Something a little bit different. I did the initial data gather on this earlier in the month, but the delay has made for good timing given the recent use of "Corsi" in the Boston Globe (and the reply at the Score).
The idea that is sorting floating around out there is that mainstream hockey writers (and simply hockey culture, if you allow me to be monolithic for a second), are not "believers" of new statistical categories and research. They do not use them to provide analysis, narrate, or explain hockey games. The mainstream lags behind the cutting-edge, and so forth.
This seemed to me to overstate things.
If puck possession is good (and lots of hockey people would argue this), and something like Corsi measures this in some way, is it really the case that mainstream hockey ignores this? If we translate Corsi as simply "shot attempts" (pucks toward net), what do we observe?
This is a small slice of research. To investigate the total mass of mainstream hockey writing requires a lot of more work, but I wanted to pilot the idea by looking at the news and stories coming out of NHL.com. The setup was simple. Google site:www.nhl.com/ice/ (this would remove a lot of noise), keyword: "shot attempts". I downloaded the Google hits pages, inserted raw data, and then hand-screened for duplicates and "penalty shot attempts". Nothing complex.
Table 1. C'mon. It's still a table. So what if it's small? ... What?
Chart with smoothed lines. Via i41.tinypic.com
1) As of August 12, Google (and my screen) counted 19 hits of "shot attempts" for 2013.
2) The chart shows what might be called the cyclical nature of norm diffusion. Basically, very few new ideas or language diffuse without resistance or stops/starts. The chart suggests this is happening. My intuition agrees, but this is very much suggestive, not conclusive. Every year after an increase in hits is followed by a small decrease. 2013 should see an increase, and given the count so far (19) it is on-track to surpass 2012.
3) In 2007, there were stories that talked about "shot attempts". They talked about individual players (Modano, Whitney, Ryder) and group efforts (Wild forwards, OHL teams London and Ottawa). Sure, four stories, but still 2007.
4) So what is the crux of the Corsi tension if "shot attempts" does not seem all that radical anymore. To me, statistical hockey analysis floats (perhaps too easily) between units of analysis. Are we analyzing individual contribution alone, individual contribution to the overall team or a particular group (e.g. defense pairings, wing-center pairings, forward lines, or 5-man PP), or simply to evaluate team/run of play? The development of Corsi Rel, Corsi Rel QoC, etc., it seems to me, was to focus on the individual. If hockey people lean individual or system, they lean system. Hockey stat people don't lean either way (I think, in a sort of agnostic way). It depends on the question, and we should foreground that. Maybe it is this non-allegiance that rubs hockey people the wrong way. I'd be curious to hear what others think on this.
5) Bottom line: So long as puck possession is widely recognized as important (e.g. see WaPo's
rumor article on the Caps signing Grabovski), and that shot attempts / "Corsi" is considered a legitimate proxy for that variable, we will continue to see an overall increase in its use. This is not all that unsurprising. This limited look confirms that.
Final note: three distinct stories used "PDO" on NHL.com. All 2013. GoogleDoc spreadsheet here.