HFBoards Today's Star Defenseman vs High-Level Competition
 User Name Remember Me? Password
 Register FAQ Members List Search Today's Posts Mark Forums Read
 Mobile Hockey's Future Become a Sponsor Site Rules Support Forum vBookie
 Notices Please do not post or solicit links to illegal game streams.

 By The Numbers Hockey Analytics... the Final Frontier. Explore strange new worlds, to seek out new algorithms, to boldly go where no one has gone before.

# Today's Star Defenseman vs High-Level Competition

 11-18-2012, 07:44 PM #1 Sureves Registered User     Join Date: Sep 2008 Location: Ottawa Posts: 10,518 vCash: 500 Today's Star Defenseman vs High-Level Competition Hey everyone, Today I decided to do some analysis on how the leagues' best defensemen (or some of them) performed when facing top level competition. This seeks to control for the fact that some defenseman face really tough competition and thus are theoretically at a disadvantage over those facing easier competition. So what I did was I looked at how Karlsson, Pietrangelo, Chara, Doughy, and Weber performed against the top 60 scoring forwards in the NHL in the 2012 season and used goals as a measure of performance. Here's how it works: I got a list of all the players the defensemen in question played against in 2012 measured in minutes and also showing how many goals that player scored on them. This was attainable for all players that the defenseman in question played against for around 17 minutes or more (ie. I couldn't see literally every player they played against). So then I looked at which of the top 60 point getters showed up on these guys lists and looked into them for each defenseman. From there, I used GF ON 60 from behindthenet.com to get the forwards' "average" GF when playing and compared that against their realized GF per 60 when playing against the defenseman in question. To avoid small sample size, what I did is I summed the total TOI versus Top60 players that each defenseman did, and summed the total goals scored against the defenseman from those same Top60 players, and divided to obtain the "average Goals per 60 Realized". From there I was able to see if the defenseman had a positive performance (measured by Goals Against) compared to what was expected. I did the exact same analysis with goals against (ie. tracked to see if the defenseman scored more goals against their competition than the competition usually allows). Here's an example of what I'm talking about in case this wasn't clear, here's Chara's chart: In other words, Chara played 605 minutes against top competition, and his competitors scored are a rate of 2.68/60mins which was 0.40 less than they normally score (ie. 3.08/60 mins) and as such, when Chara was on the ice, they shut down those players very well. Similarly, Chara's team scored 3.27/60mins and normally those players got scored on only 2.89/60mins and as such Chara did a good job offensively as well, scoring 0.39 more than expected. OK so I think you guys get the idea. One thing to note is I only allowed 1 person per team to show up on the defensemen's lists since otherwise it would kind of be double counting. I realize not all the forwards in question played together but to play it safe, I only allowed one per team (and selected the one with the higher point totals). So here are the nominal results: This basically shows you how Team X performed with Defenseman Y on the ice against those tough competitors relative to how those competitors usually perform. Pietrangelo's team really shut down the competitor's hard, which isn't altogether surprising considering the Blues broke records this year on their defensive performance, while Karlsson's team really scored on the competition much worse than they normally get scored on. I thought I'd post these results first since they are of interest, but I think it should be clear that defenseman on good teams naturally should perform better here, given the fact that their team as a whole is so good. So then I prepared results that adjusted for the fact that each team has a different overall skill level. For example, Los Angeles doesn't have that scoring power that Boston or Ottawa has (or at least had in 2012 for whatever reason), and as such, when looking at Doughty's goals for, it isn't really fair to oh wow it's so low, he must be bad! Rather, it makes sense to adjust for the players LAK has relative to the other teams in question. For example, Doughty's GF OFF (ie. the GF scored by LAK without Doughty on the ice) is only 1.9 and for Boston without Chara it's 2.68, as such it's pretty clear that the other players on the team are benefiting Chara more than Doughty in terms of GF ON. As such, I adjusted by calculating the average for each team and adjusting accordingly. This table I believe best illustrates how the specific defenseman in question performed with their duties as a shutdown defenseman. As expected, Karlsson performed the worst defensively of all of the defenseman in question, and not all that surprisingly, Pietrangelo performed the best at shutting down the opposition. Offensively, Karlsson performed by far the best as expected, with Pietrangelo performing the 2nd best. This analysis is not gospel or perfect, but I feel as though this is one of the best looks at how defenseman perform relative to tough competitors to date. I intend - depending on the reaction of how you guys feel on the faults and merits of the analysis - on performing the same analysis using Corsi as a measure as goals against is very volatile and not altogether that reliable. Thoughts? EDIT: On second though a time-weighted average would have been better than a regular average for the GFON/GAON 60 of the competitors, but oh well, doubt it would have much effect as seen below (GF ON) Karlsson: 3.06 Pietrangelo: 3.07 Chara: 3.08 Doughty: 3.06 Weber: 3.09 Last edited by Sureves: 11-18-2012 at 07:58 PM.
 11-18-2012, 07:54 PM #2 deckercky Registered User     Join Date: Oct 2010 Posts: 7,083 vCash: 500 That can't be a full list of the top 60 scorers Chara has faced. I'm positive he had icetime against the Sedins last year for example.
11-18-2012, 08:03 PM
#3
Sureves
Registered User

Join Date: Sep 2008
Location: Ottawa
Posts: 10,518
vCash: 500
Quote:
 Originally Posted by deckercky That can't be a full list of the top 60 scorers Chara has faced. I'm positive he had icetime against the Sedins last year for example.
I can tell you for sure that it was less than 19.2 minutes (the lowest amount of ice time I can see, which was against JONAS HILLER.

As I said, it doesn't give a full list, only the ones with significant amounts of ice time unfortunately.

11-21-2012, 06:29 PM
#4
Jason MacIsaac
MARS Officer

Join Date: Jan 2004
Location: Halifax, NS
Country:
Posts: 17,705
vCash: 500
Quote:
 Originally Posted by Sureves Hey everyone, Today I decided to do some analysis on how the leagues' best defensemen (or some of them) performed when facing top level competition. This seeks to control for the fact that some defenseman face really tough competition and thus are theoretically at a disadvantage over those facing easier competition.
This doesn't take poor goaltending or team defense into consderation.

 11-22-2012, 05:06 AM #5 Ola Registered User     Join Date: Apr 2004 Location: Sweden Country: Posts: 22,841 vCash: 500 I like the topic, I also often finds myself wondering about some players post-lockout vs how they would have fared against for example really loaded teams pre-lockout. But you definitely can't look at stats on that topic. I think two players like Stamkos and Weber are examples of players who on paper are looking great, but whom I don't quite find as elite as I am used to seeing elite players be from the time I followed the league before the lockout. To illustrate what I mean, I just find someone like Weber to defensively be pretty far behind guys like Blake, Pronger, Bourque and co.
11-22-2012, 06:59 AM
#6
Sureves
Registered User

Join Date: Sep 2008
Location: Ottawa
Posts: 10,518
vCash: 500
Quote:
 Originally Posted by Jason MacIsaac This doesn't take poor goaltending or team defense into consderation.
The final adjustment (table 3) does actually.

Forum Jump