DRAFTING DEFENSEMAN IS HARD

There used to be a time and place, that when doing my rankings, if faced with the choice of a similarly valued forward and defenseman I would always choose the defenseman. My theory was that defensemen are more valuable as an asset than forwards, as they play a more important position. While nothing wrong with the theory, an individual needs to be able to adapt their view when presented with new information. What has made me change this viewpoint is the fact that defensemen take a longer time to develop than forwards, their development is more volatile and the fact scouts do a better job at picking forwards than they do defenseman. With these things in mind, let us repeat what we did with the forwards and look for the correlation between a player’s junior points in their first draft eligible season and their best season in the NHL.

The criteria we are using:

– played in the CHL in their 17 year old season

– played 250+ NHL games as of the end of the 2013-2014 season

– was drafted between 1998-2006

– their best NHL season is defined as the season with the highest points per game (minimum 40 GP)

After using these criteria, it leaves 54 defensemen that include:

B. Stuart, R. Regher, M. Walker, B. Allen, M. Skoula, J. Erskine, M. Biron, J. Fischer, F. Beachemin, B. Jackman, S. McCarthy, G. Exelby, N. Schultz, R. Klesla, K. Foster, D. Hamhuis, S. Morrisonn, C. Colaiacovo, T. Gleason, L. Krajicek, J. Harrison, D. Wideman, S. O’Brien, A. Johnson, M. Jurcina, Z. Michalek, J. Bouwmeester, J. Boychuk, I. White, J. Gorges, S. Eminger, T. Daley, J. Wisniewsk, P. Ranger, C. Campoli, B. Coburn, D. Phaneuf, B. Seabrook, S. Weber, D. Byfuglien, K. Klein, K. Quincey, M. Methot, A. Picard, B. Gervais, C. Barker, J. Schultz, M. Fistric, M. Green, K. Russell, C. Franson, M. Staal, M. Vlasic, K. Letang

So just like with the forwards, I split power play and non power play points and ran the correlation. The results were:

Non-PP points: 0.192

PP points: 0.180

What we can see right away is that correlation is much weaker for defensemen than that of forwards. To determine how much weaker, it is better to look at the adjusted r-squared, as that will gives out the amount of their NHL production that is explained by their junior production. The adjusted r-squared is:

Non-PP points: 0.018

PP points: 0.014

While I was expecting a lower number than the forwards, I was not expecting such incredibly low numbers of under 2%. The statistical model for forwards is over 18 times stronger with non-power play points and 21 times better with power play points. While the model for explaining future NHL potential for forwards shows a lot of value, the one for defensemen is pretty much useless at this point. In the future, I would like to go further back with the data to get a larger sample size to draw from and see if the correlation strengthens. With defenseman developing at a slower rate than that of forwards, part of the problem may lie in the fact that a larger percentage of the defensemen used in the model have not reached their prime compared to that of the forwards. Also scoring is a more important number when evaluating forwards, as it is their primary goal as a player while for defensemen it is not.

So since points have not been a useful tool for evaluating defensemen, let us instead use their junior points and compare against games played in the NHL. The correlation is:

Non-PP points: 0.326

PP points: 0.240

The adjusted r-squared is:

Non-PP points: 0.089

PP points: 0.040

Well at least these numbers are a little more promising and useful than what we have seen previously. While the correlation is stronger for non-power play points, the main item that jumps out for me is that the adjusted r-squared for non-power play points is twice as strong as for power play points. Even strength scoring being more important than power play scoring does makes sense, especially with defensemen, as there tends to be more consistency in their even strength ice time compared with power play time. Another way we can compare the two is through the average power play and non-power play scoring in juniors of these 54 players versus their best season in the NHL:

Non-PP points: 100.3%

PP points: 58.2%

Once again, power play points are not as important, with the average player only scoring 58% of the power play points they scored in juniors. It is interesting to see that the average non-power play points, in their junior season, are essentially the same as the NHL. It just goes to show you the development of defenseman does not follow the same trajectory as forwards. For comparison, if we use our sample of 126 forwards, the numbers are:

Non-PP points: 77.6%

PP points: 48.6%

In a nutshell, junior points for defensemen are not a very useful tool in creating a statistical model for their future offensive potential and turning more to the scouting reports is going to be a key component when analyzing them. However, the stats are not void of use, as discovered in the fact that even strength scoring for a defenseman in juniors is probably about twice as important as their power play scoring. In the end, this confirms something we already knew, drafting defensemen is hard.Damn hard.

Next article April 13: Age Matters…Sort Of