The Best Defense is a Good Offense: the Professional Hockey Writer’s Association and 57 Years of Screwing the Pooch on the Norris Trophy
The Best Defense is a Good Offense
On Lidstrom’s retirement, I hypothesized that he might be one of the three best defencemen in NHL history. I did this based on one thing: Norris trophy wins.
But the more I thought of it, the more I thought that was ridiculous. The PHWA had just awarded two absolutely brutal Norris trophies: the most recent to Karlsson, and the previous year to Lidstrom. The first was a clear case of “I didn’t watch enough hockey this year but this guy scores a lot”-itis, which seems to infect the majority of PHWA voters from time to time, and the latter was a case of the “well he has had an amazing career and he might never win another” flu. I realized I had never seen Doug Harvey play and couldn’t possibly believe deep down that the three best D of all time were Orr, Harvey and Lidstrom in that order.
So I decided to figure it out for myself. Jump down to the results.
Defensive Point Shares (DPS)
For those of you who don’t know, point shares are a (very) rough attempt at calculating the value of a player for his particular team. The concept stems from Bill James’ idea of a win share.
In baseball, this is a relatively easy task, as there are encounters between the offensive player and the primary defensive player many times a game, and these include so-called “true outcomes” where only the catcher is also involved in the play.
In basketball, it is definitely harder to figure out worth, but there are still clear differences in the value of players between all-stars and 12th men, especially because the NBA has tracked minutes-played for decades.
In the NHL it is harder to figure out worth for a few reasons.
- First, as with baseball, there are two distinct types of positions, where one player plays most or all of a game (the starting pitcher, the goalie) and the others are not involved as much (the fielders, the skaters).
- Second, unlike basketball, no skaters play remotely close to a full game. Example: when Chris Pronger won his MVP, he was playing half the time every game, which is unbelievably high for a skater, and that season is an outlier in terms of ice-time norms.
- Finally, the NHL gives out points for wins and for non-wins, unlike the other leagues so you can’t just calculate win shares.
So for hockey the concept of a point-share was developed: how much a given player contributes to the points his team won in a given season. This has been further broken down into goalie shares, offensive point shares and defensive point shares:
- Goalie point shares are relatively straightforward (especially since shots were tracked since the ’60s).
- So are offensive point shares, since they are based on what is called “goals created,” a metric that attempts to calculate a player’s offensive input better than goals + assists. (Aside: the problem with goals + assists is manifold: assists are sometimes given when they are not deserved and sometimes not given when they are deserved; some players score a lot of goals by standing still and waiting for someone to give them the puck; etc.)
- Defensive point shares are more difficult. They are calculated in four different ways:
- the current way,
- the method prior to ice-time being recorded,
- the method prior to shots on goal being recorded,
- and finally, the least reliable: the method prior to +/- being recorded. Obviously, the farther back in time we go, the less reliable DPS are.
Ideally, a Defensive Point Share measures the defensive contribution of the skater by taking into account
- their share of ice-time,
- their proportion of marginal goals against,
- their “position adjustment,”
- and their +/- adjustment (as +/- is not a 5-man stat).
For me, it still isn’t perfect as it doesn’t take into account things like
- blocked shots (which are only relevant if we also know blocked shot attempts),
- hits that ended rushes,
- delay of game penalties and other penalties that penalize the team,
- and other such things, most of which have only been tracked by the NHL in recent years, and are not not uniformly recorded by the various hockey stat sites.
But Defensive Point Shares are still better than the whims of the voters, as you will see below.
Problems with Defensive Point Shares
The obvious advantage of DPS is that they allow us to assign a number to an otherwise numerically ill-defined position and say “This is what Player A contributed to The Team last season.” It makes it easier for us to look at a player’s stats and judge his contribution without watching every single hockey game he plays, something that is harder than ever given that there are 30 teams in the NHL and 82 games in a season.
The only way we can otherwise judge these players is by watching a lot of video. And frankly none of us can do that with Doug Harvey, unless we have a lot of free time and a lot of money.
There are a number of very big problems on the other hand.
- DPS is an estimation of estimations: All point shares are estimations, but DPS are particularly dependent on educated guesses – such as goals created – especially prior to the tracking of ice-time. These guesses are still better guesses than looking at a defenceman’s total points, for example, but they are guesses nonetheless.
