Bouw knows defence . . . probably


So far, the war that was furiously waged in baseball between statheads and “stats?-we-don’t-need-no-steenking-stats” traditionalists has been a relatively under-the-radar skirmish in hockey. Because of the comparative number of stats in each sport and the disparity of number-crunchers doing baseball compared to hockey, we’ll likely never see anything in the shinny world along the lines of the Sabermatricians’ revolution and the Establishment’s stick-that-abacus-where-the-sun-don’t-shine resistance movement.

Hopefully that means the two sides in hockey will be able to find a middle ground, because that’s really where everyone needs to be. For the most part, advocates of advanced stats have been frustrated by the stubborn refuse of old-school types who believe only what they see with their own eyes and refuse to consider such, to them, outlandish factors such as Corsi numbers.

On the flip side, there are proponents of advanced metrics who will steadfastly cling to a notion based on statistics while ignoring anecdotal evidence and, to a certain degree, common sense.

Now to be absolutely clear about this, colleague Jonathan Willis was NOT guilty of the latter transgression in a piece that filled this space on Tuesday. The esteemed Mr. Willis was in fact very careful to footnote his point about the respective save percentages behind Jay Bouwmeester in Florida and Dion Phaneuf in Calgary was far more involved than a cursory glance at the numbers might suggest. That example is brought up only because the naked statistical citation, sans the caveat, is a perfect demonstration of what can happen when advanced metrics are taken too far and taken too far out of context.

The save percentage stat is a useful complement to the notoriously inadequate plus-minus statistic because it brings into the play the possibility a skater plays in front of a leaky goaltender. But as Herr Willis alludes to in his article, it’s overly simplistic to declare that if Bouwmeester and Phaneuf had swapped places in 2008-09, the difference in goals against can be calculated to within a fraction of a puck.

This, ladies and gentleman, is where common sense comes in. For most hockey fans in the Western Conference, Bouwmeester the NHLer is a little like the Yeti. We’ve heard a lot about him and he sounds awfully impressive, but relatively few of us on this side of the Mississippi can provide eye-witness accounts of his post-Medicine Hat existence, let alone his rumoured awesomeness. That’s what happens when you play for a Southeast Division club that hasn’t been active in the spring since the Florida Panthers goaltending tandem consisted of Mike Vernon and Trevor Kidd.

On the other hand, Flames fans this past season got a real good look at Phaneuf’s defensive flaws, his tendency to stand around while pucks were centred to opponents camped just a few feet away from him, his habit of moving back into defensive position after an offensive rush with the alacrity and swiftness of a constipated octogenarian wearing shoes on the wrong feet and his insistence on playing neither the shot nor the pass when defending two-on-ones and, well, they can’t help but wonder if those operating practices might have had a little something to with the puny save percentage when No. 3 is on the ice for the Flames.

All of which is an incredibly long-winded way to say that the save-percentage hypothesis alone shouldn’t deter anyone from believing that the Bouwmeester isn’t a better defender than Phaneuf. Really, unless you believe that the NHL’s version of the witness protection program is preventing us from finding out that 2008-09 Panthers netminders Tomas Vokoun and Craig Anderson are the second coming of Ken Dryden and Bernie Parent, it’s reasonable to believe that Bouwmeester’s defensive acumen deserves some credit for the credit for the spiffy save percentage behind him. At least, that makes for an acceptable working theory until more empirical evidence can be gathered now that Bouwmeester is within range of the Canadian network cameras.

Not having been present to look for a possible tongue in cheek when Jonathan typed that the arrival of Bouwmeester might be a good news-bad news scenario for the Oilers in that it might necessitate the jettisoning of Phaneuf, it’s hard to know how much credence he himself puts in that theory. I’m guessing relatively little, but then again I was the idiot who thought until very late Oleg Saprykin’s hockey sense would finally kick in, so who am I to question hockey judgment?

That won’t stop me from declaring that for Flames fans, the bonus good news about Bouwmmester’s signing (and the subsequent free-agent additions) is that there’s not enough money left for Todd Bertuzzi, even at a discount. Bouwmeester’s arrival, Bertuzzi’s apparent departure and the possibility that the hiring of his former junior boss will get Phaneuf’s career back on the rails may be the best bets to improve Calgary’s goals-against numbers until somebody at the Saddledome figures out how to turn the hands of Miikka Kiprusoff’s clock back a half-decade.

