Recently, a lot of people have argued that Mike Green shouldn't win the Norris trophy because he plays (relatively) little on the penalty kill. I wanted to see how much value the Norris trophy voters actually place on penalty killing.
The NHL has officially tracked ice time since the 1997-98 season. I looked at the amount of penalty killing ice time ("PK IT"), for each defenseman who's finished in the top five in Norris voting over the past 11 seasons.
Summary
The median ranking is 17th (this is roughly an average #1 penalty killing defenseman in the thirty team NHL), however there's a huge range.
Ranking
Percentage
Top ten
27.3%
Top 30
69.1%
Top 60
78.2%
Top 120
92.7%
A surprisingly small number of Norris candidates are what I would call elite penalty killers - i.e. finishing in the top ten in PK IT. This is a smaller percentage than I was expecting. This shows that elite penalty killing, on its own, is not sufficient to make a defenseman Norris calibre. However, I don't think that's a very surprising conclusion. Adam Foote and Craig Ludwig were elite penalty killers but (rightfully) never won the Norris because their offense was so lacking.
What I found surprising is that there's a large number of defensemen - over 30% - who were Norris candidates despite ranking worse than 60th in PK IT (which approximates to not even being a top-pair penalty killer). Nearly one-third of Norris candidates are, at best, their coaches' second-pairing option on the penalty kill. This is the somewhat surprising conclusion: a defenseman can be Norris calibre even if they're rarely used for penalty killing.
Pre vs post lockout
Rank
Pre-lockout
Post-lockout
Top ten
54.3%
15.0%
Top 30
77.1%
55.0%
Top 60
85.7%
65.0%
Top 120
91.4%
95.0%
Another interesting finding is that penalty killing correlated much more closely with Norris voting before the lockout. Pre-lockout, the majority of Norris candidates were elite penalty killers, and an overwhelming majority were in the top six. Post-lockout, just 15% of Norris candidates were elite penalty killers and a bit under two-thirds were in the top sixty.
I can think of a few possible explanations:
1. There were more elite, complete defensemen before the lockout. Before the lockout we had Lidstrom and Pronger in addition to the last few elite seasons from Boruque, Chelios, MacInnis, Stevens and Blake. There's been a clear drop-off in high-end blueliner talent since 2004. The pre-lockout elite defensemen were great at everything, including penalty killing, so there was a strong correlation between Norris nominations and PK IT. Post-lockout, we just have Lidstrom and Chara (the other elite defensemen are too young or too inconsistent).
2. There are far more powerplays post-lockout, especially in 2006 and 2007. Perhaps this had lead to coaches using their defensemen for increasingly specialized roles. Pre-lockout, maybe an elite, Norris-calibre defenseman could handle 4 minutes per game as their team's #1 penalty killer. Post-lockout, with so many more powerplays, perhaps coaches have started using more role players on the PK, to help rest their stars, who would be unable to handle so much PK ice time.
In summary: penalty killing was more important to Norris trophy voting before the lockout
Finalists vs winners
Even though, in general, it doesn't look like one needs to be an elite penalty killer to be nominated for the Norris, arguably one needs to be an elite penalty killer to win.
Out of the past eleven Norris trophy winners, 8 were in the top ten in PK IT (Blake in 1998; MacInnis in 1999; Pronger in 2000; Lidstrom in 2001, 2002, 2003, 2006 & 2007).
Ten of the past eleven Norris winners were at least in the top thirty in PK IT (aforementioned plus Lidstrom in 2008 and Chara in 2009).
The only exception was Niedermayer, who was 47th in PK IT in 2007. He was 2nd among defensemen in scoring (and perhaps his reputation was buoyed by his outstanding playoffs the previous spring).
This shows that, although defensemen who don't play on the penalty kill can be nominated for the Norris, trophy winners are overwhelmingly likely to be a #1 penalty killer.
2010 outlook
The PK IT ranking of this year's likely finalists are:
Player
Rank
Pronger
4th
Keith
18th
Lidstrom
22nd
Green
81st
Doughty
84th
Traditionally, one of the well-rounded Pronger, Keith or Lidstrom would likely win the Norris. I'd wager that, pre-lockout, these would be the three finalists. However, the above data shows that the voters seem to care less about penalty killing post-lockout. This will probably allow Green, with his dominant offense, to earn a spot in the top three, and he might even win. It'll be interesting to see what happens.