- DPS are calculated per season: The issue with this is that all players who do not manage to play a full season are penalized. Obviously you and I can calculate Defensive Point Shares Per Game (DPSPG) but this has to be done for each individual player, as no one that I know of has done this for every player in NHL history in an online searchable database yet. So if you look up DPS on hockey-reference.com, the results are skewed to those who played more games.
- DPS are team-centric: One huge problem is that in order to accumulate DPS, a player has to be on a non-losing team (not necessarily a winning team).
- The more the team accumulates points (through wins or OT losses) the more DPS a player can get. The problem is that a given player could be very, very good but the rest of his team could suck. DPS obviously underrate those players. (Though, with ice-time, you can see that it does somewhat reward major defensive contributors on bad teams, as there are a number of D from bad teams appearing in the top DPS-getters in seasons since ice-time was calculated. See the chart below.)
- The farther back in time we go, the more team-centric DPS get. So you will see some teams dominate the top 10 DPS especially prior to the introduction of +/-. This is obviously not accurate, but the only cure to it I know is to do way more research than I currently have the resources for.
- DPS is a measure of value, not talent: Another issue is that DPS measure the overall contribution to the team.
- If a player doesn’t play as much for matchup reasons, or because the coach isn’t getting along with him, or because they don’t have a good enough point shot (but have every other skill you need to play D), that will be reflected in their final value. That is to say the player with the most DPS in a season (more accurately, the player with the most DPSPG in a season) is the most valuable defenceman and not necessarily the best. The NHL, and all other leagues with MVP awards, faces this problem annually when it tries to award the Hart and the Conn Smythe.
- This is further exacerbated when we don’t know ice-time. A perfect example of this is the perceived value of Brian Engblom and Rod Langway as compared to Larry Robinson for the late ’70s / early ’80s Habs. If ice-time were calculated back then, we can expect that Robinson would probably have considerably more DPS (and therefore more PS) than at least one of those two (whoever wasn’t playing with him that season). But because we don’t have ice-time, we get a result that Robinson’s contributions are considered relatively equal on the defensive end, at least some years, despite his far greater offensive contributions. Without ice time, contributions are assessed less fairly because their ice-time is assumed to be equal. And obviously there is no way that all three players played the exact same amount of ice-time every year they were together on the Habs, even though their DPS suggest that they did.
- If we go even farther back in time, we see teams with 5 defencemen in the top 20 DPS some seasons, something that is clearly absurd and impossible. (Unless, of course, some team was playing all six D 10 minutes per game, and I think we can agree on the unlikeliness of that)..
But despite all these problems, it still remains the best method we have for evaluating the most valuable – not the best – defenceman in a given year. So it seemed like a good idea to me to look at Norris trophy wins as they relate to DPS since it was first awarded in 1953-54.
A few notes before I present the data
I have included the few years prior to 1954 so you can see why Red Kelly was given the first Norris even though he might not have deserved it (depending on which metric you judge him by). The award was created for him because, in the early ’50s, he was far and away the best defenceman in the league.
- I have included all leaders each year:
- for goals,
- and +/- and ice-time when available,
- in order to show two things:
- First, that DPSPG and offensive ability regularly don’t correlate at all
- and, second, that offensive contributions often skew total point shares way too heavily in favour of offensive players.
- I have calculated DPSPG for players who made it onto my list for each year.
- I have somewhat jokingly handed out the George Boucher Award – named after the career leader in DPSPG prior to the introduction of the Norris, see the spreadsheet – to the player that leads the league in DPSPG. These numbers are obviously not to be wholly believed, as somebody might have a higher DPSPG, but only played 2/3rds of a season and so doesn’t appear on the DPS leader board which I used to create my data set.
- I have judged the defencemen by two ratings: PS rating and Weighted Point Shares Per Game:
The PS Rating is a simple rating that evaluates where the player finished on the leader boards for DPS and PS that year. I chose DPS and PS, rather than OPS, because we are awarding trophies for defense, not offense or all-around play.
The PS Rating essentially counts DPS’ rank 1 1/2 times and OPS’ rank half a time. 2 is perfect (1st in DPS and 1st in PS) and anything higher than 2 should be judged as to how far away from 2 it is. (I could have divided all numbers by 2, so that 1 is perfect but that is something nitpicky that I can’t be bothered with. Feel free to do so yourself.)
