Shut Down Index (SDI) Results 2011-2012

So earlier this season, around the midway point, I started trying to devise a measurement tool to assess defensive skill for defenders. I decided after reading the postings on the Heavy Lifter Index written over at PuckProspectus.com that I could craft something comparable for blue liners, with a few minor tweaks. I ended up writing two postings on what I eventually termed the Shut Down Index (SDI). I have proposed the idea of an index value that compares the play of individual defenders to the league average for NHL D men that play in a minimum number of games. This is done across a select group of categories, with the original categories chosen as follows:

SDI Situational: Corsi REL QoC; Corsi REL QoT; OZ%

SDI Result: Corsi ON; Penalty Differential per 60.

*All data courtesy behindthenet.ca*

In analyzing the numbers I decided that if one is looking for Shut Down Defenders then it makes more sense to look at their Defensive Zone percentage rather than their Offensive Zone percentage, so for this most recent iteration I switched OZ% for DZ%, and I think the numbers make a bit more sense.

To explain in a bit more depth, the player's SDI is determined by determining how many standard deviations above or below the mean a given player is in each category, and adding the results together for SDI Sit, and SDI Res, and then determining the difference between Results and Situational. That is to say, players are rewarded for playing more difficult competition, with worse team-mates from a possession perspective, and from regularly beginning in a worse starting position. They are also rewarded for producing a higher corsi rate (shot attempt differential), and having a good penalty differential (taking fewer than they draw).

As a result of the weight given to starting position and the differential between quality of opposition and team mates, this statistic could (perhaps unfairly) penalize players that play on good possession teams, with high quality team mates, who have fewer starts in the defensive zone, resulting in a lower SDI Sit component. That being said, the results aspect should compensate, as players playing in less stressful situations can exceed average performance in SDI Res scores, and boost their overall SDI rating accordingly.

One of the aspects that makes this type of statistic valid and/or meaningful is repeatability. That is to say, veteran players who are handed similar ice time regularly produce comparable results over time. Players may see their SDI shift drastically, but this is often a result of changes in usage. For example, Jan Hejda has the highest average SDI rating over the past 5 seasons amongst active defenders. Despite this fact, his last season in Columbus saw him used in more of a typical top 4 role and less of a shut down role, with approximately 50% OZ starts, and facing less difficult competition. As a result, his SDI score dropped drastically. Were it not for this single season of more typical usage, Hejda would easily rank as the top Defensive D man to play in the NHL over the past 5 years, rather than second on the list (only Kurt Sauer has a higher average rating and he only played in the first 2 seasons of the analysis before retiring).

As you can see by examining the list below of the top 20 Active Defenders ranked by average SDI ratings over the past 5 seasons of play (minimum 2 years of NHL experience), many of the names listed are top tier D men around the NHL, consistently regarded as solid defensively. Some play in easier starting situations with better team-mates (Doughty and Seabrook for example), but their results push them into elite company from a defensive standpoint.

*Radek Martinek played under 30 gp in two of the past five seasons due to significant injury, thus his numbers are limited.*

Below I have listed the top 20 players from the 2011-12 season in SDI score, and it should be noted that there is obviously multicollinearity affecting the results (players that play together a lot are heavily influenced by their defense partner). That being said, the interesting part of these results is it may provide insight into who the top shut down blue liners of the future may be - keep an eye on some of the younger names on this list long term.

It is also interesting to note that the average age of the top 20 D men this season by SDI is a relatively young 26.25 years of age, which indicates that the idea that shut down defenders may not need to be as grizzled and veteran as initially thought. That being said, it does indicate that "prime" age for defenders is slightly older than goal scorers (24-25) for instance.

So as a last question - you're probably wondering how badly the Leafs did on the season. Well in the table below I have listed all of the Leafs blue liners that suited up in at least 30 gp this year, along with their PDO, SDI Res, SDI Sit, SDI, Age, NHL SDI Rank and lastly their 5 year running average.

It's fairly obvious that the Leaf's defensive issues stem from problems in the bottom 4, but before everyone turns around and blasts Luke Schenn for his horrid showing - finishing in the 19th percentile is pretty bad - please note that his average score is well above this season's result, and he actually ranked above the following D men you might be familiar with:

Scott ranked dead last amongst D men, but for all intents and purposes that stems from his usage and not his results. Steve Eminger and Ryan Whitney appear to not really be NHL level defenders. Having an average SDI score below -2.0 probably means they should either be in the AHL or playing another position (forward in the case of Whitney perhaps as he has offensive skills).

As a result of this statistic corresponding to yearly averages and standard deviations in each category, it should be noted that approximately 50% of D men will be above 0 while the other 50% will be below zero in any given season. Player's scores seem to be severely impacted by changes in usage, injury, and defense partners.

If I was to explore how this statistic could be usefully translated into a team building tool, I would probably actively avoid any defenders in free agency that had SDI average scores under -0.750. I would also largely seek to add D men whose average scores were above +0.75. Using that as your ball park numbers, you're looking at roughly the top 50 shut down and the weakest 50 D men defensively in the NHL.

Thus, entering Unrestricted Free Agency this year the following D men fit the description of useful defensively:

From a reliability standpoint in terms of health, Jackman and Allen were the healthiest of the group this past season, but they're also the most expensive. I would suggest doing something similar to what the Blue Jays have done in recent years with their relief corps, and sign 2 or 3 of these guys to $1-2 million contracts, and see how they do in training camp. If they play poorly or get injured then they can be demoted or put on LTIR.

I would take a long look at Jackman, Allen, and Zanon as stabilizing influences. I would also pursue Lundin, Rome and Stoner as potential breakout D men who have floated between the minors and the NHL for a few years, despite success from a defensive standpoint. Obviously their lack of offense has held them back, but the Leafs need better defense, not more offense.

Toss your thoughts in the comments section, and if you have any questions I can also be reached on twitter via @SteveBurtch.