Why I Could Care Less About (Goalie) Wins and Losses

I’m sure everyone here is sick of me preaching doom and gloom about the Oilers’ coming season. I’m equally sure that they’re sick of my rather pessimistic predictions about the Khabibulin contract. This article isn’t meant to continue those trends, but rather to make a general point that is inspired by one particular line of reasoning about Khabibulin: that he’s a winner.

It’s always seemed to me that a goaltender’s win-loss record is well beyond his control; it just makes sense. A logical extension of this is winning the Stanley Cup; since the goalie alone can’t determine wins and losses, it seems that winning and losing is a bad way of judging goaltender ability.

An extreme example of this occurred during the 2005-06 season. An NHL team had two goaltenders; goaltenders who had posted the following stats lines:

  • Starter: 24-9-7, 2.29 GAA, .919 SV%
  • Backup: 6-17-1, 3.00 GAA, .910 SV%

The backup obviously had poorer numbers, but that was still an exceptional save percentage. With this team out of a playoff spot, they elected to swap the backup to a competitor for draft picks; the fact that the backup was a pending free agent undoubtedly helped them to make that decision.

In any case, there were some serious doubts about the backup. Not only was his win/loss record shoddy, but despite being 36 years old he had but a single playoff round win to his name – and even there, his performance hadn’t been great (posting a .903 SV%).

The backup was Dwayne Roloson, of course, and here (in part) is what Tyler Dellow of mc79hockey.com had to say about the trade on the day it took place:

Now, from the Oilers perspective, they’re getting a guy who has just put up solid numbers for the last three years. I’ve included Roloson’s numbers in a table to the left. His relative save percentage has been outstanding for the past three seasons after a very mediocre start to his career-1019, 1024 and 1015. The last time I did the Oilers as a team, they were at 980-even if Roloson falls to league average, he’d be a significant improvement over what they were getting.

Some of the dumber commentary I’ve seen on this (like the Wild’s press release, and every second comment on Calgarypuck), makes a big deal about Roloson’s record compared to Fernandez. His significantly higher goals against average has also been mentioned. I wouldn’t read too much into it. Roloson has, for some reason, had the Wild play significantly worse in front of him. On my numbers, the expected even strength goals against per 60 (based on shot volume, distance and whether or not it was a rebound) for the Wild when Roloson was in net was 3.07; when Fernandez was in net, it was 2.63. That’s a pretty significant difference. The Roloson figure of 3.07 (which is designed to reflect the team in front of him as opposed to the goalie) is a worse figure than that put up by every team except Atlanta, the Islanders and the Penguins this sesaon. I’m not so sure that he’s had the Wild’s usual stellar play in front of him in games he’s played this season. I don’t generally pay attention to wins as a goalie stat but I’d assume that the Wild have given better offensive support to Fernandez.
Much has been made of Roloson playing for the Wild and the idea that he’s just putting up numbers playing behind a stifling defence. I’m not sure what to make of this myself-while it’s true that he put up worse than league average numbers prior to the last three years, he was consistently excellent in the minors and spent much of his NHL time playing for teams that weren’t particularly good defensively. As I noted, there are indications that they haven’t played particularly well in front of him this season and he’s still put up good numbers.

Comparing him to the trio we’ve been using all year, I have him as being 3 goals below his expected goals against at even strength this year, again, with expected goals against being determined by shot volume, rebounds and distance. He’s faced 536 ES shots as of my last update. That’s a hell of a lot better than the guys that the Oilers have been using-as of my last update, Jussi was -15 on 633 shots, Ty was -5 on 198 shots and Morrison was -3 on 248 shots. This indicates to me that Roloson has been significantly better this season than the guys he’ll be replacing.

My conclusion is that if he puts up the save percentage he has so far this season for the remaining 20 games and we assume he plays 80% of the games, I figure he makes the Oilers 12 goals better. That’s a ridiculously huge number, worth in the neighborhood of 4 points in the standings.

