Earlier this season, the SBNation blog for the Wild, Hockey Wilderness, had a post opining whether Wild defenceman Ryan Suter should garner consideration for the Norris Trophy. The article made an interesting case, highlighting that Suter was among the league leaders in a number of key statistical categories that would be of importance to a defenceman.
While the comments devolved into something of a farce, the post itself was well done. But I had one specific problem that was never properly addressed; why are these numbers important?
Since then, a number of different SBN blogs (not to mention other places) have taken a crack at trying to predict the Norris Trophy this season. In almost every case, the numbers have a tendency to skew towards being favourable to a player on that team, thereby making a stronger case for the hometown player and also introducing an unintended element of bias into the article.
The ability to properly place the stats in context is a problem that you come across in several areas, especially when a blogger, or newspaper writer, or TV analyst uses their platform to make a case that "This Player should be in the ____ Trophy discussion!" The only stats ever mentioned are the ones that highlight the case for that specific player, leading an unbiased observer to infer that those are the key factors to consider for an award.
Consider Erik Karlsson's Norris Trophy last season; 78 points in 81 games, playing 25 and a half minutes a night was an incredible season, to be sure. But just over a third of his points came on the power-play, and only 30 seconds of his ice time was spent on the penalty kill. Karlsson's supporters focused on his astronomical offensive totals, while detractors focused on the fact that his defensive credentials were virtually non-existent?
So what if we had an agreed-upon criteria for the Norris Trophy?
The Norris Trophy is awarded "to the defenseman who demonstrates throughout the season the greatest all-round ability in the position", according to NHL's website. All-around, to me, is the key word in that definition; the player judged to be the best defenceman in the league needs to be able to contribute in all three zones of the ice.
So, if you're going to use all available statistics to make a decision on who should win the Norris, I think it's important to make it clear how you weight one stat versus another up front. I came up with this methodology before accumulating any data; my goal is to remove as much bias as possible from this process.
I concentrated on the three main facets of being an NHL defenceman, and then found suitable statistics to build my formula. The three overall categories are:
Offensive Play - 30%
Being able to chip in on creating offence is an important aspect of playing defence in the NHL. Without it, you really only provide value in one segment of the game (taking away goals), which minimizes your overall impact.
To judge offensive contributions, we'll use the following stats.
Even-strength points per game (15%)
Power-play points per game (10%)
Zone shift (5%)
The value in even strength and power play production should be obvious to discovering who the best offensive defenceman are. Zone shift is included to help address a defenceman's ability to move the puck in a positive direction.
(Note: I am not a huge fan of Zone Shift, because the stat has several problems. However I think we need something to try and measure the value of being able to move possession up ice, and we've mitigated the risk of bad data influencing results by giving it a very low percentage).
Defensive Play - 30%
To be an effective all-around defenceman, playing strong defence should be just as, if not important, than being good offensively. Defence is more difficult to prove, so we'll incorporate possession metrics here to help capture who keeps the puck away from their goaltender.
Shots Against - 15%
Fenwick % - 10%
Penalty-Kill Time-on-ice - 5%
Shots against gets primary billing here because it is a purely defensive activity; Fenwick includes an element of offence. Combined with the zone shift statistic we should see strong possession players be fairly represented.
Usage - 40%
The reason usage is given top billing is simple; the award is supposed to recognize the top all-around defenceman in the league. A coach has 120 minutes of ice time to divide between six players on his roster; how he decides to allocate the ice time, how he uses them in different situations, says a lot about the trust he places in each of them. Therefore we want a Norris Trophy winner that plays a lot of minutes for his team and who plays in key situations.
Even-strength Time-on-ice - 25%
Quality of Competition - 15%
Before we dive into the actual results, we need to come up with a player pool; who are the people that we should be considering for the award?
Now, we could simply pick the ten or fifteen people that we consider to be the most worthy, but again, the whole point of this was to try and figure out who should win the Norris without any bias. So let's set a pretty basic filter of what we would consider to be a top defender in the NHL; someone who plays a minimum of 22 minutes a night, and someone who played in at least 75% of the games this season (36; which means Kris Letang is not being considered).
Ladies and gentlemen, your contenders for the 2013 Norris Trophy.