GUEST COLUMNIST: A look at goalies who washed out of the NHL

You know, one of the most fun things about this blog is that other people who don’t (yet) have their own blogs do incredible work, and then I – having smartly volunteered to publish their work here! – get a sneak peek at that work as part of the process of getting it ready to post as a guest column.

And if you’ve read SwedishPoster’s recent Magnum Opus on drafting or FrJohnK’s post on team defense and goaltending, you know what incredible work it is.

Today’s guest columnist is WheatNOil, who ironically is one of the guys who nagged me most that I should start a blog!  Ha!

After my previous post looking at goalies through the lens of how their cumulative sv% stabilized over their careers, Wheat and I started corresponding.  I had noted of course that my previous work had been intended as more of a ‘longitudinal study’, and suffered from both small sample size and from selection bias, as I was looking at just a handful of mostly successful goalies.  Wheat, in the comments section, asked exactly the right question, which is whether or not the same patterns held true for goalies who had washed out of the NHL.  

And to his extreme credit, he went and did the work to find out!  Props also to TheGreatMutato, who ran some additional charts as well (you’ll find those in the comment section of that last goalie post), and inspired some of the follow on work.

So for your edification, here is WheatNOil‘s look at goalies who washed out of the NHL, whether their early NHL careers might have helped predict that, and what that might tell us about early career goalies today.


Washed Out Goalies

aka WheatNOil gets curious and goes searching for random things

Search strategy:

  • Used Rob Vollman’s Goalie Data Set from 1998 to 2014 to find goalies who have got a foothold in the NHL but were unable to maintain their hold, washing out
  • Goalies must have played a minimum of 50 games but no more than 200 games
  • Goalies must have played in their 50th game prior to age 30 (giving them a fair chance to hit 200 games before their inevitable age related decline and limit the study to “young” promising goalies)
  • Goalies are no longer active, thus having “washed out” (defined as not having played a game in the NHL in this past season)
  • Must have played their first game since the 1997/98 season
  • Save percentage data from war-on-ice (since 02/03) and nhl.com (pre 02/03); note even-strength save percentages pre 02/03 are not readily available on a game-by-game basis, so the season-end EVSv% had to be used… hence some charts don’t have curves that go right to the Y-axis

Findings

  • 17 total goalies meet this criteria
    • Dan Blackburn, Jani Hurme, Erik Ersberg, Dany Sabourin, Jeff Deslauriers, Hannu Toivonen, Mika Noronen, Sabastien Caron, Milan Holmqvist, Jani Hnilicka, Pasi Nurminen, Michael Leighton, Mikael Tellqvist, Joey MacDonald, Jussi Markkanen, Curtis Sanford, Pascal Leclaire
  • 2 (Dan Blackburn and Jani Hurme) had extremely serious injuries that end their NHL careers and thus were excluded. Blackburn was unable to use his glove hand ever again. Hurme missed two seasons due to injury and then went back to Europe. Neither by eye look like they would impact the conclusion
  • Of the 15 remaining goalies, 5 played less than 100 games & 10 played more than 100 games
  • Of the 5 that played less than 100 games
    • 4 out of the 5 actually had lower career cumulative save percentages than 50 game save percentage, though a couple were close (not surprising given they didn’t play many games beyond 50 to affect their save percentage)
    • Only Ersberg had an even-strength save percentage that was in the ball-park of some of the young goalies the Oilers are looking at now through 50 games (50 game Sv% of 0.924… for reference, Scrivens was also 0.924, Talbot 0.931, Lack 0.922)
    • Ersberg, started his career like a house on fire and was starting to come back to Earth (though still doing very well). The Kings sent him down to the AHL and he refused to report. So they cancelled his contract and he went back to Europe where he had a successful career in the KHL. There may have been a goalie there.
  • Of the 10 that played more than 100 games but less than 200
    • 9 of the 10 had equal or higher cumulative career save percentages than their 50 game save percentage… so even most of the guys that washed out may have been stabilizing out at a higher save percentage
    • Only 1 out of the 10 had a 50 game even-strength save percentage close to the goalies in our study. It should be noted that Ev-Sv% has risen gradually over the years. However, a quick glance at a history of Ev-Sv% indicates that none of the other goalies were at or above league average
    • The 5 goalies who played between 100 and 125 games had 50 game save percentages of around 0.900… so fairly poor, even accounting for changes in league average save percentage over the years
    • The 5 goalies who played more than 125 games had save percentages greater 0.907, so better but well below league average (though Sanford was close) except for…
    • Jussi Markkanen. *shudder* Jussi Markkanen is the exception to the rule. A 50 game save percentage of 0.923 (!) is head and shoulders above all the washed out goalies except for the promising Ersberg. None crashed and burned worse than Markkanen though. He was flying high and should have been a promising young goalie. None of the other goalies who met the criteria for this study followed his pattern.

