A couple months ago, I stumbled upon Basketball-Reference.com’s NBA MVP Award Tracker. The tracker, “ranks candidates based on a model built using previous voting results.” So rather than analyzing who was the league’s most valuable player, it projects who, according to past voting patterns and tendencies, is most likely to be named the MVP. I subsequently tried to parlay this to the NHL awards. Projecting the Hart Trophy in this way seemed next to impossible, as goalies and players are both eligible for it. The only award that I could foresee using this method for was the Norris Trophy – awarded to the league’s top defencemen – because of it’s specificity and the multitude of effective stats that would potentially be at my disposal.
For the Norris, I first looked at the past five winners of this trophy. I took their statistic totals, looked at where these totals ranked them (with respect to that season’s defencemen) and noted how much their statistic totals deviated from league average totals. Through trial and error, I found that that being a leader in points, time on ice, Corsi For (CF%) and Corsi For Relative Team (CF% RelTm) percentage in close situations, and Corsi Against Per 60 minutes (CA60) were most indicative of being chosen as a winner for the Norris (points factored in most, by quite a large margin, while the other categories factored in almost equally behind). With these numbers, I then combined the percent deviation from average for the different categories and scaled the different stats according to past results. The reason for using these percent deviations was to adjust for different scoring environments in different seasons, as well as varying numbers of games played in a season (i.e. the lockout-shortened season).
|Lg Averages (min 40 GP)||4.96||16.87||21.83||19.83||49.87||55.27|
|% Deviation from the Average||20.97||226.02||179.43||24.31||15.90||13.52|
|Lg Averages (min 25* GP)||2.86||9.43||12.29||19.85||50.09||54.5|
|% Deviation from the Average||284.62||186.32||209.19||17.13||17.19||16.51|
|Lg Averages (min 40 GP)||4.35||16.42||20.77||19.95||49.73||55.3|
|% Deviation from the Average||336.78||259.32||275.54||26.92||10.20||-1.63|
|% Deviation from the Average||257.94||177.44||194.54||18.24||2.35||0.16|
|Lg Averages (min 40 GP)||4.62||16.93||21.55||19.73||49.89||54.75|
|% Deviation from the Average||203.03||224.87||220.19||34.82||14.25||16.16|
Thus, when the percent deviations are added, we end up with a figure that we are referring to as Norris Points. Now, I had to apply this method to this season’s defenceman. Here are the results (for length concerns, I’ve only included the top fifteen, but the full list can be found in the Google Sheet at the bottom of this page):
|Rank||% Deviation from Pts Avg||% Deviation from ATOI Avg||% Deviation from CF% Close Avg||% Deviation from CF% RelTM (Scaled)||% Deviation from CA60||Norris Points||Player|
As you can see, the top three defenceman in Norris Points also happen to be the three Norris nominees, which we will now discuss:
Erik Karlsson is the epitome of an elite offensive defenceman. Three of the past four season he has led all defenceman in points – the year he didn’t, he lacerated his Achilles tendon. Karlsson’s unparalleled speed and stickhandling among defencemen ensure that he’s always in the play, no matter which side of the ice the puck is in. Although he is no slouch on the back end, Karlsson’s falls short of the trophy this year as his fellow nominees were positive players in their defensive game. There has been one instance in the past five years that a defenceman won the Norris while being below average in CA60 (Corsi Against per 60 minutes): Karlsson’s 2011-12 season where he was an absurd 275% above the average points for blue-liners – as shown above, this season, Karlsson is only 199% above average.
Drew Doughty is easily the most all-around talented defenceman of this group, as he was well above the average in all categories measured. From the past five years, we’ve seen two types of winners of the Norris Trophy. Those who had tremendous all-around seasons and those who excelled in points. Although even during these all-around seasons, winners didn’t have a deviation from the points average that was below 179% – Doughty’s is 108%, and as points is, seemingly, the most looked-at and important category to voters, that isn’t going to cut it.
PK Subban is basically the middle-man between his fellow nominees, the best of both worlds if you will. He doesn’t have the issue of being below average in CA60. Moreover, he is 171% above the average with regard to points. When it comes down to it though, the Norris Trophy is awarded to the defenceman that stands out against his peers, and the past winner list can attest to that. CF% relative to team captures this rather well, as it presents the amount of offense generated by a player compared to his teammates. Subban’s CF% RelTm of 7.6 is over twice as much as both Karlsson and Doughty. Without Subban and the offensive acumen he provides, Montreal is a vastly different team. It is for these reasons that we believe that PK Subban will be the Norris Trophy winner of the 2014-15 NHL season.
Adam is a student at McGill University. You can follow him on Twitter @adam_m3318.
You can follow Hit the Cut on Twitter @hitthecutblog.
Compiled stats can be viewed on this Google Sheet