Tuesday, September 18, 2007

Shots Matter

James Mirtle presented a fine and finely detailed look at some goaltending history in this post. Being the sort who has a certain fascination with the behavior of numbers, The Peerless offered a comment, which we share with you below:

Thanks for that excellent look at goaltending trends. One of the things that has occurred to me, anecdotally, is that shots matter from a defensive point of view. At this level of play, there just isn't that much difference in save percentage among goaltenders in the NHL (last year, the top 20 goalies in save percentage with at least 50 games ranged from .902 to .922).

But looking at one goaltender to whom I pay attention -- Olaf Kolzig of the Capitals -- the effects of shots is telling. Kolzig faced 1,771 shots last year (33.4 shots per 60 minutes of play). For his trouble, he had a goals against average of 3.00, 20th in the league among goalies with at least 50 games, while posting a .910 save percentage. In order to shave that GAA to, say, 10th (2.58, Rick DiPietro) facing the same number of shots, Kolzig would have had register a save percentage of .923. Had he done so, he'd have led all goaltenders who played 50 or more games in that statistic.

Here is another way to look at it. Marty Turco and Kolzig had the same save percentage in the regular seson (.910). But while Kolzig had his facing 33.4 shots per 60 minutes, Turco had his facing 24.9 shots per 60 minutes. The arithmetic results in a 0.77 differential in their respective GAA (more than 25 percent better in Turco's case). Which team's goalie -- and team -- was more successful?

That's a long way round to the point that "shots" seem the most underrated statistic in the sport. Which brings to mind another factor in this -- what has been the trend in blocked shots and missed shots? And, have goals scored as a share of "shots taken" (as opposed to "shots on goal") changed over time?

We're mindful of a rather standard response to the "shots" argument, that there are "shots" and there are "shots." By this we mean that goaltenders might face a lot of easily stoppable shots that inflate a shots on goal statistic. In any given game, that is certainly possible, but data being what they are over a series or a population, we suspect that over the course of a regular season that sort of thing evens out -- that the 6-8 easy shots goaltender A might face today are evened out by a similar number goaltender B faces next week or goaltender C faces next month.

I think it useful to pay particular attention to the comparison of Kolzig and Turco, above, and apply the analysis we used in the first example. For Kolzig to have achieved Turco's GAA (2.23) while facing the number of shots he did last year, he would have had to post a save percentage of .933. That would have led all goaltenders, according to the NHL's statistics. And Turco finished fourth in GAA.

Wayne Gretzky famously said that you miss 100 percent of the shots you don't take. It's merely another way of saying, "shots matter."