This is probably such an old thread that you will miss it, but Beane is trying his hand at statistical analysis of soccer while working for Lew Wolf’s other team, the San Jose Earthquakes.

Linked the original source material at the link above as well as some initial thoughts.

One-touch epée fencing
Air pistol target shooting
Show jumping
Swimming
Cross-country running

Now that’s an awesome sport.

(newly minted fencing geek)

See, if I were more statistically inclined, I could have written that sentence several posts ago, and saved everyone a lot of headaches. hahaha. That’s what I’m trying to get at, really.

Getting worked over by King Felix for a strikeout is not the same as striking out against Horacio Ramirez. A strikeout is not “purely” the result of any one player’s actions. And on a ball in play, there’s no reason to give all the credit to the fielders–Ichiro is going to have a different BABIP than Richie Sexson in the long run. So even there, you have to split credit between the hitter, the fielders, and the pitcher.

But see, this is exactly what I’m trying (apparently without success) to argue – however large or small those effects may be, I don’t see that they necessarily have predictive value in soccer, the way they would and do in baseball.

They don’t necessarily have predictive value. My argument is simply that they could and you don’t know until you check. Define some sort of passing percentage, and hire some college interns to keep track of it. Then divide the games up into two subsets (odds and evens or something like that) and check to see how passing percentage in one set correlates to passing percentage in the other set. Maybe it’s awful, but it’s not like it’s impossible to find out if you have the resources.

The issue, though, is that you’re not going to get the kind of differential data that allows detailed situational analysis by person… it’s a thorny problem.

I still think you could get somewhere with even the passing issue, though.

I can give you reams of anecdotal evidence, but I can tell by your tone that nothing I give you will satisfy you. But here’s one quick example – you could have a guy that’s the best passer on your team, and he could have a crap day, or even a bad stretch of games, because all of his normally right-on-the-money passes are not gathered by their intended recipient.

Is that the fault of the passer, or is that the fault of the recipient, who couldn’t get to where he needed to be, or is it to the credit of the defender, who stopped the recipient from getting somewhere? In baseball, if a batter strikes out, it’s purely his fault, and if a batter makes an out on a ball in play, it’s credited to the person who makes the out.

In soccer, though, how would it be possible to ascribe responsibility? Again, an anecdote. A midfielder has the ball. HE surveys the field, and he sees one of the forwards making a run in a specific direction, and in his head, Mr. Midfield Maestro says “OK, forward, you’re going over there, I’ll put the ball there”.

But then, in that same split second, say that Mr. Forward decides to take a different route, due to some perceived or real obstruction. Mr. Midfielder passes the ball as intended, but Mr. Forward wasn’t there – whose fault is that? The forward, for not being able to read the midfielder’s mind, or the midfielder, for not being able to predict the forward’s intentions?

These types of things happen dozens of times a game, which is why I think it’d be nearly impossible to quantify. Again, doesn’t mean it’s not worth trying, but I’m not sure what it would tell you.

But you don’t know how large those effects are until you look at the numbers.

But see, this is exactly what I’m trying (apparently without success) to argue – however large or small those effects may be, I don’t see that they necessarily have predictive value in soccer, the way they would and do in baseball.