Statistical view of Lidstrom's last few seasons

 Uncategorized  Comments Off on Statistical view of Lidstrom's last few seasons
May 312012

Lidstrom has officially announced his retirement from the NHL ending what was one of the best, if not the best, NHL career by an NHL defenseman, and for that matter one of the best careers of any NHL player.  Bobby Orr may have been a better and more dynamic defenseman in his prime but Lidstrom was the dominant defenseman during his time in the NHL and had the longevity that Bobby Orr never had.  In fact, Lidstrom was still a dominant defenseman in the NHL into his 40’s.  Since I have the data readily available, let’s take a look at Lidstrom’s impact on the Red Wings the past 5 seasons, or ages 37-41.

I believe the best way to evaluate a player is that players impact on his teammates, often called WOWY (with or without you).  Here are Lidstroms teammates WOWY goals for percentage for each player who played 300 minutes of 5v5 zone start adjusted ice time with Lidstrom over the past 5 seasons.


Player Season w/ Lidstrom wo/ Lidstrom w/ – wo/
White 2011-12 62.9% 52.5% 10.4%
Howard 2011-12 60.6% 61.1% -0.5%
Datsyuk 2011-12 67.6% 61.3% 6.3%
Franzen 2011-12 78.8% 55.8% 23.0%
Bertuzzi 2011-12 68.8% 61.1% 7.7%
Zetterberg 2011-12 58.8% 57.5% 1.3%
Filppula 2011-12 54.5% 61.0% -6.5%
Average 2011-12 64.6% 58.6% 6.0%


Player Season w/ Lidstrom wo/ Lidstrom w/ – wo/
Howard 2010-11 56.4% 48.4% 8.0%
Stuart 2010-11 53.2% 50.0% 3.2%
Zetterberg 2010-11 50.0% 48.2% 1.8%
Franzen 2010-11 56.7% 54.7% 2.0%
Bertuzzi 2010-11 50.0% 51.5% -1.5%
Datsyuk 2010-11 55.8% 58.1% -2.3%
Holmstrom 2010-11 52.0% 45.7% 6.3%
Average 2010-11 53.4% 50.9% 2.5%


Player Season w/ Lidstrom wo/ Lidstrom w/ – wo/
Howard 2009-10 58.3% 47.2% 11.1%
Rafalski 2009-10 57.9% 53.6% 4.3%
Datsyuk 2009-10 63.8% 52.5% 11.3%
Zetterberg 2009-10 55.1% 50.0% 5.1%
Bertuzzi 2009-10 59.0% 39.6% 19.4%
Holmstrom 2009-10 63.3% 50.0% 13.3%
Osgood 2009-10 45.8% 32.5% 13.3%
Cleary 2009-10 51.6% 45.2% 6.4%
Average 2009-10 56.9% 46.3% 10.5%


Player Season w/ Lidstrom wo/ Lidstrom w/ – wo/
Rafalski 2008-09 63.5% 50.0% 13.5%
Osgood 2008-09 62.0% 46.4% 15.6%
Datsyuk 2008-09 58.8% 73.3% -14.5%
Conklin 2008-09 61.5% 57.6% 3.9%
Zetterberg 2008-09 55.6% 48.8% 6.8%
Hossa 2008-09 68.4% 66.7% 1.7%
Cleary 2008-09 48.4% 45.0% 3.4%
Franzen 2008-09 73.7% 60.4% 13.3%
Average 2008-09 61.5% 56.0% 5.5%


Player Season w/ Lidstrom wo/ Lidstrom w/ – wo/
Rafalski 2007-08 69.6% 50.0% 19.6%
Datsyuk 2007-08 77.6% 58.5% 19.1%
Osgood 2007-08 72.2% 48.6% 23.6%
Zetterberg 2007-08 69.8% 58.3% 11.5%
Hasek 2007-08 73.0% 58.3% 14.7%
Holmstrom 2007-08 78.1% 48.3% 29.8%
Average 2007-08 73.4% 53.7% 19.7%

As you can see, almost every player Lidstrom has played significant minutes with over the past 5 seasons had better results when Lidstrom was on the ice with them than when he wasn’t.  Of the 36 player seasons listed above, only 5 times did a player peform better without Lidstrom than with and only twice could you call the difference substantial (Filppula’s 6.5% in 2011-12 and Datsyuk’s 14.5% in 2008-09).  To me that is a sign of greatness, especially considering a lot of the players Lidstrom plays with are great players in their own right.

