Rating the Flames Defensively


Measuring a player’s defensive contributions statistically is currently one of the most difficult things to do. You can’t just look at goals (or shots) against, because the great defensive players are up against the toughest opponents and in their own end, while the weaker ones are against softer opponents mostly in the opposing zone. Calculating for this is very challenging and even if you could, is there really a way to compare how effectively a player in one situation did relative to someone in another?

Many people have tried to create catch-all statistics, like Tom Awad’s Goals Versus Threshold (GVT), Iain Fyffe’s Point Allocations, Alan Ryder’s Player Contributions (PC) and Justin Kubatko’s Point Shares (PS), to name just a few. And while they’re a step up on plus/minus (or even Corsi, which is a shot-based plus/minus statistic), they leave a lot to be desired, often leading to crazy conclusions like Anton Babchuk is Robyn Regehr’s defensive equal, as we explored earlier this year.

Advertisement - Continue Reading Below

Regardless of the shortcomings of the existing systems, it’s still worthwhile to look back and see how the Flames stack up against one another defensively.  We’ll also look at Vollman’s Folly, a name hastily applied to a system we designed to prove that a better catch-all defensive statistic was possible, something it was never intended to actually be. 

Vollman’s Folly works by figuring out how many shots a team prevented relative to the 1992-93 San Jose Sharks (a replacement-level team defensively), and assigning credit to the players relative to the amount of defensive ice-time they played, which was crudely estimated based on their special teams usage.

Let’s start with the defensemen.

Flames 2011-12 Defensemen

Both the Defensive Point Shares system and Vollman’s Folly correctly concluded that Jay Bouwmeester was Calgary’s best defensive player, but Defensive GVT incorrectly gave Scott Hannan a huge nod. This was basically the only point of contention between DPS and DGVT, which otherwise agreed completely.

Advertisement - Continue Reading Below
Defenseman      GP DPS  VF DGVT
Scott Hannan    78 3.0 62.0 5.1
Jay Bouwmeester 82 4.8 64.1 3.8
Mark Giordano   61 3.4 27.8 3.6
Chris Butler    68 2.8 55.1 3.2
Cory Sarich     62 2.5 37.9 2.3
T.J. Brodie     54 2.4 10.0 2.1
Derek Smith     47 1.7  9.3 1.2
Anton Babchuk   32 1.3  4.8 0.9

Is Mark Giordano Calgary’s 2nd best defensive option on the blue line?  If so you’re probably happy with DGVT, and if you even believe he’s Bouwmeester’s equal then you’re probably happy with DPS because his 3.4 DPS in 61 games would work out to 4.6 over Bouwmeester’s 82. 

On the other hand if you believe Mark Giordano is a much improved and underrated defensively but doesn’t contribute quite as much in his own end as Calgary’s other top-four defensemen Scott Hannan and Chris Butler, then VF might be worth a closer look.

Otherwise the only bone of contention could be among the depth defensemen. While DPS and DGVT see Cory Sarich and T.J. Brodie as basically equivalent and only a short step below Chris Butler, VF has Sarich up ahead of Giordano, and Brodie in the back of the pack with Derek Smith and Anton Babchuk. 

Now, let’s see what the picture is like up front.


Jarome Iginla as Calgary’s 2nd best defensive forward?  That’s how DPS sees it. Fortunately DGVT adds a bit more sanity to the picture where Calgary’s beloved captain barely edges out Matt Stajan and Roman Horak for 6th.

Forward          GP DPS  VF  DGVT
Alex Tanguay     64 1.9 11.9  3.4
Curtis Glencross 67 1.0 20.9  2.7
Lee Stepmniak    61 1.2 14.3  2.1
Olli Jokinen     82 1.5 11.7  1.9
Tom Kostopoulos  81 0.6 25.7  1.8
Jarome Iginla    82 1.7 11.8  1.6
Matt Stajan      61 0.9 18.0  1.5
Roman Horak      61 0.9  6.0  1.5
David Moss       32 0.4  9.2  1.3
Blake Comeau     74 1.0 25.5  1.2
Blair Jones      43 0.5  8.2  1.1
Mike Cammalleri  66 1.1 10.5  1.0
Paul Byron       22 0.4  1.9  0.7
Mikael Backlund  41 0.2  9.7  0.2
Lance Bouma      27 0.1  6.9  0.1
Krys Kolanos     13 0.1  1.0  0.1
G. Desbiens      10 0.0  0.6 -0.3
Tim Jackman      75 0.0  5.5 -0.8

Alex Tanguay is Calgary’s best defensive forward, at least according to both DPS and DGVT, which certainly makes more sense than VF’s assertion (Tom Kostopoulos, who probably didn’t play a regular shift against a 20-goal scorer at even-strength all season). While he played a slightly tougher role, the same thing could be said for Blake Comeau, who comes in 2nd.