The reason why PS Rating is good is because it is relative to the year. It takes no account of variations in scoring from era to era and it somewhat nullifies the skewing of OPS that occurs especially in the 1970s, 80s and early 90s.
The problem with PS Rating is that it is not per game and so greatly disfavours those with injuries and / or suspensions.
Weighted Point Shares Per Game (WPSPG)
Weighted Point Shares Per Game are exactly what the name says; calculated like so:
(DPS + (PS/2)) / games played.
The overall point share is treated like the OPS is treated when calculating PS. This somewhat skews the OPS effect that overvalues offense in the ’70s, ’80s and ’90s. But arguably it does not do so enough during the ’80s and early ’90s. If some kind of adjusted PS data was available based on the adjusted goals created of each player, then this would be a more accurate process.
The valuable thing about WPSPG is that it is per game, and that it is weighted in favour of defense, if only slightly.
The problem is that it is not weighted in favour of defense enough. Even with the weighting, Paul Coffey still comes out on top in WPSPG for his absurd 1985 season – where, by DPS, he was the 18th most valuable D in the league and so it makes me cringe that he could win the Norris – and Mike Green still comes out on top in 2009, despite the utter lack of defense played by him that season.
The only way around the problems of both methods – that I can see – is to calculate a PS Rating that takes into account a weighing of DPSPG and PSPG rankings. The problem with that is that I would have to calculate DPSPG and PSPG for much of the league, to be truly fair. And that’s something I don’t have the resources for at the moment.
So, without any further delays, here is the spreadsheet. If this is hard to read, you can see the whole thing in your browser here.
Note: all stats are from hockey-reference.com and those I calculated myself are based solely on their data.
Interpreting the data
Getting it Right: The PHWA Picks the Most Valuable/Best D as Norris Trophy Winner
The first thing we can glean from this list is that the PHWA gets it wrong a lot of the time. If we combine both rating methods, the PS Rating and WPSPG, and say that whenever they agree the player absolutely won the Norris, the PHWA got it right 24 times in 57 years, or 42% of the time. This is principally due to the ’80s and ’90s, where they got it wrong most years.
That wouldn’t be so bad, I guess, if we were talking about Pacific Coast League batting averages, but we are talking about the Professional Hockey Writers Association’s assessment of the best hockey defenceman of the year. This is their job. And frankly, 42% is pretty bad – though not nearly as bad as I thought it would be when I started working on this. Not even half of the time did they choose the player who should have been the consensus most valuable D.
Here is what they got absolutely right:
- Doug Harvey in ’56, ’57, ’58
- Tom Johnson in ’59
- Pierre Pilote in ’64 and ’65
- Bobby Orr in ’69, ’70, ’71, ’72, ’74 and ’75
- Larry Robinson in ’77
- Denis Potvin in ’79
- Ray Bourque in ’88 and ’90
- Chris Chelios in ’89
- Paul Coffey in ’95
- Al Macinnis in ’99
- Chris Pronger in ’00
- Nick Lidstrom in ’03, ’07 and ’08
- Scott Neidermayer in ’04
We can say that those are our sure-thing seasons, where there is absolutely no controversy: Both metrics I am using chose the player and so did the PHWA. Every single one of the above trophies is absolutely deserved and, in my view, close to unimpeachable when talking about Hockey Hall of Fame admissions, greatest players of all-time and the like. (How we weigh those wins in those discussions is another question for another article.)
Getting it somewhat right: The PHWA Picked Either The Most Valuable or the Best D when it Awarded the Norris Trophy
Additionally, the PHWA picked the Norris winner chosen by at least one of our rating methods a further 8 times. Which, if we’re feeling charitable, brings their percentage up above 50%, to 56%, which is still a (Canadian) D grade (appropriate, I know), though it is a pass (in Canada).
Those trophies were:
- Red Kelly in ’54 (by WPSPG)
- Bobby Orr in ’68 and ’73 (WPSPG)
- Denis Potvin in ’78 (WPSPG)
- Paul Coffey in ’85 (WPSPG)
- Ray Bourque in ’87, ’91 and ’94 (PS Rating in ’87 and ’91, WPSPG in ’94)
- Brian Leetch in ’92 (WPSPG)
You can see my concern over the WPSPG rating system, as every single one of those players was helped over the hump by their outstanding OPS.