That’s a long quote, but it’s worth reading. It’s also illuminating because Tyler (even more so than me) has been accused of having a consistently pessimistic view of the Oilers. That isn’t true; he’s consistently judged goaltenders by the available market and by their save percentage over the past few seasons in relation to the rest of the league.

The same things that made Roloson appealing in 2005-06 are what he’s using to judge goaltenders now. As we saw, save percentage worked well then and I think it will continue to work well in the future. Honestly, a goaltender’s win/loss record – even a goaltender’s playoff record – holds very little weight with me if he’s been sub-par save percentage-wise.

Chris Osgood is, in many ways, the perfect example of this. People who watch the numbers agree with people who watch the games: he’s a fairly average tender playing for a great team. However, he’s won enough Stanley Cups that there’s now almost a ‘Cult of Osgood’; he was considered a leading candidate for the Conn Smythe trophy in the event of a Wings victory, and people talk about how he turns it on for the playoffs.

I just don’t buy it. He’s played well (not great, but well) behind a phenomenal team, but he isn’t one of the better goaltenders in the NHL.

In any case, the point here isn’t to bash Khabibulin (who had a remarkably good save percentage last season) but just to point out that winning and losing is beyond a goaltender’s control – and it’s wrong to judge him (for good or ill) based on the number of wins or losses he’s picked up along the way.

Final point: the NHL record for most losses is currently a record held by two goaltenders. Curtis Joseph is one of them, and the other is one Gump Worsley. Worsley owns the rather pedestrian career record of 335-352-150, along with a couple of other things (four Stanley Cup rings, two Vezina trophies). I realize I’m being as subtle as a sledgehammer here, but looking at the win/loss record alone tells you virtually nothing about a goaltender.

  • Hippy

    Tyler wrote:

    I’m Willis’ Northern BC crony. The stats thing is a coincidence.

    Coincidence or not, the pattern is clear.Tyler wrote:

    I missed the part where you explained how many times you’ve seen the Hawks play in the past four years.

    I missed the part where you 'fessed up about being shoulder to shoulder with JW in your belief in advanced stats so that we can at least start with an honest baseline about where you're coming from.

  • Hippy

    I can't stand stat geeks. Guys that never played the game figure they can become "gurus" just by analyzing stats for hours. Bottom line is: Just win baby! The Bulin wall was a great signing and a HUGE upgrade!

  • Hippy

    West Coast Oil wrote:

    I never had a chance to respond to JW’s responses to me on the last blog.
    My entire point previously was that many of the posts here have become doom and gloom.
    My job means I deal with stats all the time and the one thing I know regarding stats is that you can always find a stat to support your argument because in most cases they are subjective.
    You cannot say your argument is correct or that your choice would be better based solely on statistics.
    This is like saying you can be a doctor because you know this medicine cures this ailment half the time.
    There are many factors that go into decisions beyond just the stats.
    We do not know the nuances of people and situations.
    I guess I prefer the good old emotional response on occasion with a stat thrown in.

    That's exactly the way I see it, westcoast. Thanks for putting it into the words that I haven't been able to.

    @ Tyler & Jonathan Willis:

    This is what I'm talking about, and I'm sure you guys will roll your eyes at another emotional argument rather than an empirical one, but it's also the reality of the situation with most of the fans out there.

  • Hippy

    West Coast Oil wrote:

    My job means I deal with stats all the time and the one thing I know regarding stats is that you can always find a stat to support your argument because in most cases they are subjective.

    This isn't actually true. You can find a stat that will support your argument most of the time if you don't mind making fallacious arguments. Used properly, stats are pretty useful. All they are are records of the number of times something objective occurred.

  • Hippy

    Tyler wrote:

    I missed the part where you explained how many times you’ve seen the Hawks play in the past four years. I’m sure if you’re bagging on Willis for relying on the stats and explaining that some of you don’t need to, you must have seen a lot of Hawks games. What are we talking here?