Advertisement - Continue Reading Below

Curtis Glencross is probably the team’s actual best defensive forward, based on people’s subjective opinions, as well as Sutter’s choice to use him against top opponents and on the first penalty-killing unit. He’s ranked 2nd by DGVT, 3rd by VF, and 6th by DPS.

DGVT could very well be the winner here, not just for ranking Curtis Glencross so highly and for avoiding overestimation of Iginla, Kostopoulos and Comeau, but for wisely ranking Lee Stempniak 3rd just ahead of Olli Jokinen.

The Verdict

Defensive GVT did a surprisingly good job ranking Calgary’s players defensively, except for overestimating Scott Hannan (or underestimating Jay Bouwmeester). The top forward ranking of Alex Tanguay, Curtis Glencross, Lee Stempniak, Olli Jokinen and Tom Kostopoulos all ahead of Jarome Iginla, Matt Stajan and Roman Horak seems quite reasonable.

Defensive Point Shares may have placed Jay Bouwmeester on top, but overestimated Mark Giordano, left the defensemen in far too tight a pack, and made a complete mess of the forwards.

The simplistic Vollman’s Folly system underestimated Mark Giordano, overestimated Cory Sarich, and placed depth options Blake Comeau and Tom Kostopoulos ahead of legitimate top-line shadows like Curtis Glencross.

A usable catch-all defensive statistic might be further from our grasp than we had hoped, but in the case of the Calgary Flames defensive GVT is certainly getting closer.

Advertisement - Continue Reading Below

  • MC Hockey

    Cool article Robert. And you explained it well enough for most people to follow so props to you! My thoughts:

    I guess I agree that DGVT is the best stat to measure “Defensive talent” but it’s overestimation of Hannan is probably not way off. Here is why:

    First, versus other D-man on Flames, he surprisingly does OK at driving play, coming in 4th after Butler, Bouw, and Gio if you use the “Corsi Rel QofC” stat from behindthenet.ca.

    Second, he also does NOT benefit from goalies playing well in front of him (perhaps they relax for the old vet), because from behindthenet comes his On Ice Sv% being worst on Flames and his PDO being second worst (JayBo is worst…again that relaxation factor in play?).

    Third, since he scores very few points, that helps explain his minus 10 if you think Plus-Minus has any value (I know..it probably does not!).

    Finally, if you ignore his age and be “a bit subjective” he is good from a technical point of view in defending (positioning, blocking shots, etc) and headmaning the play out of the zone (makes good and smart first passes, etc).

    Thus unless you tell me DGVT is supposed to “correct” for differences in quality of opponent or other “complicating factors” then perhaps it is an accurate view of Hannan? Perhaps you can reply?

    • MC Hockey

      Corsi Rel QoC is quality of competition BASED ON possession of opposing players, not possession itself. Hannan’s was high simply because Sutter rightly preferred him to Sarich, Brodie, and Smith at even strength. If you want to see who drove the play, use CorsiRel or CorsiOn.

    • Yeah, actually Hannan had the worst relative corsi rate on the blueline…and he playd with Gio, who has been the best possession blueliner on the team for a couple years running.

      These holistic stats discussed by Rob tend to be relatively blind to possession, circumstances and percentages. They roll everything into one which is useful because it makes it simpler, but they also lose nuance in the process.

      The best thing we can say about Hannan this year is he was decent on the PK. I’d say he’s a third pairing defender at best this point in his career.

  • MC Hockey

    Thanks guys…misunderstanding some stats obviously. And I am not a huge Hanna fan but do feel he is decent but would still rather see him ove Sarich but also feel a young guy over Hannan is better!