  • The problem with your argument Jean is that everyone in Florida had high EV SV% figures behind them. Bouwmeester's .936 looks gaudy until you notice that the team number was .932. Vokoun and Anderson were terrific last year, regardless of who was out there. Bouwmeester played about 40% of the EV ice time. If you don't think that it was goaltending, then you're almost forced into arguing that the other Florida defencemen were exemplary as well. Either that, or Bouwmeester was able to influence the 60% of the EV ice time he was on the bench via mind control. If so, Darryl Sutter is a genius for signing him 😉

    FLorida's EV SV%, via Vic Ferrari's excellent timeonice site.

    For comparative purposes, here's Calgary's.

  • Point taken, Robert.

    It may have not been clear from the piece, but I really tried to avoid talking in absolutes when talking about the Panthers situation because like so many others who follow the Western Conference, I just didn't see Florida play enough.

    My conclusion, subject to review and reconsideration, is that Bouwmeester's defensive ability must be a contributing factor to the elevated save percentage behind him. How much of a factor is it compared to the quality of the Panthers' goaltending and the DeBoer/Martin influence on the team's playing style? I'm all ears as far as that goes, but the fact Bouwmeester's save-percentage number is a little higher than the team average despite the fact Bouwmeester's qualcomp number was highest on Florida might be evidence he knows what he's doing in his own end.

    Meanwhile, if it's important to hedge your bets when discussing the Panthers, it's somewhat easier to be more assertive when speaking to the Flames' situation. Yes, Miikka Kiprusoff is not the man he used to be but I feel pretty confident in saying that Phaneuf was the author of a goodly portion of his own poor save percentage figure because of various issues. The save percentage numbers for guys like Sarich (.929) and Regehr (.922) and even the much-maligned Aucoin (.908) would seem to support that contention, especially since those three (as well as Mark Giordano in 58 games) all had higher qualcomp's than Phaneuf.

    Bottom line, it's not my intention to completely dismiss the difference between the Bouwmeester and Phaneuf numbers, but I find it hard to believe that Vokoun/Anderson are that much better than a lower-quality Kiprusoff.

  • but I find it hard to believe that Vokoun/Anderson are that much better than a lower-quality Kiprusoff.

    Well, we'll have to agree to disagree on that, Jean, at least as far as Voukoun goes.

    Since Kiprusoff became a starter full time in 03/04, here are their overall SV%:

    Kiprusoff .933 .924 .917 .906 .903

    Voukoun .909 .919 .920 .919 .926

    What's that line from Planes, Trains and Automobiles? "You're going the wrong way!" In 02/03, Voukoun's first full year as a number one goalie in Nashville, his SV% was .918, BTW. The guy's been good since before Kiprusoff came to Calgary, and he's markedly better now.

    For a bit of context, here are their teams' Shots Against league rankings and shots for/against differential numbers over the same 5 years:

    03/04 CGY – 5th SA, +1.9 SD
    NSH – 17th SA, -1.3 SD

    05/06 CGY – 6th, +0.4
    NSH – 26th, -3.2

    06/07 CGY – 21st, -3.2
    NSH – 24th, -3.6

    07/08 CGY – 13th, -0.3
    FLA – 29th, -2.5

    08/09 CGY – 16th, +2.4
    FLA – 30th, -5.3

    Voukoun has played for teams that have given up a very high number of shots relative to the league, and have been regularly outshot by significant margins. Florida had the worst shot differential in the league last season as well. If there's some evidence of a great defensive system in Florida, I guess I'm not really seeing it. Kiprusoff has had better team defence in front of him, and since 05/06 has done less with it.

    Look, I certainly expect Jay Bouwmeester to make the Flames better, and Dion Phaneuf was very poor last year, as well as injured. Those points are completely correct. I just think that if Miikka Kiprusoff has a better SV% next season, almost all of it will be up to him.

  • Jean,

    The preliminary stuff done on this subject suggests that individual players don't exert a lot of influence over their goalie's SV% – it's in the +/- 0.001 area I believe. Team effects, goalie quality and just plain old randomness are apparently far weightier factors (which is, I admit, counter-intuitive).