Regular readers of are probably aware of my rating system that takes into account on-ice results, quality of teammates and quality of competition and takes into account all of the information shown in the tables above.  My ratings are HARO+ for offensive rating, HARD+ for defensive rating and HART+ for total overall rating.  Here are Lidstrom’s ratings over the past 5 seasons.

2011-12 1.060 1.097 1.079
2010-11 1.198 0.896 1.047
2009-10 1.157 1.014 1.086
2008-09 1.145 1.205 1.175
2007-08 1.397 1.474 1.435

Lidstrom’s 2007-08 season was truly a remarkable season.  Anything better than 1.00 is very good and above average but to have both offensive and defensive ratings in the 1.400 range or higher is truly remarkable and he deservedly earned the Norris Trophy for top defenseman in the NHL that year.  He hasn’t matched his 2007-08 season since but his numbers have still been quite good and remarkably consistent for an aging NHL player, save for his defensive rating in 2010-11.  Lidstrom won the Norris Trophy in 2010-11 but I did not believe he deserved it as there were more deserving defensemen (specifically, Zdeno Chara).  But regardless, Lidstrom has been and still is a top tier defenseman.

Over the past 4 seasons there have been 68 defensemen to play 4000+ minutes of zone start adjusted 5v5 ice time.  Of those defensemen Lidstrom ranks 9th in HARO+13th in HARD+ and 4th in HART+.  The only defensemen more valuable overall over the past 4 seasons are Chara (excellent offensive and defensive play), Willie Mitchell (excellent defensive play which we are clearly seeing these playoffs) and Marc-Edouard Vlasic (excellent defensive ability, good offensive play).  Of the 68 defensemen there are only 6 defensemen who have both a HARO+ and a HARD+ rating above 1.00 and Lidstrom is one of them (Chara, Matt Carle, Christian Ehrhoff, Toni Lydman, and Keith Yandle are the others).  Having both top end offensive ability and top end defensive ability is what makes Lidstrom so great, and this is just looking at his years at the end of his career.  Statistics from his prime years are probably ridiculously good.

The big question is how will the Red Wings possibly replace Lidstrom.  Everyone is pointing to Ryan Suter who is set to become an unrestricted free agent this summer and there is little doubt that the Red Wings will show some serious interest.  So how does Suter stack up against Lidstrom?  Let’s look at his ratings over the past 5 seasons.

2011-12 1.012 1.080 1.046
2010-11 1.004 1.271 1.137
2009-10 1.084 0.955 1.020
2008-09 0.860 0.914 0.887
2007-08 1.041 0.991 1.016

Suter had a poor 2008-09 season but otherwise has been very good as well with ratings quite close to Lidstrom’s but probably a small step back.  Of the 68 defensemen discussed above, Suter ranks 30th in HART+ over the past 4 seasons but that was dragged down by his poor 2008-09.  Over the past 3 seasons he ranks 20th (to Lidstrom’s 16th) in HART+ among defensemen with 3000 minutes of zone start adjusted 5v5 ice time with Lidstrom being the better offensive defenseman and Suter being the better defensive defenseman.  All said, Suter would be a quality replacement for Lidstrom but I would expect the Red Wings offense to take a bit of a hit with Lidstrom not being in the line up.


Should Leafs give up on Armstrong?

 Uncategorized  Comments Off on Should Leafs give up on Armstrong?
May 182012

Over at, Cam Charron did a corsi-based analysis of Colby Armstrong and came up with mixed conclusions regarding his performance over the past several seasons.