So we are already witnessing the skewing that comes with OPS even when using weighted shares. If we further skewed for DPS, it is likely that that at many of these players – especially Coffey, Bourque and Leetch – wouldn’t be chosen by our metric.
A good try: The PHWA Awarded the Norris Trophy to a Top D, Just Not the Best
The PHWA picked one of the top candidates in the league – in this case “Top” meaning someone who fell within the top 3 D of the league before the expansion, and top 5 afterwards – a further 17 times.
Those trophies are:
- Harvey in ’60 and ’61 (by WPSPG)
- Pilote in ’63 (WPSPG)
- Laperierre in ’66 (WPSPG)
- Howell in ’67 (WPSPG)
- Potvin in ’76 (WPSPG)
- Robinson in ’80 (both methods)
- Langway in ’84 (both)
- Coffey in ’86 (both)
- Chelios in ’93 and ’96 (both in ’93, PS Rating only in ’96)
- Leetch in ’97 (both)
- Lidstrom in ’01, ’02 and ’06 (both)
- Chara in ’09 (both)
- Keith in ’10 (both)
That’s 38 we’re at now. We could add this to their success percentage, but I don’t go for that. They should get it right, not close.
Why should they get it right? Because this stuff matters. This is what many players have as their legacy. Especially with defencemen, it is impossible to accurately judge a player without seeing him play. The Norris trophy is the main way we do this years after the fact, since most of us don’t have the luxury to watch video of the ’61 Habs all day long.
So the above 13 players were reasonably okay choices, but more deserving players were left off, and their legacies are suffering as a result.
More on that in a second.
Screwing the Pooch: The PHWA Awards the Norris to Someone Who Doesn’t Deserve It
So that brings us to the remaining 8 trophies: where the PHWA completely screwed the pooch on the award:
- Harvey in ’55 and ’62
- Carlyle in ’81
- Wilson in ’82
- Langway in ’83
- Rob Blake in ’98
- Lidstrom in ’11
- Karlsson in ’12
In all of the above cases, there were not just a couple better choices but many better choices. In fact, in most of the above cases there were numerous better choices. Certainly an interesting article could be written on the how and why the of the PHWA picking such terrible candidates and undeserving players as Randy Carlyle and Rob Blake – the only players prior to Lidstrom in ’11 to win the Norris with a minus on the year – but today we are more concerned with who they should have picked.
Re-Awarding the Norris Trophy
No-doubt Norris Trophy Seasons
In addition to the 24 seasons in which the PHWA actually awarded the Norris trophy correctly – at least in terms of the most valuable defenceman, based on both PS Rating and WPSPG – we can also pinpoint the years when there was a consensus candidate that wins both the PS Rating and WSPSPG, and therefore who should have been awarded the Norris, at least by our method.
Those players are:
- Red Kelly in ’55 (.16 WPSPG, 2 PS Rating) instead of Doug Harvey (.14 WPSPG, 10 PS Rating)
- Pierre Pilote in ’60 (.13 WPSPG, 2 PS Rating) instead of Harvey (.12 WPSPG, 12 PS Rating)
- Pilote again in ’61 (.14 WPSPG, 3 PS Rating) instead of Harvey (.13 WPSPG, 17 PS Rating)
- JG Talbot in ’62 (.15 WPSPG, 2 PS Rating) instead of Harvey (.11 WPSPG, 20 PS Rating)
- Tim Horton in ’63 (.12 WPSPG, 2 PS Rating) instead of Pilote (.12 WPSPG, 24 PS Rating)
- Guy Lapointe in ’76 (.19 WPSPG, 3 PS Rating) instead of Denis Potvin (.16 WPSPG, 17 PS Rating)
- Ray Bourque in ’84 (.19 WPSPG, 2 PS Rating) instead of Rod Langway (.13 WPSPG, 17 PS rating)
- Marke Howe in ’86 (.21 WPSPG, 4 PS Rating) instead of Paul Coffey (.17 WPSPG, 9 PS Rating)
- Bourque again in ’93 (.16 WPSPG, 4 PS Rating) instead of Chris Chelios (.14 WPSPG, 6 PS Rating)
- Vladimir Konstantinov in ’97 (.15 WPSPG, 6 PS Rating) instead of Brian Leetch (.15 WPSPG, 9 PS Rating)
- Sergei Zubov in ’01 (.151 WPSPG, 3 PS Rating) instead of Nicklas Lidstrom (.145 WPSPG, 4 PS Rating)
- Rob Blake in ’02 (.17 WPSPG, 4 PS Rating) instead of Lidstrom (.14 WPSPG, 12 PS Rating)
- Lubormir Visnovsky in ’11 (.151 WPSPG, 8 PS Rating) instead of Lidstrom (.098 WPSPG, 82 PS Rating)
- Shea Weber in ’12 (.152 WPSPG, 7 PS Rating) instead of Erik Karlsson (.136 WPSPG, 29 PS Rating)
Some of these are indeed close but if we are to take the metrics seriously then we now have 38 Norris trophy winners with another 19 to sort out.