    As an illustration, the Flames player who had the best SV% @ 0.956 last year was #27…Andre Roy.

  • Kent wrote:

    The preliminary stuff done on this subject suggests that individual players don’t exert a lot of influence over their goalie’s SV% – it’s in the +/- 0.001 area I believe. Team effects, goalie quality and just plain old randomness are apparently far weightier factors (which is, I admit, counter-intuitive).
    As an illustration, the Flames player who had the best SV% @ 0.956 last year was #27…Andre Roy.

    I would not argue for one second that randomness is a huge factor in that category. In fact, I would be surprised if it was anything but. I guess the point I'm labouring to make is that I can't accept that any negative Phaneuf defensive numbers can be explained away merely by poor goaltending behind him. We've also seen enough anecdotal evidence of Phaneuf's defensive lapses to know better. That's why the after-the-fact injury excuse rings hollow — somehow Phaneuf had enough steam to rush the puck up ice, enter deep into opposition territory and get a good scoring chance but he didn;t have enough to take one more stride and tie up the guy loose in the slot?

  • Just you wait till Jonathan Willis wades in on this argument. He is currently powering up his array of Cray Supercomputers and formulating a dizzying algorithmic response.

    btw What's it like signing a free agent?

    *leans on hands and wonders*

  • @ Robert Cleave:
    I'm not prepared to claim Kiprusoff is better or even the equal of Vokoun at this stage of their careers. The numbers you cited say otherwise and I simply haven't seen enough of Vokoun as a Panther to make an intelligent claim one way or another.

    That said, for whatever it's worth, I do recall Vokoun briefly lost his No. 1 job this past season and I did see enough of him when he was Nashville that I can't bring myself to believe that there's enough of a difference between he and Kiprusoff to adequately explain a .040 difference in save percentage, which is the only reason the comparison was brought up in the first place.

    Anyway, it looks like we have enough common ground in this discussion that I don't think the various details matter too much.

    Just thinking though, if Bouwmeester and Phaneuf get paired together and are never separated for even one shift, how is anyone going to debate their relative defensive merits in a year's time if their numbers will all be exactly the same?

  • Willis and I parsed some of the distinctions you guys are talking about in this thread. Short version is, OF COURSE players can and do have an impact on a goalie's expected save percentage, but that effect is bloody tough to discern from actual save percentage, because so many other things (mainly puck luck) drown out that impact.

  • One thing I'm realizing is that I'm a fence-sitter when it comes to making hockey conclusions, which I suppose is the height of weenie-ness. I don't mind resorting to statistics (especially if it helps my argument) but there are instances where I fall back on "I know what I know and no number-crunching is going to convince me otherwise."

    Not sure how much of a hockey fan Aaron Levenstein was but his line, "Statistics are like a bikini — what they reveal is suggestive, but what they conceal is vital" contains an important lesson for all of us.

  • Lefebvre wrote:

    Just thinking though, if Bouwmeester and Phaneuf get paired together and are never separated for even one shift, how is anyone going to debate their relative defensive merits in a year’s time if their numbers will all be exactly the same?

    We won't be able to, so we'll arguing over semantics again 😉

    Thanks for the replies, Jean.

  • Here's the link that Kent was referring to above (I think he was, anyway).

    Jean, you're right when you guess that I was talking a little bit tongue-in-cheek; I'm fairly confident that Bouwmeester's a better defender than Phaneuf is by a fair degree, particularly given that he didn't have the luxury of Robyn Regehr going out to play the real bad guys most of the time.

    Mostly, my point was that if Calgary needs to move Phaneuf to get this deal completed, they'll be selling at a really low price. Phaneuf had a rough year, but at least some of it was attributable to the goaltending behind him, and with even league-average tending while he was on the ice he would have put up substantially better numbers in the +/- department. My other point was that Jay Bouwmeester got outshot by a ton in Florida – probably because he was the best defenseman on a lousy team – but that this sort of thing can't continue on a stronger Flames squad.

    And I can't complain about my treatment here – somehow you saw what I was trying to say despite the poor way I chose to express it.