So, causes? What caused a player with pretty good possession statistics in Atlanta to completely fall off the map in the last two seasons? System? Trust? Role? A flaw in advanced statistics when players move teams? Or was it just all the injuries that made it a lot tougher on Colby than we think?

I don’t know what the answers to those questions are, but instead of trying to answer then I thought I would take a look at Armstrongs underlying goal numbers look like.  Let’s first start off with a high level view by looking at his HARO+, HARD+ and HART+ ratings.

2011-12 Toronto 235:24 0.449 0.784 0.616
2010-11 Toronto 588:54 1.274 0.823 1.048
2009-10 Atlanta 837:31 1.118 0.948 1.033
2008-09 Atlanta 900:34 1.214 1.016 1.115
2007-08 Pit/Atl 766:46 1.160 0.812 0.986

Save for this past season, where he simply didn’t play enough to get a reliable rating, his HARO+ rating is awfully consistent and remarkably good.  Defensively he had one good season in Atlanta but generally speaking has been extremely sub-par.  The end result is his HART+ numbers are fairly solid and a net positive player overall.  Now lets look at his WOWY stats for players he has played 150 minutes with during a single season.  We’ll start with GF20 data.

Player Year w/ Armstrong wo/ Armstrong Diff
Boyce 2010-11 1.561 1.100 0.461
Versteeg 2010-11 1.148 0.537 0.611
Kane 2009-10 0.931 1.266 -0.335
Slater 2009-10 1.051 0.815 0.236
Peverley 2009-10 0.980 0.826 0.154
Reasoner 2009-10 0.628 0.656 -0.028
White 2009-10 1.049 0.782 0.267
Reasoner 2008-09 0.850 0.719 0.131
Peverley 2008-09 1.469 0.663 0.806
Christensen 2008-09 0.266 1.067 -0.801
Kozlov 2008-09 1.399 0.702 0.697
Perrin 2008-09 0.914 0.648 0.266
Kovalchuk 2008-09 1.390 0.953 0.437
Crosby 2007-08 1.699 1.125 0.574
Malkin 2007-08 1.474 1.147 0.327
Perrin 2007-08 0.944 0.634 0.310
Average 1.110 0.853 0.257

Of the 16 player seasons, there were only 3 where the player had a worse GF20 with Armstrong than without.  That’s pretty good and on average the improvement was 0.256, or about 30%.  He even seemed to make elite offensive players such as Croby, Malkin and Kovalchuk better.  It makes me wonder if Armstrong is contributing in the same way that the players I identified in my “Mixing Toughness with Skill” article did.  Armstrong himself is not an elite offensive player, but the things he does on the ice (retrieving pucks, causing distractions on the ice, drawing attention to himself, etc.) allow the skilled players to do more.

Now let’s take a look at GA20.

Player Year w/ Armstrong wo/ Armstrong Diff
Boyce 2010-11 0.739 0.880 -0.141
Versteeg 2010-11 1.059 0.832 0.227
Kane 2009-10 1.008 1.266 -0.258
Slater 2009-10 0.901 0.815 0.086
Peverley 2009-10 1.224 1.071 0.153
Reasoner 2009-10 0.733 1.006 -0.273
White 2009-10 0.525 0.956 -0.431
Reasoner 2008-09 0.464 0.790 -0.326
Peverley 2008-09 0.851 0.900 -0.049
Christensen 2008-09 0.888 1.115 -0.227
Kozlov 2008-09 0.600 1.130 -0.530
Perrin 2008-09 0.686 1.105 -0.419
Kovalchuk 2008-09 1.137 1.139 -0.002
Crosby 2007-08 0.809 0.783 0.026
Malkin 2007-08 1.053 0.918 0.135
Perrin 2007-08 0.944 0.965 -0.021
Average 0.851 0.979 -0.128

For GA20, negative numbers are good as they indicate fewer goals against.  Interestingly, in 11 of the 16 players seasons the players saw their GA20 drop when playing with Armstrong, though six of them occurred during his previously identified good defensive season of 2008-09 (he didn’t have any consistent line mates that year).  As an average, Armstrong’s teammates saw an a 0.128 drop in GA20, or about 13% which isn’t too shabby.

