at shot ratios (Corsi, Fenwick, etc...) as a proxy for puck possession or offensive zone time. These are fairly consistent quantities from year-to-year, and teams retain nearly 60% of their performance:

Compare that to ball possession in soccer:

EPL teams have virtually zero year-to-year variation in ball possession. If you know how a team did last year, then you've got a pretty good idea of how they'll do next year. There are other elements - goaltender talent, shooting talent, and a whole lot of luck - but teams are essentially stuck where they are. Here are teams that been in the EPL over the last four seasons:

Obviously there are structural reasons why Chelsea, Man United, Arsenal and Liverpool dominate play, but most teams are trapped in a very small range of performance regardless of year-to-year changes they make to their teams. It's unfortunate that we don't have passing and possession data for the NHL - we could see if this is an equivalently-persistent skill in hockey. It seems unlikely, but having pass information as opposed to just shots as we do today would leave us much less at the mercy of luck.

In terms of ball possession when the game is tied, Everton is 4th in the English Premier League. But somehow, they're in last place in the table, with zero wins in six games. What's going on?

In five of their six games thus far, Everton has, for the most part, controlled the play when the game was tied, which has given them a huge lead over their opponents in the shots department. With average luck, they'd probably be 5th in the table today. But they gave up two goals on three shots against Blackburn and Aston Villa, which made it fairly difficult for them to come away with many points in those two games. At the same time, they dominated Wolves, Man United and Fulham and didn't win any of those games.

Make no mistake - Everton is a good team. You're welcome to think that Tim Howard will keep giving up goals at this rate, but I'd bet against it. At the same time, their next three games are against tough opponents, and it isn't until October 30 that they get an opportunity to absolutely dominate a team (Stoke). Despite their dominance, they could very easily find themselves going into that game 0-4-5.

Update:

Here are the league-wide possession numbers through Saturday's games. Everton isn't quite as good as Chelsea, but it's close:

Nine of the weekend's first ten games have been played, with only one upset so far (Everton losing to Blackburn Rovers.)

Territory-leading teams took 16 points, while losing teams took 7.

Here's the main reason why I stress the importance of controlling territory in soccer. Territory percentage is essentially a measure of where a team has the ball - it takes the number of completed passes and weights them by the difficulty of completing each pass. The plot below is for the last four seasons of the English Premier League. I've taken the road team's territory percentage in every game while the score was tied and plotted it against their winning percentage - we could have looked at this as a logistic regression, but I'm showing it as bucketed data.

If we look at this data at the season level, we see a very strong correlation between territory percentage and winning percentage:

Of course, a lot of that comes from Arsenal, Chelsea, Liverpool and Manchester United having been that much better than the rest of the pack over the last four seasons. If we remove them from the sample and look only at games between the other sixteen teams in the Premier League, the r^2 at the season level drops precipitously. What doesn't change is performance at the game level: teams that control passing and territory when the game is tied also end up winning.

One thing is for certain: controlling territory (as you might have suspected) is a true talent at the team level. Here I'm showing a team's territory percentage in even and odd minutes in tied road games:

This is essentially a perfect fit. There's no indication that passing and controlling territory - when the game is tied - has a major luck component to it. There are other ways to verify that, and I'll discuss that in the future. Next time, I'll take a look at how a team's ability to control territory impacts performance in head-to-head matchups.

You'll have to take my word for it for the moment, but one of the best indicators of a team's ability is the percentage of territory that it owned on the road with the score tied. What do I mean by 'percentage of territory'? Essentially, this is a measure of both ball possession and location during a game. The more passes you complete into dangerous areas when the game is tied, the better your team is.

Here's the 2009-10 table for the EPL, ranked by road territory%:

I'll delve more into the details of these advanced metrics in the coming days...

Overall, there were 270 cards handed out over 64 games:

That's 4.08 yellows and 0.14 reds per game, which is quite a bit higher than the EPL, for example.

Spain displayed a surprising level of dominance throughout the World Cup. Whereas other teams might periodically not dominate possession from the moment the game started, the Spaniards never blinked.

It seems surprising that they were less dominant during group play, putting up middling numbers (for them) against the Chileans and a weak Honduran side. But once they got up against good opponents, they made sure to control the play.

One sign of a team being way better than its opponents is its ability to control possession even when they have the lead. Normally, a team will play defensively when it's up a goal and concede the ball in order to prevent higher-quality scoring chances. Spain never really gave the ball up when it was beating most of its opponents:

It wasn't until they came up against the Germans that they actually needed to back off a bit. Despite their loss to the Swiss and a few shaky moments against Paraguay, this looks like the best team's textbook path to winning the World Cup.

If anyone could have challenged Spain, it would have been Argentina or Brazil, but each of them eliminated themselves long ago with a bad game. The Dutch and the Germans simply weren't up to the task.