Conclusion: four lessons
1. Elite penalty killing, on its own, is not sufficient to make a defenseman Norris calibre.
2. A defenseman can be Norris calibre even if they're rarely used for penalty killing.
3. Penalty killing was more important to Norris trophy voting before the lockout.
4. Although defensemen who don't play on the penalty kill can be nominated for the Norris, trophy winners are overwhelmingly likely to be a #1 penalty killer.
Last edited by Hockey Outsider: 04-04-2010 at 05:18 PM.
Source: NHL.com detailed ice time tables
Note: players are ranked based on total, rather than per-game, PK IT
Great work, as usual, HO.
However, since you do make the point that players are ranked on total SH TOI, don't you think that is a major reason why it seems that some defensemen are rather strangely low in PK time rankings, considering Detroit has always been one of the most disciplined teams in the league, as well as Boston in Chara's Norris-winning year?
Howcome? Wouldn't it make more sense to rank them based on who played more per game?
Great work though, interesting stuff!
Thanks! Three reasons I used total (rather than per game) PK IT:
1. Defensemen have zero value if they're not playing. A defenseman who plays 3:00 on the PK every game has contributed more than a player who plays 4:00 on the PK in 45 games. (I suppose you can argue that the second defenseman is actually his coach's top choice, but he contributed less to his team's success).
2. If I used per-game stats, there's no clear games played threshold. Obviously I'd need to filter out the players who only played one or two games, but I wasn't sure where to draw the line.
3. Most Norris trophy finalists play close to the full season anyway, so it shouldn't affecting their rankings too much.
Quote:
Originally Posted by Blades of Glory
Great work, as usual, HO.
However, since you do make the point that players are ranked on total SH TOI, don't you think that is a major reason why it seems that some defensemen are rather strangely low in PK time rankings, considering Detroit has always been one of the most disciplined teams in the league, as well as Boston in Chara's Norris-winning year?
Thanks! Yes - this could definitely impact the results. (Not so much for Lidstrom, who finished 1st, 1st, 3rd, 4th, 4th, 5th, 6th, 11th, 17th and 21st in PK ice time during his past ten nominations, but it's a testament to his greatness that he still gets a ton of PK duty on such a disciplined team with so little PK ice time to go around).
For example, Chara's Bruins were shorthanded roughly 10% fewer times than the average team. If we assume that he played 10% more PK minutes on a team with average discipline (this is a very general assumption, of course), he would have jumped from 23rd to 15th.
How much does this affect the results overall? I'm not sure - it's hard to say. If you think that Norris candidates play on teams that are, on average, more disciplined, then it could certainly skew the average. My gut feeling is that this might sometimes move defensemen up or down a few spots, but it won't transform a 2:00 PK'er on one team into a 3:30 PK'er on another.
How much does this affect the results overall? I'm not sure - it's hard to say. If you think that Norris candidates play on teams that are, on average, more disciplined, then it could certainly skew the average. My gut feeling is that this might sometimes move defensemen up or down a few spots, but it won't transform a 2:00 PK'er on one team into a 3:30 PK'er on another.
Scott Stevens and Scott Niedermayer are definitely hurt by this metric during the dead puck era.
This is great work though, and I'm not sure if there's a better method than what you did. The only possible improvement I can think of would be to just list a team's total number of minutes shorthanded or at least rank, so we can easily put your numbers in context with a glance.
Last edited by TheDevilMadeMe: 04-04-2010 at 08:40 PM.
Success rate killing penalties would be an interesting metric when balanced against time on ice.Overall team vs specific player
Example Chris Pronger - if his team would have a higher success rate killing penalties with him in the penalty box then with him as part of the PK or with him not on the ice.
Ten of the past eleven Norris winners were at least in the top thirty in PK IT (aforementioned plus Lidstrom in 2008 and Chara in 2009).
The only exception was Niedermayer, who was 47th in PK IT in 2007. He was 2nd among defensemen in scoring (and perhaps his reputation was buoyed by his outstanding playoffs the previous spring).
I think Niedermayer appears to be an exception, but when you look at the situation he was in he actually seems to follow the same pattern as the other winners...
The Devils were the very least penalized team in the NHL the year he won the Norris, and Niedermayer actually logged the most amount of SH TOI among all Devils defensemen by a roughly 30% margin. He was also on the same team as Scott Stevens, and although Stevens only played 38 games that year, Niedermayer surely lost at least some SH time to him (Stevens averaged the most SH TOI/GP among Dmen on that team). Being second to Scott Stevens is hardly a bad thing. The combination of Niedermayer being on the least penalized team in the league that also had probably one of the top 5 best defensive defenseman in the league (Stevens) caused his ranking to be low. I have a feeling that had he been on another team his ranking would have been much higher.