    And my mistake for not getting to that part of your question. I've seen Khabi in net for the Hawks maybe…10 times in the regular season over the last couple seasons, and more so in the playoffs (esp. this past season vs. CGY and VAN).

    I see Khabi as an upgrade over Roloson in terms of the difficulty/quality of shots faced and saved (which I noticed without having to consult the numbers, but the numbers did agree with me as JW argued a few weeks ago). Sure, I could be wrong and the stats could further cement that fact, but I'm willing to take that chance and take the plung if I'm incorrect.

    There has to be a better way of meshing these statistical arguments and the observation (LT's "seen him good") arguments…maybe it's already being done and I'm not paying close enough attention? I know that many of the stats you gents use are very useful and effective, but some others leave me scratching my head and wondering just how necessary they are.

  • Hippy

    I never had a chance to respond to JW's responses to me on the last blog.
    My entire point previously was that many of the posts here have become doom and gloom.
    My job means I deal with stats all the time and the one thing I know regarding stats is that you can always find a stat to support your argument because in most cases they are subjective.
    You cannot say your argument is correct or that your choice would be better based solely on statistics.This is like saying you can be a doctor because you know this medicine cures this ailment half the time.
    There are many factors that go into decisions beyond just the stats.
    We do not know the nuances of people and situations.
    I guess I prefer the good old emotional response on occasion with a stat thrown in.

  • Hippy

    I missed the part where you explained how many times you've seen the Hawks play in the past four years. I'm sure if you're bagging on Willis for relying on the stats and explaining that some of you don't need to, you must have seen a lot of Hawks games. What are we talking here?

  • Hippy

    Tyler wrote:

    Tyler wrote:
    I’m Willis’ Northern BC crony. The stats thing is a coincidence.
    How many Hawks games have you watched in the past four years? I’m talking about games you’ve actually watched at least two periods of, not seen some highlights or watched for ten minutes while flipping channels.

    I'm not talking about Khabibulin individually, I'm talking about the argument Willis has against goaltenders in particular.

    I liken it to professional football where passing statistics for quarterbacks are seemingly paramount over a win-loss record. As a goaltender, you're front and centre on a hockey club, just as a QB is in football. I would argue that the position has more to do with winning and losing a game than any other position, just because of how often that player is involved with the game at critical junctures.

    Maybe the stats back up what I'm arguing, maybe they don't, but it's the way I view the game. I'm sure you're just as tired of the non-empirical arguments as I am of the over-analyzed statistical arguments, but neither of us are going anywhere and that's ultimately my point. When it comes to debating with an audience that generally doesn't completely understand and/or completely accept the statistics-weighed approach, seemingly redundant posts have a way of appearing to come across as over-thought with stats that are doctored to fit any argument.

    I do believe that one can make numbers say just about anything…politicians make it happen each and every day, don't ya know!

  • Hippy

    GSC wrote:

    Tyler wrote:

    Yeah Willis. You might as well just ditch all of these “historic examples” and “evidence”. Your argument is lacking in truthiness.
    And, of course, it’s another stats crony to the rescue.
    You go by statistical formulas that most of us don’t need to consult, we know enough by what our eyes tell us.

    I'm Willis' Northern BC crony. The stats thing is a coincidence.

    How many Hawks games have you watched in the past four years? I'm talking about games you've actually watched at least two periods of, not seen some highlights or watched for ten minutes while flipping channels.

  • Hippy

    Hey, hey, hey – this isn't about Khabibulin here. Seriously; that just got me thinking about this topic, but I've pretty much said my piece on the Russian.

    All I'm saying is that wins and losses, which are brought up so often when evaluating goaltenders, are basically useless.

    Save percentage isn't good enough, by itself, but it's infinitely more useful than W/L.

  • Hippy

    Jon's pointing out that numbers are subjective in their interpretation – that's all. 'What our eyes tell us' is even more subjective…and at times inaccurate.