So what do we do about the other 19 years?
Norris Wins by PS Rating vs. Norris Wins by WPSPG
With 38 trophies accounted for, we can see how differently hockey history would be if the PHWA had picked the relatively best defenceman each year (the PS Rating) or had picked the most valuable defenceman by (weighted) point shares.
I have re-embedded my spreadsheet here so you don’t have to scroll up.
Obviously those are two very different lists since there are 19 trophies different from one to the other.
For example, by one metric Tim Horton has four trophies, instead of 0, and by another he has 1 instead of 0. Either makes a big difference in terms of how people might think of him now.
Those lists include all 38 trophies where the metrics agree, believe it or not.
Can we figure out what to do with those 19 trophies where the metrics don’t agree so we don’t have two such divergent lists?
PS Rating Vs. WPSPG
Both ratings have their drawbacks but I think we can look at each season without a definitive candidate and, using our judgment decide which we should prefer, the PS Rating or the WPSPG; rather than just claiming that one works for every season without a consensus candidate.
So now we look at the player with the best PS Rating and the highest WPSPG per season and figure out who should win between the two.
Please note that all rankings are relative only to those players I calculated the rankings for and therefore it is slightly possible that a more deserving player who played fewer games slipped through the cracks.
1954 Norris Trophy: Red Kelly
- Tim Horton: 4 PS Rating (1st), .148 WPSPG (2nd)
- Red Kelly (Norris winner): 10 PS Rating (tied for 3rd), .173 WPSPG (1st)
Horton wins the Boucher award for this year but only leads Kelly by a lot of DPS because Kelly played significantly fewer games. Whereas Kelly has a significantly better .173 WPSPG.
We give it to Kelly, so they did get it right this time.
1966 Norris Trophy: JC Tremblay
- Tim Horton: 5 PS Rating (1st), .116 WPSPG (4th)
- Jacques Laperriere (Norris winner): 13 PS Rating (6th), .131 WPSPG (2nd)
- JC Tremblay: 9 PS Rating (2nd), .131 WPSPG (1st)
This is an easy one for me: Tremblay.
1967 Norris Trophy: Pierre Pilote
- Harry Howell (Norris winner): 7 PS Rating (4th), .127 WPSPG (3rd)
- Doug Mohns: 9 PS Rating (3rd), .165 WPSPG (1st)
- Pierre Pilote: 3 PS Rating (1st), .139 WPSPG (2nd)
Though Mohns is clearly way ahead in WPSPG his numbers are pretty inflated by his large OPS (for the era). This is one that could be seriously helped by ice-time or even +/-.
If we do it by adding the rankings, Pilote gets it.
1968 Norris Trophy: JC Tremblay
- Bobby Orr (Norris winner): 40 PS Rating (14th at the very best), .153 WPSPG (1st)
- JC Tremblay: 5 PS Rating (1st), .141 (3rd)
Orr’s terrible PS Rating is a result of his only playing 46 games due to injuries. But Orr’s DPSPG is significantly weaker than Tremblay’s and it is only his very large OPS that inflates his WPSPG past Tremblay’s.
Though I firmly believe Orr is the best hockey player in history, I think he did not deserve the Norris this year.
1973 Norris Trophy: Bobby Orr
- Bobby Orr (Norris winner): 7 PS Rating (tied for 2nd), .22 WPSPG (1st)
- Serge Savard: 4 PS Rating (1st), .165 WPSPG (2nd)
Orr played 11 fewer games, had a barely lower DPSPG and absolutely killed Savard on the PSPG and WPSPG. Yes, it’s because of his offence but it’s not like the Bruins were a bad defensive team that year and its not like Orr’s DPSPG is bad in any way shape or form.