This shows that, although defensemen who don't play on the penalty kill can be nominated for the Norris, trophy winners are overwhelmingly likely to be a #1 penalty killer.
By looking at your numbers quickly, it appears that even being a finalist requires elite penalty killing. Of all the years you listed, I only count 4 players that finished in the Top 3 in Norris voting to have a ranking worse than 30th...
Niedermayer 04' (who may be a victim of circumstance as previously posted)
Pronger 07' (only 66 games played)
Phaneuf 08'
Green 09' (extremely large offensive numbers too hard to ignore?)
Phaneuf is the only one that there isn't a decent explanation for. It seems that the 4th and 5th place finishers may be skewing the results a bit.
BTW, forgot to say this before, but great work, very interesting analysis.
Phaneuf is the only one that there isn't a decent explanation for.
Hype. Nothing more.
One factor to consider is how much time these defensemen themselves spend in the box. Lidstrom is almost never penalized, therefore available for almost every Detroit PK. At the other end of the spectrum, Pronger is making routine trips to the sin bin and forcing the coach to call upon others. Not sure how much this would alter the results, but it probably has some effect.
I personally think Green will win it anyways. But to kill penalties is a valuable asset as well. Keith and him will battle it out and I think Doughty will be the 3rd finalist for sure. But it should surprise no one if it goes to Green, he has still had a fabulous year and his all around game has improved
I think Niedermayer appears to be an exception, but when you look at the situation he was in he actually seems to follow the same pattern as the other winners...
Great analysis. Let me quantify your argument.
The 2003-04 Devils were a freakishly disciplined team and that really skewed his numbers. The Devils were shorthanded 232 times, 26% less than the league average of 293.
If we pro-rate Niedermayer's PK ice time to the league average, he'd jump to 299 minutes. This would rank him 13th place. This doesn't even take into account your point about how he's fighting to PK ice time with one of the best penalty killing defensemen in history. Obviously this isn't an exact since. However, given the circumstances and his excellent penalty killing in other years, I think we have to considered Niedermayer in '04 an elite penalty killer, even if his ice time in that particular season is a bit low.
Quote:
Originally Posted by JohnnyD
By looking at your numbers quickly, it appears that even being a finalist requires elite penalty killing. Of all the years you listed, I only count 4 players that finished in the Top 3 in Norris voting to have a ranking worse than 30th...
Niedermayer 04' (who may be a victim of circumstance as previously posted)
Pronger 07' (only 66 games played)
Phaneuf 08'
Green 09' (extremely large offensive numbers too hard to ignore?)
Correct - those are the only four finalists who ranked less than 30th.
Niedermayer - as discussed above, is a victim of circumstances. Very likely would have been a top 10-20 penalty killer on an average team.
Pronger - due to time missed. Ranks 30th in PK per game (min 60 games played). Even that's a really low total by Pronger's standards.
Phaneuf - ???
Green - dominant offense made up for lack off PK / defense (1st defenseman since 1993 to score 30+ goals)
I personally think Green will win it anyways. But to kill penalties is a valuable asset as well. Keith and him will battle it out and I think Doughty will be the 3rd finalist for sure. But it should surprise no one if it goes to Green, he has still had a fabulous year and his all around game has improved
Just to be clear, I'm not saying that Green, as a third-tier penalty killer, doesn't deserve to win the Norris. However, if he wins, he'll be the first defenseman to ever win the Norris while being anything less than a top thirty* penalty killer.
* Niedermayer really was a top thirty penalty killer when he won the Norris trophy, see discussion above.
If he wins, Keith would be a slightly below average penalty killer by Norris trophy standards.
I'm not saying that Green won't win - but if he does, he'll be the worse penalty killer to win the Norris, by a huge margin, over the past twelve years (i.e. since they started keeping track of ice time).
If he wins, Keith would be a slightly below average penalty killer by Norris trophy standards.
I'm not saying that Green won't win - but if he does, he'll be the worse penalty killer to win the Norris, by a huge margin, over the past twelve years (i.e. since they started keeping track of ice time).
Pronger had 22-23% more PK time available to him thanks to Hartnell, Carcillo, and co. Given Lidstrom and Keith's percentage of their respective teams' PK time and had said teams had as much total PK time as the Flyers, they would have both been top 10 in total SH TOI slotting in at 7th and 8th.