Now lets take a look at how this pans out in GF%.

Player Year w/ Armstrong wo/ Armstrong Diff
Boyce 2010-11 67.9% 55.6% 12.3%
Versteeg 2010-11 52.0% 39.2% 12.8%
Kane 2009-10 48.0% 50.0% -2.0%
Slater 2009-10 53.8% 50.0% 3.8%
Peverley 2009-10 44.5% 43.5% 0.9%
Reasoner 2009-10 46.1% 39.5% 6.7%
White 2009-10 66.6% 45.0% 21.7%
Reasoner 2008-09 64.7% 47.6% 17.0%
Peverley 2008-09 63.3% 42.4% 20.9%
Christensen 2008-09 23.1% 48.9% -25.8%
Kozlov 2008-09 70.0% 38.3% 31.7%
Perrin 2008-09 57.1% 37.0% 20.2%
Kovalchuk 2008-09 55.0% 45.6% 9.5%
Crosby 2007-08 67.7% 59.0% 8.8%
Malkin 2007-08 58.3% 55.5% 2.8%
Perrin 2007-08 50.0% 39.6% 10.4%
Average 55.5% 46.0% 9.5%

Only 2 times did a player have a worse GF% with Armstrong than without.  Evander Kane saw a marginal drop in 2009-10 and Erik Christensen saw a significant drop in 2008-09.  Most other players saw significant improvements in their GF%, including Kovalchuk, Crosby and Malkin so it seems that Armstrong is a net positive player.

Looking at the above numbers, I think you can firmly put me in the lets not trade away Armstrong just to dump his salary camp.  It is quite possible that the proper thing to do with Armstrong is, if he can get healthy, promote him to the second line with Grabovski and MacArthur and he might make them even better.  Interesting concept.


Henrik Zetterberg Player Analysis

 Uncategorized  Comments Off on Henrik Zetterberg Player Analysis
May 162012

I have written a few controversial pieces here at (for example, my post on Luke Schenn back when Leaf fans thought Schenn was the best thing since sliced bread) and I suspect this might generate some controversy as well because of the conclusions made about Zetterberg’s weak defensive ability.  I also want to do this post to show how I believe we should be doing player evaluation because I believe that most people evaluate players in a poor way.

The first thing I believe is that we must evaluate players based on goals and not corsi/fenwick/shots because there is ample evidence that players can influence shooting percentage (for example here and here) and there is some evidence that players can influence save percentage (for example here).  Because of this, conducting a corsi based analysis will not give you a complete view of a players ability, and I think you will see some of that with Zetterberg.  Furthermore, I believe to get a more full and complete understanding of a players abilities we need to evaluate the players defensive ability and offensive ability separately which is what I will do.

As we are dealing with goal data which can fluctuate from season to season it is best to conduct a multi-season analysis to observe the greater trends, not what could be somewhat luck driven single season results.  Let’s start by looking at Zetterberg’s goals for per 20 minutes (GF20) and goals against per 20 minutes (GA20) on-ice stats and see how they rank league wide and on the Red Wings.

Year GF20 Team Rank League Rank
2011-12 1.157 3 of 11 21 of 312
2010-11 0.931 6 of 12 99 of 321
2009-10 1.073 1 of 10 34 of 319
2008-09 0.860 7 of 12 135 of 318
2007-08 1.193 2 of 12 16 of 312
2009-12 (3yr) 1.054 27 of 295
2007-12 (5 yr) 1.045 23 of 274

The data above is for 5v5 zone start adjusted ice time and the ranks are among forwards with 500 minutes of ice time for single seasons, 1500 minutes for the 3 year average and 2500 minutes for the 5 year average.  As you can see, from the league-wide rankings we are taking the top 9 or 10 forwards on each team so we are getting the majority of the players who get regular shifts in the NHL.  As you can see, offensively Zetterberg performs quite well and while there is some year to year fluctuation, the overall trend is that he has awfully good on-ice offensive numbers.  Now let’s take a look defensively.

