[SJFC4ever]

http://msn.skysports.com/skysports/...1027137,00.html

I Walker Leicester 19 73%
P Robinson Leeds Utd 18 75%
T Sorensen Aston Villa 16 76%
P Jones Southampton 15 88%
J Lehmann Arsenal 15 88%

NB I don't understand this table, as they list it in order of #saves, rather than save%.


I think Goalkeeper is a position where stats could be used fairly well. It is a position that is largely independent of the others, since the keeper is the only player permitted to handle the ball (inside the penalty area).

Save % is an obvious stat, but I think you would have to adjust for something like save difficulty, which would be based on range, angle, velocity and height.

Of course, one problem is skill shots (like a chip). Therefore perhaps a better system would be a scoring system for save difficulty.


Perhaps a comparable area is hockey goalie stats. I know that a major stat in that is GAA (goal against average, per match). But I think this would be a very weak stat in football as the goalkeeper can play behind a very strong team, and have an abnormally low GAA. For instance, Barthez probably had a very good GAA last season, but Ferguson still (justifiably) looked to replace him.


Obviously there are other aspects to goalkeeper performance, such as distribution (throwing and kicking) and dealing with cross balls and through balls. But saving shots is obviously the major part. (I would have thought!)

[SJFC4ever]

Further to the bit about hockey goalie stats...

http://sports.espn.go.com/nhl/statis...alies&sort=sv%

Hmm, as it so happens the goalie with the lowest GAA did have the highest save%. And the goalie with the 2nd lowest GAA had the 3rd highest save%.

The problem with this is that you are dealing with very high save% numbers. There is only a .55 difference between the best (Turco) and the worst (Irbe). This may be less of a problem in football, since save% tends to be lower (anywhere between 70% and 90%).

[the101er]

I Walker - no wins
P Robinson - one win
T Sorenson - one win
P Jones - one win
J Lehmann - four wins

Are we measuring goalkeeper performance or shots on goal? This looks like a simple chart of who has faced the most shots.

I agree that there is more to goalkeeping statistics than shots faced or stopped. For example: part of being a goalkeeper is organizing the defence so that shots don't happen in the first place.

It would be interesting to check how many times the ball has come into the penalty area against each of these goalkeepers, and see how many goals resulted. Something, like "red zone" efficiency in American football.

[microbrew]

There might be something there with that redzone statistic? We can perhaps plot time of possession within 30 yards vs. goals scored?

Going back to the goalkeepers. Well, what stats do we have on goal keepers?

GAA
Record (wins, losses, ties)
Shots Faced
Shots Saved
Corner kicks given up

Then we can measure the keepers physical attributes: wingspan, height, lateral quickness, leaping ability

Some more subjective stats:
One-on-one percentage (leading to saves or shots on goal)
Classify the goals given up based on diffculty: unstoppable, own-goals, defensive breakdowns, out-of-position, blunders, etc.
Rebound control
Distribution (goal kicks, punts, throws and their quality)
Consistency (GAA, save percentage, etc. on losses, wins and ties)

And I'm not getting anywhere, fast. The only stuff I've seen is this punch/catch statistic:
http://www.mlsnet.com/content/03/analyze0801.html
http://www.mlsnet.com/content/03/analyze0808.html

The trick here is to ask the right question. The catch/punch correlation is possibly both indicating and masking a deeper truth.

As for ice hockey: How do you anaylize, say, Martin Broduer? You can slice up Brodeur's stats in many ways, one of which is that he plays behind of the stingiest defenses. You can also mention his excellent puck handling skills too. Or that he plays a lot of games.

[The Double]

Please, under no circumstances, post Keller's perfomance charts after Spurs' visit to the Bridge on Saturday

[Kevin in Louisiana]

I like the red zone stat, but it might place a little bit too much emphasis on the quality of the defense. Still, it could be helpful. I think it could be helpful to determine offensive efficiency, to see which teams are able to not just get into the area but convert their chances.

I think doing something along the lines of player performance charting but focusing on the GK's could be useful. It would be a bit quicker as you'd be focused on 2 players instead of 11 or 22 and maybe the isolation of the GK position would produce more useful stats.

[ChrisE]

I think this is an interesting subject, but I think the original poster is thinking far too deeply on the subject. First of all, although the difficulty of individual saves will vary widely, I think it's pretty safe to assume that, on average, the shots that goalkeepers face are of approximately the same difficulty (they do, after all, play in the same league). Furthermore, even if I thought that the shot-difficulty stat was that important, it is totally beyond our (posters on this board) ability to measure, both because it is a highly subjective stat and because nobody is going to sit down and grade every single shot taken in the course of a season. I think that shot percentage does a pretty good job of measuring that aspect of a goalkeeper's ability (clearly, a far better job than simply shots blocked).

Hirdt's article on the catch/punch statistic is probably one of the most surprising that I've read on mlsnet, because it seems to clearly indicate that a keeper's ability in this category can profoundly impact a game.

Tim Howard and Mark Dodd are the only keepers that I ran across (I only did the major keepers) who had career c/p's above 4 and save % above 70. Brad Friedel is the only goalie with career c/p's above 3.5 and a save percentage above 75, though Joe Cannon is close. [Correction, Jon Busch's qualifies in this group, and his numbers are remarkably close to Big Brad's. Take that for what you will.]

During his torrid season last year, Adin Brown recorded the second highest C/P ration in league history, while during 2003, Tim Howard recorded the third best.

Although I think that simple shot blocking ability is more important than the catch/punch, it seems that these two statistics alone can do a pretty good job evaluating keepers. (for the record, Thornton and Tony Meola come out very similar, Thornton at 73.1 and 2.98, Tony at 73.4 and 3.09, indicating that it might help account for keeper styles as well.

[Fah Que]

A few random comment

Catching is better than punching. Tim Howard catches a lot of balls which other keepers would decide to punch.

GK stats can be one of the most misleading. Last year most of Brad Friedel's save were either flying across the goal saves or saves against balls that are absolutely ripped. This year most of Tim Howard's saves were slow rollers or balls that bounces a few times before it gets to him.

I think GK stats reflect a lot on the defense in front of the keeper.

I think the best way to judge a GK is still the traditional fashion, which is polling the peers and coaches.