Pronger was on the ice for 55.9% of his team's SH TOI.
Keith was on the ice for 54.4% of his team's SH TOI.
Lidstrom was on the ice for 54.3% of his team's SH TOI.
Granted I haven't done the math on everyone in the league, but given where Keith and Lidstrom sit and given that both the Wings and Hawks are bottom 4 in total number of PKs they're at a pretty big disadvantage (if you want to call it that) of PK time available to them.
Nashville and New Jersey are the only teams with fewer PKs than the Wings and Hawks and their top PK guys sit at 45th and 59th in overall SH TOI respectively.
Just to compare, I looked at where the Norris winners from '68 to '97 ranked in estimated shorthanded icetime. This is not an officially tracked statistic, but is largely based on various factors like the number of PPA goals a player was on the ice for compared to his team and how much time the team spent shorthanded. The ranking is where they rated in PK icetime per game:
Year
Player
PK TOI Rank
1968
Bobby Orr
16
1969
Bobby Orr
1
1970
Bobby Orr
9
1971
Bobby Orr
4
1972
Bobby Orr
5
1973
Bobby Orr
1
1974
Bobby Orr
5
1975
Bobby Orr
4
1976
Denis Potvin
2
1977
Larry Robinson
59
1978
Denis Potvin
3
1979
Denis Potvin
1
1980
Larry Robinson
9
1981
Randy Carlyle
19
1982
Doug Wilson
44
1983
Rod Langway
4
1984
Rod Langway
8
1985
Paul Coffey
67
1986
Paul Coffey
1
1987
Ray Bourque
15
1988
Ray Bourque
17
1989
Chris Chelios
13
1990
Ray Bourque
5
1991
Ray Bourque
15
1992
Brian Leetch
34
1993
Chris Chelios
3
1994
Ray Bourque
9
1995
Paul Coffey
51
1996
Chris Chelios
2
1997
Brian Leetch
17
Mike Green finished this season tied for 104th per game, which would put him far behind all over post-expansion Norris winners.
Granted I haven't done the math on everyone in the league, but given where Keith and Lidstrom sit and given that both the Wings and Hawks are bottom 4 in total number of PKs they're at a pretty big disadvantage (if you want to call it that) of PK time available to them.
That's definitely a valid point. That's why I try to look at broad categories ("top thirty") rather than focusing too much on their exact ranking. Even then this method isn't foolproof (see the Niedermayer comment above).
Just to be clear though, I didn't say that Pronger deserved the Norris.
Quote:
Originally Posted by reckoning
Year
Player
PK TOI Rank
1977
Larry Robinson
59
1982
Doug Wilson
44
1985
Paul Coffey
67
1992
Brian Leetch
34
1995
Paul Coffey
51
Thanks for the data! Very interesting. Let's look at the players with the least PK TOI:
- Robinson's ranking seems surprisingly low. Then again, he's only a distant third among Habs defensemen in PK goals against (Lapointe and Savard are the top two). Does anybody know what happened here? It looks like Robinson was relegated to the second defense pairing.
- Not too surprising about the others. These were all dominant seasons from dominant offensive defensemen (Coffey finished 5th and 6th in scoring in '85 and '86 respectively; Wilson scored 39 goals in '82; Leetch scored 102 points in '92).
- Robinson's ranking seems surprisingly low. Then again, he's only a distant third among Habs defensemen in PK goals against (Lapointe and Savard are the top two). Does anybody know what happened here? It looks like Robinson was relegated to the second defense pairing.
I think it's just that Savard and Lapointe were the top penalty killing pair in Montreal, and had been for years. Montreal had the best PK in the league in that year with Savard and Lapointe, so there was no need to change anything. Robinson took on a larger PK role in the following years.
I don't think any defenceman today is in a similar situation, stuck behind two HHOFers. Certainly Mike Green is not. He's more similar to the other defencemen in the group as young offensive defencemen, IMO. Wilson, Coffey, and Leetch all went on to get more (estimated) time on the PK in future seasons. Will Green do the same?