Year GA20 Team Rank League Rank
2011-12 0.840 10 of 11 193 of 312
2010-11 0.969 11 of 12 278 of 321
2009-10 0.970 9 of 10 275 of 319
2008-09 0.795 7 of 12 158 of 318
2007-08 0.638 10 of 12 65 of 312
2009-12 (3yr) 0.924 261 of 295
2007-12 (5 yr) 0.845 184 of 274

Defensively, he is not only generally among the worst on his team, he is generally speaking among the bottom half of the league, or worse.  Over the past 3 seasons he ranks 261st of 295 players in terms of GA20 which puts him in the bottom 12 percent which to many I think is surprisingly bad.

Continue reading »

Offense doesn't win in the playoffs…

 Uncategorized  Comments Off on Offense doesn't win in the playoffs…
May 092012

With just one series remaining to be won in the second round there have been 11 playoff series won so far in these playoffs so I decided to take a look at which teams won those playoff series and the results seem to be quite telling.

The team that had the better regular season goals against average has won 7 of the 11 playoff series and lost 4 of them.

The team that had the better regular season save percentage has won 6 series and lost 3 while two series featured teams with nearly identical save percentages (PHI-PIT, LA-VAN).

Most remarkably, the team that had scored the most regular season goals has just one series win in the playoffs.  That one win was New Jersey over Florida.

Here is a table summarizing the results.

Series Most Goals Won Best GAA Won Best Sv% Won
NYR-OTT No Yes Yes
PHI-PIT No No Tied in Sv%
NJ-FLA Yes Yes No
VAN-LA No Yes Tied in Sv%
PHX-CHI No Yes Yes
SJ-STL No Yes Yes
PHX-NSH No Yes Yes
NJ-PHI No Yes Yes
Record 1-10 7-4 6-3

If the Rangers defeat the Capitals they will become just the second team who scored more regular season goals to win a playoff round but they only scored 4 more goals than the Capitals and they held significant leads in GAA and save percentage.

Other observations:

  • 5 of the top 6 teams in regular season save percentage made it to the second round.  Only Vancouver did not, but they lost to a team in the top six.
  • None of the bottom 7 teams in save percentage made it to the second round and only one of the bottom nine did – Philadelphia who defeated a team with an equally bad save percentage in Pittsburgh.
  • Four of the top 5 teams in goals against average made it to the second round (Vancouver did not) and two are already in the conference finals and the Rangers could make it three.
  • None of the bottom 9 teams in GAA made the second round and if the Capitals go on to lose to the Rangers, none of the bottom 21 teams will make it to the conference finals
  • Of the top 10 teams in regular season goals scored, only Philadelphia and Nashville made it to the second round, and both defeated top 10 goal scoring teams in the first round to do so.
  • None of the top 10 regular season offensive teams have made it to the conference finals.

It is really quite remarkable that only 2 of the top 10 offensive teams made it to the second round and none to the conference finals while six of the top 10 defensive teams made it to the second round and possibly (if Rangers win) 4 of the top 9 defensive teams will make it to the conference finals.  Defense is clearly winning these playoffs.  After a few seasons of offense dominating post lockout the league has once again transitioned back to a defense-first league.

The western conference finals sees the Kings face the Coyotes.  The Coyotes had a marginally higher regular season save percentage (92.5% to 92.4%), while the Kings had a better goals against average (2.03 to 2.32) so defensively the edge probably goes to the Kings.  The Coyotes scored more goals than the Kings (210 to 188) which so far these playoffs is a negative so I guess that suggest the Kings are the favourites in the second round.

If the eastern conference finals is between the Rangers and Devils, the Rangers scored more regular season goals (222 to 216), had the higher regular season save percentage (92.0% to 90.7%) and had the better goals against average (2.18 to 2.45).  Rangers would have to be the favourites.

If Washington comes back to defeat the Rangers, the Capitals marginally out scored the Devils (218 to 216), had a marginally higher save percentage (90.9% to 90.7%) but had a worse goals against average (2.45 for NJ to 2.72 for Washington).  The Devils should probably be considered the favourites over Washington.