    I like the signing of Khabi, personally. Very safe if you think about it – he's a proven commodity, will play less so we can finally develop JDD and Dubnyk – and can always be moved at the deadline to a contender if necessary. Win-win in my mind.

  • Hippy

    Tyler wrote:

    Yeah Willis. You might as well just ditch all of these “historic examples” and “evidence”. Your argument is lacking in truthiness.

    And, of course, it's another stats crony to the rescue.

    You go by statistical formulas that most of us don't need to consult, we know enough by what our eyes tell us.

  • Hippy

    a lot of the time, neither does the save percentage unless youre figuring out the "value" of each shot. shots from the perimeter are easier to save than shots in the wheel house. that one writer/researcher (forget his name off hand) tried to figure out that system of shot quality to see who the actual "Good goaltenders" are. but even that doesnt isnt the clincher. theres astronomical calculations that can help lead to success or failure. thats why hockey is great, because its chaotic. thats why most people consider just getting into the playoffs "the chance" they need to dance with order and chaos. the numbers dont mean crap. a bounce of the boards can change a series. or a dumb play by Ty Conklin can put the Oilers behind the 8-ball in a series. or have the Super Human Red Wings who have so many fabulous CORSI stats and other crapola lose in a series up 3 games to 1.

    stats in hockey are decent for helping create a possible model of "whats to come" / possible capacity but they tell nothing of the actual context of whats to come. GOOD.

    Osgood, individually, proved he ups his game and focus when the reward has more meaning. thats all you need. if it wasnt for Brad Stuart in game 7, Osgood would have probably won the Conn Smythe. maybe you should write up a report on Brad Stuart blowing the cup for them with bad passes and pinches and figure out some stats around it. "dumb plays = loss. yum."

  • Hippy

    GSC wrote:

    You’re going to extraordinary lengths to make yourself appear correct, and it’s unnecessary.
    Some, like myself, simply aren’t going to agree with you in arguments like this. Putting up article after article in defence of yourself isn’t going to sway those who disagree with you, me included.

    Yeah Willis. You might as well just ditch all of these "historic examples" and "evidence". Your argument is lacking in truthiness.

  • Hippy

    You're going to extraordinary lengths to make yourself appear correct, and it's unnecessary.

    Some, like myself, simply aren't going to agree with you in arguments like this. Putting up article after article in defence of yourself isn't going to sway those who disagree with you, me included.

  • Hippy

    I think there is way more to the story the GAA and Save % when it comes to goalies.

    Finding consistent goaltending is very hard. Only a few select individuals have the mental side to the game that allows them to compete mentally game in and game out and one season after another. (Jim Carrey anyone?)

    How would we have felt if the Oilers handed the reigns over to an unproven starter like Anderson (one season doesn't make him a proven starter).

    The fact that everyone took a pass on Biron on July 1 speaks HUGE volumes as to what he was looking for and also to why the Flyers we're so willing to walk away and go with former problem child Emery.

    Another consideration when signing a UFA has to be player character and their dressing room presence and what their reputation does to the players in front of them. I believe teams play better when they are confident in their goaltender and reputations have a lot to do with that.

    Also when talking about JDD, he has to be good enough and athletic enough for the Oilers to endure the three headed monster and be willing to give him the back up duties this year. I think everyone can agree that his development has been unfairly compromised by this team time and time again and I for one am looking forward to him getting a chance to prove himself.

  • Hippy

    I know I've been on your @ss for this topic lately, but your telling me you'd rather have Roman Chechmanek or Manny Fernandez for a higher sv% than Eddie Belfour, bc/ W/L don't matter?

  • Hippy

    […] here is sick of me preaching doom and gloom about the Oilers’ coming season. Read more: Why I Could Care Less About (Goalie) Wins and Losses :a-general-point, article, article-isn, general-point, his-control, one-particular, rather, […]