Goalkeeping Numbers; or, How Good is Kasey Keller?

But he seems like a pretty solid option (I thought there was a decent case for calling him the Sounders’ man of the match last Saturday), and Boss is going to get playing time when the schedule picks up after the break. (I don’t think that the Sounders really meant for Keller to get as many USOC starts as he did last year, but our backup keepers kept getting injured.)

I know keeper isn’t the most physically demanding position, but at Keller’s age, I doubt it’s ideal to have him starting twice a week very often. CCL play could be interesting—do the Sounders maybe start Keller on the road in CCL play because he has experience playing in those countries from WCQ and then start Boss in MLS and/or USOC games? It won’t be something we’ll see much of if we don’t get past Metapan, but DCU got past Firpo last year, so I’d say we certainly stand a chance against Metapan.

Keller’s biggest advantage would seem to be his experience organizing the defense in the back. I don’t know enough to say whether that’s been going well or not, but our overall defensive numbers have still been pretty good considering that we’ve had a few bad giveaways in our defensive third.

…is to adjust it for score. I’m sure that scoring rates are different (probably significantly different) depending on the score and the time of the game. As dominant has Ricketts has been—and I have been very, very impressed by him—LA has also been giving him a lot of leads, and with the lead, a team can play more conservatively on defense, and by playing more conservatively on defense, the keeper will probably not see as many high quality shots on goal.

At one point I was looking at stats on MLS’ old site where they had goal scoring broken up into 15-minute chunks of the game. If you look at the league as a whole, scoring is lowest in the first 15 minutes, it increases until halftime, then decreases right after halftime, and increases to its peak in the last 15 minutes of the game. Which makes sense to me—I think that in general defenders probably tire a little more than strikers, and at the end of the game, it’s more likely that one team is going to be pushing harder for goals than the other, which should increase their scoring some and also increase the opportunities for counterattacks.

Anyway, I don’t have a decent database of goal scoring to do the heavy lifting on that, but the information is all out there—all you need to know is when each goal is scored during each game. From there, you could get even fancier if you wanted by adjusting for number of players on each side, but that’s probably a smaller effect on the overall GAA since it doesn’t often happen that a team players a man down.

…I’d like to see the figures for goals allowed during the run of play. It seems to me that an unusually-large percentage of the goals scored on Keller have taken place during set-pieces where one of our defenders totally failed to mark his opponent, and there was little the keeper could do. For run-of-play numbers, and with the exception of the total flub on L.A.’s first goal, Kasey has seemed to me quite solid.