Last week, Dan Rosen of NHL.com wrote an article on the early race for the Norris Trophy, which is given to the league’s best defenseman annually. His anointed winner, at this point, was Minnesota’s Ryan Suter, who finished a close second in the voting last year to Montreal’s P.K. Subban. Suter is a workhorse, there’s no doubt about it. He average 27:17 of ice time last year in all 48 games, and this year has played an astounding average of 29:33 per game through 25. Suter always played second fiddle to Shea Weber in Nashville when the two were a pairing prior to last year, and Suter has gained national attention for his solid play as well as his mentoring of young Jonas Brodin, who was in the conversation for rookie of the year last season.
But the issue with Rosen’s piece, and of the opinions of many main-stream hockey writers, is that time on ice isn’t a performance metric; it’s a usage metric. Ice time exemplifies the trust a coach has in a player, but that could just as easily be a result of a shallow pool of NHL defensemen or a particular style of coaching rather than an actual indication of strong play. There was a time when that we had to trust the opinions of hockey “experts” and the eye test on matters like this, but that time is long gone. While defensive analysis is far from perfect at this point, there are ways to measure the performance of Suter compared to other strong defensemen like last year’s fellow finalists Subban and Kris Letang; Rosen’s other Norris nominees so far this year, Oliver Ekman-Larsson and Marc-Edouard Vlasic, and the highly regarded Weber, Zdeno Chara, Duncan Keith, Erik Karlsson and Drew Doughty.
The main-stream way to measure defensive performance would be +/-, but the issues with that as a metric are a) It includes shorthanded and empty net goals, which disadvantages players that play in offensive situations, and b) the sample size when it comes to goals is too small to nullify the impact of luck and variation on the results. How do we solve this? Corsi is a statistic that measures shot attempt differential while a player is on the ice at even strength. There are enough shot attempts during a season that the impact of luck is minimized, and early hockey analysts found that shot differential correlates extremely well with possession and scoring chance differential, and thus with goals and wins. What about the fact that winning teams will often go into a defensive shell and thus shoot less (otherwise known as “score effects”)? We will only measure this when the score is within one goal in the first two periods, or tied in the third. This is called Corsi Close.
So how does Ryan Suter stack up against these other nine NHL’s elite defensemen in Corsi Close? He is last. When Suter is on the ice, his team has accounted for only 49.8% of the total shot attempts by both teams, meaning that he can’t even break even in possession. This is despite his coach sheltering him with 56% of his non-neutral zone post-whistle shifts starting in the offensive zone, the highest number among the ten. So to recap, in comparison with the other top defensemen I mentioned, Suter starts the highest portion of shifts in the offensive zone, and still manages to post the worst possession numbers. Does that sound like a Norris Trophy worthy defenseman, or a guy whose coach simply relies on him a little bit too much? If early season numbers are any indication, the obvious answer is the latter.