  • @mchockey

    Thanks for the kind words.

    Kent is right, these catch-all statistics ignore possession, competition, etc.

    Kent’s also right about Hannan. He was very impressive on the penalty kill but at even-strength he was average at best.

    As for a low goalie save percentage, remember that Bouwmeester and Hannan face players with much higher shooting percentages than the players T.J. Brodie, Cory Sarich and Derek Smith usually face.

    And a word about some of these stats:

    1. Corsi is simply a player’s plus/minus except that it’s shot-based and calculated per 60 minutes.

    2. Therefore, Corsi can be effected by the same things as plus/minus: team, opponents, situation, score, etc!

    3. Quality of Competition is basically the average Corsi of one’s opponents.

    4. Therefore, it is not a measurement of the average opponents offensive abilities.

    5. More accurately it is partly a measurement of how well opponents drive the play and the situations they generally play in (including what we mentioned in point #2).

    6. Furthermore, it is affected by all the same things we mentioned in point #2.

    7. Relative Corsi, which is just that player’s Corsi minus the team’s Corsi when he was on the bench, eliminates some of the team-based bias, but not much else.

    Long story made short, shot-based statistics are very useful but they’re also very complicated and I’ve seen them misused by people whose initials aren’t KW all the time. In fact, I probably made a few mistakes right here.

  • ChinookArchYYC

    Hey Robert, I got your link to Dorkapolusa the other day. Thanks, that pretty well covered what I was looking for.

    Is there any value in comparing defensive pairing with how they can stop top perferming forwards (RelCorsi) on each team? In other words, how good is Bouwmeester/Butler at preventing Zettebreg/Dadsyuk over 60?

  • Arik


    That’s going to be too small of a sample size- especially outside the division with only a few games played and even less where the Coach gets the line matches he wants.

    Personally I’m not much of a QualComp fan in general, I think zone starts are a better indicator of usage and have more effect on a player’s underlyings, but that’s just me.

  • Derzie

    Great article. Easy to follow. My personal opinions are quite inline with DGVT. It also shoots holes in the love for Backlund’s D work. Hannan and JBo with the top defense stats makes good sense to me based on watching every game this season.

  • @ChinookArch


    And fortunately there’s always a Flamesnation fan to beat me to the correct answer.

    Ask yourself what Quality of Competition really is – the average Relative Corsi of a player’s opponents. And Relative Corsi is the attempted shot-based plus/minus per 60 minutes, relative to their teammates.

    So a player with a high Quality of Competition is facing players with high Relative Corsi.

    Who are players with high Relative Corsis? Players who are on the ice for a lot more attempted shots for than against, relative to their teammates. And who are they?

    Players who drive the play forward, either because they’re great two-way players, or more likely players who are getting the most favourable match-ups (in the offensive zone and against weaker opponents, for example).

    So basically a player with a high Quality of Competition is someone who typically plays against someone who gets played in favourable situations.

    In many cases those are the best offensive players (like the Sedins), but all those layers of onion make it easy to misuse and misinterpret Quality of Competition numbers.

    It’s probably best to use Quality of Competition like Kent normally does: as a rough ranking of where a player falls on the depth chart, in terms of playing against top opponents.

    Specifically, for the Flames, it would be Glencross, Jokinen, Moss, Iginla, Backlund, Cammalleri, tanguay, Comeau, Stempniak, Jones, Bouma, Stajan, Horak, Kostopoulos, Kolanos, Jackman, Desbiens, Byron, Nemisz which, I’m sure you’d agree, looks right!

  • RKD

    From an observational point of view, the Flames were very poorly defensively.

    Though Jay-Bo and Chris Butler played against top opposition every night that tandem looked very soft.

    Giordano never looked to regain his full form after coming back from his hamstring injury.

    Brodie and Smith looked pretty good until they got injured. Sarich and Hannan were shut down guys, but lacked mobility and were caught standing still.

    Babchuck was all around poor, did have a couple stretches of somewhat decent play near the end of the season.

    I look at goals for, despite a great season by Kipper the Flames surrendered 226 goals against in 82 games. Which works out to an average of 2.75 goals per game. They only scored 202 goals an average of 2.46. Some of the forwards were leaning heavily in double digit minus ratings.