1978 Norris Trophy: Larry Robinson
- Brad Park: 5 PS Rating (2nd), .174 WPSPG (2nd)
- Denis Potvin (Norris winner): 8 PS Rating (3rd?), .184 WPSPG (1st)
- Larry Robinson: 4 PS Rating (1st), .173 WPSPG (3rd)
I added Park just to show how wide open this year was. Cases could easily be made for all three players. Since we are trying to pick a player with either the best PS Rating of his year, or the best WPSPG of his year, Park is automatically excluded, but it’s hard to look at those numbers and say he doesn’t belong on the list.
My choice is Robinson, who was on the best defensive team, by far (we’re talking 30-40 fewer goals that season).
1980 Norris Trophy: Jim Schoenfeld
- Ray Bourque: 3 PS Rating (1st), .1625 WPSPG (2nd)
- Larry Robinson (Norris winner): 9 PS Rating (4th), .161 WPSPG (3rd)
- Jim Schoenfeld: 5 PS Rating (2nd), .163 WPSPG (1st)
This is a really close one.
It brings up a topic I keep alluding to as well: how major awards shape our memory. Most of you are probably thinking “who’s Jim Schoenfeld?” I know I was. Well, he was one of the top 2 D in 1980, that’s who he was. (Provided, of course, he was playing a lot of minutes.) And I’m going to give it to him because he played for the best defensive team that year. Again, 30-40 goals fewer than his competitors’ teams. He also has the hands down best DPSPG at .1 (a rare thing, especially in the 80s).
1981 Norris Trophy: Larry Robinson
- Randy Carlyle (Norris winner): 88 PS Rating (19th, at best), .095 WPSPG (20th, at best)
- Rod Langway: 5 PS Rating (1st), .152 WPSPG (2nd)
- Larry Robinson: 11 PS Rating (3rd?), .165 WPSPG
Carlyle’s Norris is probably the most egregious mistake the PHWA ever made when awarding the trophy. There is no justification for it. If you look throughout the spreadsheet you’ll see how his advanced numbers aren’t just embarrassing within 1981, but are embarrassing throughout the history of the awarding of the Norris trophy. I don’t think a single other player won the Norris with such terrible defensive credentials.
Oh wait, Rob Blake did in 1998. Oops.
Anyway, Carlyle’s ’81 Norris Trophy is still a giant embarrassment to the PHWA. So the decision between the two Habs comes down to ice-time, which we cannot know. I give it to Robinson for the completely subjective reason that I think he was the better player. Kidding.
The actual reason is that we know Langway was not an offensive player. Moreover, we can at least assume that Langway and Robinson played together a lot this year and can also assume that Langway’s better than normal OPS came from playing with Robinson. (This regularly happens to defensive defencemen throughout NHL history.) I do have some regrets about giving it to a guy with over a 10 PS Rating on the other hand…
1982 Norris Trophy: Larry Robinson
- Rod Langway: 12 PS Rating (3rd?), .1667 WPSPG (1st)
- Larry Robinson: 6 PS Rating (1st), .163 WPSPG (2nd)
- Doug Wilson (Norris winner): 60 PS Rating (21st at best), .113 WPSPG (15th at best)
Wilson’s absolute advanced numbers are less embarrassing than Carlyle’s but his relative rankings are just as pathetic. “He scored 39 goals, he clearly was a great defensive player” runs the logic. The same logic that gave Karlsson the trophy this year. The only people who can vote that way must not watch a lot of hockey.
Again I am inclined to give it to Robinson given the closeness of the two Habs’ WPSPGs. Incidentally, Langway’s two Norris trophies can be seen as makeups by the PHWA after not giving it to him in ’81 or ’82, when he might have actually deserved it, for all I know.
1983 Norris Trophy: Ray Bourque
- Ray Bourque: 6 PS Rating (2nd), .183 WPSPG (1st)
- Rod Langway (Norris winner): 68 PS Rating (21st, at best), .9 WPSPG (23rd, at best)
- Mike O’Connell: 5 PS Rating (1st), .114 WPSPG (9th at best)
Langway’s WPSPG is the worst for a Norris winner for the seasons I calculated it. This award was clearly a mulligan. (To be fair to Langway, he was suddenly on a not so great team.)
Bourque wins this handily, having a far superior WPSPG to his teammate. (We can make the Robinson-Langway argument here again for Bourque and O’Connell.)
