“It’s funny because our game looks at numbers just like other games [...] but as much value as we assign to puck possession and how essential it is to winning, we really don’t have a numerical value for it that everyone can agree on. Remember when [A's general manager] Billy Beane started emphasizing on-base percentage in baseball? It wasn’t just a curious number; it changed the game. It redefined the type of player you wanted on your team. It’s coming in hockey; we just have to figure out how.”
That’s Ken Holland from Brian Cazeneuve’s SI column on the importance of puck possession. I don’t know if the Red Wings GM is being purposely obtuse with Cazeneuve or if he really isn’t aware, but the numerically inclined amateurs who cover hockey have been employing a quantified version of possession for years. It’s called Corsi.
I frequently discuss Corsi in my own articles here at theScore and elsewhere, but he’s a brief overview of the metric:
Developed by Buffalo coach Jim Corsi, it was originally designed to track goaltender performance. Corsi is a differential or ratio of all shots at the net (shots, missed shots and blocked shots) for and against while a player is on the ice (usually at even strength). It is a proxy measure for zone time: high Corsi players spend more time in the offensive zone and vice versa.
Gabriel Desjardins does a more thorough overview of Corsi here. I also have a look at the Calgary Flames individual possession rates and scoring chance numbers from this past season in this post. Of note: the correlation between Corsi and scoring chance ratios was just south of 0.7 in the 79 games I counted this year (a number which is both strong and statistically significant). Others have found even higher correlations between Corsi and chances in the past. Objective NHL’s JLikens determined the strength between outscoring and outshooting over large samples in this investigation. These links are merely scratching the surface on the body of work that has been done to establish Corsi’s value as a metric.
Cazeneuve’s piece highlights many of the reasons possession has become more important in a post-lock-out NHL: the crack-down on obstruction, the rush to imitate the success of the Detroit Red Wings, etc. He interviews more than a few notable NHL GM’s including Holland, Mike Gillis and Steve Yzerman, none of whom apparently agree on a method to reliably track puck possession, even though they universally agree it’s a key to winning.
This means one of two things: either the league’s executives are actually behind the curve in terms of quantitative hockey analysis or they are purposely misdirecting Cazeneuve because they don’t want to share any secrets (or some mix therein). The latter possibility is hinted at in the piece by Doug Wilson, who cheekily offers a “no comment” to the question of whether the Sharks track possession or not.
If it’s the former, however, then I’m pretty confused. As mentioned, there’s a vast body of on-going research on possession that is growing and widely published on the web for free. At this point, NHL executives and GM’s would have to be either purposefully ignorant of this research or so isolated in their organizations that they are woefully unaware of anything beyond their own borders.
If it seems unlikely that a bunch of outsiders and amateurs would be a step ahead of hockey’s higher-ups when it comes to something this fundamental, but take another look at Holland’s quote at the opening of this post. He mentions the Oakland A’s Billy Beane and his reliance on new stats* (including on-base percentage) to find efficiencies in baseball markets in order to compete. What goes unsaid is the fact that Beane was actually standing on the shoulders of Bill James, the father of baseball Sabremetics who was both an amateur and an outsider himself.
Nobody in the league can agree on just how to quantify puck possession. But everybody knows that hockey’s most elusive statistic is essential to winning the Stanley Cup.
So says the subheading of the Cazeneuve piece. He’s only half right.
*(Beane’s discovery of and reliance on Jamesian “new stats”, of course, was the basis for the well-known Michael Lewis book Moneyball which I highly recommend, even if you have no interest in baseball itself)