A Theory on NHL Defenseman Metrics

This article is going to postulate a theory with respect to establishing the inherent value of a defenseman, and further more try to establish a rudimentary scale for grading such players. For hockey fans interested in some interpretative analysis, buckle up because this should be a doozy. It’s also an attempt to explain why your favorite hockey team signs washed up Senators defenseman who make opponents forwards look a combination of Gordie Howe and Wayne Gretzky.

When Michael Lewis wrote Moneyball, he introduced the world to sabermetric theories as Billy Beane applied them to the Oakland Athletics. The specialized analysis that Beane used to find undervalued players allowed him to remain competitive in a market where his teams would be completely over-matched by opposition resources.

With the new NHL salary cap, all teams are theoretically on an even playing field. No team (I’m looking at you Rangers) can use its financial resources to dominate the free agent market with its advantage in revenue. This places a premium in scouting and statistical analysis, which is an area where I feel the NHL is at a distinct disadvantage. Although some people, like Gabe Desjardins at Behind the Net have applied sabermetric principles to hockey, overall the sport is severely lacking in quantifiable data. His research takes into account the quality of play, which is something I’m not quite sure how to apply; Perhaps an “average shots by opposing players” stat? Theoretically, there’s no reason why we can’t grade defense.

Defense is the position on the ice where there is the least amount of quantifiable data regarding a players effectiveness. For forwards (although assists, and to a smaller extent goals are a team dependent stat,) there is at least a method to determine the value of a player. However, in the defensive sphere, the only stat that can be found is hits and blocked shots, which would be like tracking a pitcher’s usefulness by his curveball’s speed.

What I’m proposing is an additional metric to grade defenseman. Frankly, I’m not sure this is a stat that the NHL itself would be interested in keeping at first. From a scouting perspective, though it would seem when investing limited resources into players the clubs could assign some of their scouts to track this metric with respect to potential targets (For example, Mark Streit versus Wade Redden) or draftees (like Victor Hedman), in order to get an idea of how much a player is worth on the open market. If something like this caught on, perhaps the NHL could eventually assign its official scorers to track the play of defensemen over the course of the night, but the financial implications would prove difficult if immediacy was required. If grading players based on tape , this endeavor would become much more feasible

The first part in establishing a metric is determining the true goal of a defenseman. I’m proposing that the first, primary goal of a defenseman is to eliminate shots from his quadrant of the ice. Viewing the amateur graphic below, I’ve divided the ice into segments. Essentially, in Segments 1 and 12, the defenseman’s primary goal should be to eliminate high-quality shots, with the oval in the center being both partners areas. Blocked shots and contested shots also have to be factored in. Offensively, I believe that the goal of the defenseman, should be keeping the puck in the zone, and preventing breakaways. Assists also are the mark of a solid defenseman, and should be given weight in any scoring metric. The neutral zone regions theoretically are the least important, but will be acknowledged in the “preventing breakaways” category.

The problem with using +/- is that those are team-specific stats, like RBIs and runs in baseball. No matter how well you’re playing your position, if a poor goalie or attack is paired with you, your +/- is going to significantly suffer. What this metric would attempt to do is isolate the defenseman as much as possible, in the absence of a supporting cast. Takeaways obviously would hold an increased value, as well as pinning an opponent away from your end of the ice.
I do acknowledge that teams with weaker defensive forwards would be penalized somewhat in this, but I’m not sure any way exists to prevent that (I’m open to ideas). What I’m trying to do is extract scoring and goalie performance from how defenseman are viewed, and for now this seems like the most efficient way to do that.

The primary use of the data would be, in the tradition of the A’s, to find undervalued defenseman in a capped league. Accordingly, I’m proposing the following scoring system as an additional way to grade defensemen.

* Uncontested Shot -3
* Contested Shot -1
* Give Up Breakaway -5
* Giveaway – 2
* Giveaway in Defensive Responsibility Zone -5
* Blocked Shot +3
* Takeaway +5
* Successful Pin/Check (opposing team loses puck) +1
* Successful Outlet Pass +1
* Assist +3
* Goal +3

Obviously this is still rudimentary, but I’d love to hear any other hockey fan’s thoughts.

Raise a Glass

Put the Glass Down

Tags: defenseman, first derivative, firstderivative is a dork, geekfest, nhl, sabermetrics, stats
Categories: Daily Specials, Domestics, Non-Alcoholic

This entry was posted on Tuesday, June 16th, 2009 at 2:42 pm and is filed under Daily Specials, Domestics, Non-Alcoholic. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.