1985 Norris Trophy: Mark Howe
- Ray Bourque: 11 PS Rating (3rd), .1568 WPSPG (2nd)
- Paul Coffey (Norris winner): 19 PS Rating (8th, at best), .1588 WPSPG (1st)
- Mark Howe: 10 PS Rating (tied 1st), .1493 WPSPG (3rd)
- Larry Robinson: 10 PS Rating (tied 1st), .1467 WPSPG (6th, at best)
Coffey’s WPSPG is highly inflated by the offense-only team he played for.
Bourque also played for a team that wasn’t exactly defense-first that year.
Though the field is as wide open as any other year, I like Howe for it. He played for the best defensive team of the bunch and, if you combine the rankings, he easily comes out on top.
1987 Norris Trophy: Mark Howe
- Ray Bourque (Norris winner): 3 PS Rating (1st), .1686 WPSPG (2nd, at best)
- Mark Howe: 4 PS Rating (2nd), .1703 WPSPG (1st)
This one is extraordinarily close and, in the absence of ice-time, can again only be decided by looking at who played for the better defensive team.
And that was Howe.
1991 Norris Trophy: Ray Bourque
- Ray Bourque (Norris winner): 4 PS Rating (1st), .1605 WPSPG (2nd)
- Al MacInnis: 5 PS Rating (2nd), .1609 WPSPG (1st)
This is an easy one for me. MacInnis had one of those absurd offensive seasons that so few defencemen have had, one all but Orr and Potvin never had during the crazy ’80s and early ’90s. It skews his numbers.
Bourque, hands down.
1992 Norris Trophy: Chris Chelios
- Chris Chelios: 7 PS Rating (1st), .1413 WPSPG (2nd)
- Brian Leetch (Norris winner): 14 PS Rating (3rd, at best), .1463 WPSPG (1st)
Again, an absurd offensive season skewers the results.
1994 Norris Trophy: Scott Stevens
- Ray Bourque (Norris winner): 10 PS Rating (3rd), .173 WPSPG (1st)
- Scott Stevens: 5 PS Rating (1st), .1627 WPSPG (2nd)
I like Stevens for this. Played for the better defensive team. Again, ice-time would have helped make the decision clearer.
Combining the metrics gives it to Stevens.
1996 Norris Trophy: Vladimir Konstantinov
- Chris Chelios (Norris winner): 19 PS Rating (7th, at best), .1389 WPSPG (7th, at best)
- Vladimir Konstaninov: 7 PS Rating (2nd), .1735 WPSPG (1st)
- Nicklas Lidstrom: 4 PS Rating (1st), .1556 WPSPG (3rd)
Yet another trophy decision that would be significantly aided by the tracking of ice-time.
I am inclined to give it to Konstaninov on numbers and hearsay, and Lidstrom based on the fact that I have at least watched Lidstrom play a ton and know how good he was.
I guess I’ll give it to Konstantinov without the ice-time to confirm the rightness of my decision.
1998 Norris Trophy: Larry Murphy
- Rob Blake (Norris winner): 89 PS Rating (20th, at the very best), .0914 WPSPG (20th, at the very best)
- Larry Murphy: 4 PS Rating (1st), .1494 WPSPG (2nd)
- Chris Pronger: 6 PS Rating (2nd), .1537 WPSPG (1st)
This is, along with Carlyle’s, Wilson’s and possibly Lidstrom’s’ 11th trophies, one of all the time mistakes for the awarding of the Norris.
Between Pronger and Murphy, Murphy was on the better defensive team.
I give it to Murphy.
2006 Norris Trophy: Nicklas Lidstrom
- Nicklas Lidstrom (Norris winner): 5 PS Rating (2nd), .1569 WPSPG (3rd)
- Mathiue Schneider: 7 PS Rating (3rd), .1674 WPSPG (1st)
- Sergei Zubov: 4 PS Rating (1st), .1615 WPSPG (2nd)
It seems like a crime to give Schneider the Norris over Lidstrom, when Lidstrom played way more than he did. (But this definitely raises questions as to how DPS should be calculated.)
Zubov should have it with the numbers only but Lidstrom did play for the better defensive team.
I give it to Lidstrom, but I am slightly concerned that I am doing so out of favouritism.