Just to compare, I looked at where the Norris winners from '68 to '97 ranked in estimated shorthanded icetime. This is not an officially tracked statistic, but is largely based on various factors like the number of PPA goals a player was on the ice for compared to his team and how much time the team spent shorthanded. The ranking is where they rated in PK icetime per game:
Year
Player
PK TOI Rank
1968
Bobby Orr
16
1969
Bobby Orr
1
1970
Bobby Orr
9
1971
Bobby Orr
4
1972
Bobby Orr
5
1973
Bobby Orr
1
1974
Bobby Orr
5
1975
Bobby Orr
4
1976
Denis Potvin
2
1977
Larry Robinson
59
1978
Denis Potvin
3
1979
Denis Potvin
1
1980
Larry Robinson
9
1981
Randy Carlyle
19
1982
Doug Wilson
44
1983
Rod Langway
4
1984
Rod Langway
8
1985
Paul Coffey
67
1986
Paul Coffey
1
1987
Ray Bourque
15
1988
Ray Bourque
17
1989
Chris Chelios
13
1990
Ray Bourque
5
1991
Ray Bourque
15
1992
Brian Leetch
34
1993
Chris Chelios
3
1994
Ray Bourque
9
1995
Paul Coffey
51
1996
Chris Chelios
2
1997
Brian Leetch
17
Mike Green finished this season tied for 104th per game, which would put him far behind all over post-expansion Norris winners.
Just for fun, I tried approximating ice time for a 14-year chunk of Phil Russell's career for an ATD bio using a fairly elementary formula. Do you have a file or something that has all your 1968-1997 calculations? I'd like to see it.
Interesting numbers. Yikes, I knew Tomas Kaberle wasn't much in terms of a penalty killer or defensive play in general, but wow, 21 goals per 60 minutes short-handed? You definitely don't see any other supposedly "elite" (as Toronto fans tell us) defenseman anywhere near him on that list. That's mind-blowing.
Now the Green thing is tough to judge. It's hard to imagine a player as dynamic as him not winning at least 1 Norris in his career. Paul Coffey is not a good comparison, no matter how popular it is on this site. Coffey won the Norris because his offensive production was Orr-esque. It was too much to ignore. I know most of you already know this, but Green is not comparable to Coffey just because both they have similar flaws.
Green falls more in the Phil Housley category, with great, but not spectacular, offensive numbers. Not to mention, he plays Housley-esque defense, which usually costs him in the playoffs, thus his rather ugly -7 in just 21 playoff games. Housley never won a Norris, and his closest finish was a distant third in 1992. I could see that happening to Green, honestly, especially with such a dominant group of all-around defensemen ready to take hold of the Norris for the next decade. You have Erik Johnson, Doughty, Myers, Weber, and Suter, to begin with.
Me thinks Green got such more Norris consideration last year because it was a down year for defensemen. Lidstrom wasn't himself, and only Zdeno Chara really exerted himself as elite. Now, with Doughty, Keith, and Lidstrom all at the top of their game, I think Green will suffer. I've seen enough of him to feel bad if I gave him a vote for the Norris. He's horrifying defensively. It's not just giving the puck away, it's more the fact that he is the worst decision-maker that I have seen in a long, long time from his position. He happens to play in a system designed by a coach who thinks that he is Glen Sather, Ovechkin is Gretzky, and Green is Coffey, so if it worked for the Oilers, it has to work for the Capitals. The problem is that he doesn't have a Messier or Tikkanen to back-check, and he doesn't have Grant Fuhr to bail out the team for Green causing an odd-man rush with an ill-timed pinch or big hit.
Last edited by Blades of Glory: 04-14-2010 at 01:21 AM.
Interesting numbers. Yikes, I knew Tomas Kaberle wasn't much in terms of a penalty killer or defensive play in general, but wow, 21 goals per 60 minutes short-handed? You definitely don't see any other supposedly "elite" (as Toronto fans tell us) defenseman anywhere near him on that list. That's mind-blowing.
He played a total of ~8 min shorthanded during the whole season and was on ice for 3 GA. All those numbers say is that Ron Wilson isn't using him to kill penalties. Link: http://www.nhl.com/ice/playerstats.h...Name=timeOnIce
Not a great year for penalty-killing defensemen. Lidstrom becomes only the second Norris winner to rank less than 25th in PK TOI (Niedermayer was only 47th in PK TOI but was fighting for ice time with Stevens, and was on a freakishly disciplined team - last year's Wings were essentially average in times shorthanded, so it's not like there wasn't a lot of PK TOI to go around).
I realize I'm not saying anything new, but Visnovsky and Yandle finished in the top five due to their offense.
...Funny timing to see this thread coming up again.
Just caught a declaration of Denis Potvin -- now back in the Outaouais -- who said that Erik Karlsson shouldn't get Norris consideration due to the fact he isn't playing on the PK, or playing very few minutes (he's 6th on his team for total minutes...)