2009 Norris Trophy: Nicklas Lidstrom
- Zdeno Chara (Norris winner): 13 PS Rating (4th), .1369 WPSPG (5th, at best)
- Mike Green: 16 PS Rating (tied for 5th), .1743 WPSPG (1st)
- Duncan Keith: 10 PS Rating (3rd), .1429 WPSPG (4th)
- Nicklas Lidstrom: 8 PS Rating (2nd), .1449 WPSPG (2nd)
- Dennis Wideman: 7 PS Rating (1st), .1437 WPSPG (3rd)
This is another wide-open one.
I’m going to bring subjectivity into the fray for a moment: Those of us who have seen Wideman play all know that it is because he was playing with Chara this year that his advanced numbers are great. If we weigh both rankings he and Lidstrom come out on top together, and everyone knows that Lidstrom is a million times better as a player than Wideman . (If “everybody knows” isn’t good enough for you, please compare Wideman and Lidstrom on hockey-reference.com.)
Lidstrom gets it.
The New Norris Trophy “Count”
So now we have a completely new list of Norris Trophy winners:
Why does this matter?
So what? Why does anyone outside of the hockey geek community care who should have won the Norris (or any other award for that matter)?
Because the trophy wins shape our collective memory:
- Doug Harvey is immortalized as one of the three greatest D in NHL history, when the reality might have been that he was one of the greatest of his era, and that’s it. (Provided, of course, we can take the concept of point shares – especially point shares prior to tracking of ice-time – seriously.)
- Looking at this new list, we can make a far more sense of Mark Howe’s induction into the Hockey Hall of Fame last year then many were able to make at the time.
- And looking at this list I will be able to make a better case for including Zubov – a regular target among HHOF message board debates…because he’s Russian? – in a future Hall of Fame class.
- Larry Robinson seems to be closer to his rightful place in history, too.
This kind of revisionist thinking is sadly necessary when it comes to NHL awards – especially defence-based awards like the Norris and the Selke – because frankly the voters can’t be trusted to make the right decision. Whether it’s because it’s a vote, because there isn’t enough time in the day for any one person to watch all the players all the time, or for some other reason, trophies are routinely awarded to undeserving candidates. And not just to the wrong top candidate, but candidates who should have never been in the conversation in the first place.
When I eventually write my book, I plan to apply logic similar to that above to all seasons in NHL history, to look at who should have won the Norris – and all the other major awards – throughout, even before they were invented. The subsequent revised list will hopefully help us better understand hockey history without trying to find video of NHL games from 1943.
What does this mean for Lidstrom?
So I started this whole thing by referring to a post of mine where I claimed Lidstrom was one of the Top 3 D ever. This list claims he is tied for that honor, at least if Norris trophies is how we judge it (instead of, say, career PS Rating, or career WPSPG).
But I actually think this new list makes his case even stronger and that – if going by these adjusted Norris trophies alone (I will call mine the “adjusted” for the moment) – he is 2nd only to Bobby Orr. Allow me to explain.
Our metrics bumped Bourque’s count up by one, while knocking both Lidstrom and Harvey down a few pegs. On the surface, maybe that makes it look like Bourque has been elevated and Lidstrom demoted. I beg to differ.
The Norris is a relative award like all season-end awards. The awarding of a a trophy does not immediately make a claim for the person as “among the all-time elite” if, for example, everyone else was terrible that year. Bourque amassed his real and “adjusted” Norris trophies in the highest scoring (i.e. least defensively competent) era in NHL history
I’m not sure that 6 deserved trophies for this era really adds to any argument claiming Bourque’s status as the second greatest D ever. Sure, Bourque was great – perhaps even greater than we give him credit for – but his competition wasn’t exactly on the level. The same reasoning applies to Mark Howe’s 3 adjusted-trophies.
A similar argument can be made against Pilote – Lidstrom’s other main competition – and I suppose against Harvey, if we were willing to grant the merit of his 7 actual Norris trophies. Pilote deserved 5 adjusted-trophies in a six team league. He had way, way less competition than Lidstrom and, though he played in in a relatively low-scoring era, it was only relatively low scoring to the expansion era that followed it (and not compared to the era before).
Lidstrom played much of his career in one of the lowest scoring eras in NHL history and with more competition for Norris trophies than at any other time in league history. For me, his deserved 5 adjusted-trophies are far more impressive than Bourque’s 6, Pilote’s 5, and even Robinson’s 4.
Originally published on rileyhaas.com in 2012.