Summary

  • The goalies that washed out prior to their 200th game did not carry impressive save percentages at the 50 game mark, which is not surprising
  • Goalies that got to at least 100 games had raised their cumulative save percentage at or above their 50 game mark despite fluctuation prior to this. Could this 50 game cumulative save percentage be a bit of a “baseline” for a goalie’s intermediate future? This would offer some support G Money’s article of selected elite goalies (though Mason appears to be an exception).
  • Obviously more studies need to be done. Maybe the next step is to look at goalies that had 50-game cumulative save percentages of 0.920 or greater (or league-average or greater to account for differences through the years). Clearly most goalies that start off that well get longer careers and so the games played window will have to be extended well past 200 to see if they burn out later.  (Editor’s note: After he did the work on this topic, WheatNOil pulled those numbers too, and that will be the next post, coming soon)
  • EXCEPTION: The curious case of Jussi Markkanen, which should serve as a cautionary tale.  Compare Markkanen’s curve to Scrivens’ curve.  If Scrivens is like the goalies selected by G Money or even if he’s like the goalies that ended up washing out, he should rebound. Therefore, he’s a good bet to rebound… UNLESS he’s Jussi Markkanen. That’s why there’s no guarantee… especially if you’re an Oilers fan.

Editor’s Note:

FWIW, my conclusion out of this is pretty simple: if you’re an NHL GM looking to hire a young goalie, I would suggest you STAY AWAY FROM ANY GOALIE WITH A LOUSY SV% AT 50 GAMES.  

Now, this look at goalies still incorporates selection and survivor bias, because Wheat did a retrospect look at goalies who we already knew washed out.  So you might argue that maybe there are goalies who had poor sv% at 50 games who went on to become stars.  We’ll look at that next.  

The answer in general: no.  If they start poor, they usually end poor.  

So stay away.  It’s a fools game to bet that they’re going to improve.


Charts For the Goalies Included In this Study

(ordered by games played)

Erik Ersberg – Ended up with a solid KHL career

Games = 54, 50 Game = 0.924, Career = 0.925

Dany Sabourin – tough beginnings

Games = 61, 50 game = 0.908, Career = 0.904

Jeff Drouin-Deslauriers – First Oiler on the list

Games = 61, 50 Game = 0.911, Career = 0.910

Hannu Toivonen – Had an okay few games and then fell well below .900

Games = 66, 50 Game = 0.903, Career = 0.892

Mika Noronen – fell off a cliff around game 60

Games = 77, 50 Game = 0.907, Career = 0.889

Sabastien Caron – pretty consistently not good enough

Games = 105, 50 Game = 0.899, Career = 0.901

Johan Holmqvist – stabilizing out to be pretty poor and sub-.900

Games = 110, 50 Game = 0.895, Career = 0.898

Milan Hnilicka

He hit his 50th game pre 02/03 so can’t get game-by-game even strength data from war-on-ice, only year-end even-strength data from NHL.com; finished a year at 38 games so that’s the closest estimate, not a useful chart but included for completeness

Games = 117, *38 game = 0.902, Career = 0.911

Pasi Nurminen – stabilizing out by game 110?

Games = 123, 50 Game = 0.903, Career = 0.905

Michael Leighton – Lack of volatility in the early days, slowly rising from game 80 but not quite good enough

Games = 125, 50 Game = 0.898, Career = 0.906

Mikael Tellqvist – started to improve by the end and may have ended up as a serviceable back-up or better

Games = 128, 50 Game = 0.909, Career = 0.914

Joey MacDonald – trending down by the end, his backup career stretched over multiple seasons and he was 34 by the last one

Games = 141, 50 Game = 0.907, Career = 0.911

Jussi Markkanen

Gah! Solid save percentage without TOO much volatility early on… collapsed during the 2005/06 season and didn’t recover. Collapsed around the 90 game mark. This is the one cautionary tale on expecting Scrivens to bounce back.

Games = 156, 50 Game = 0.923, Career = 0.906

(Extra strange true story: I use ShareX and imgur to capture and upload these images for embedding.  I was able to do that with ALL of the images you see here – except for this one.  This one gave an error on upload.  It was the exception.  Figures, doesn’t it?)

2015-06-18_11-40-04

Curtis Sanford – Fairly consistent with slow improvement, probably a decent back-up; 50-game save percentage was only 3 points off league average at the time

Games = 159, 50 game = 0.913, Career = 0.915

Pascal Leclaire – Was he rebounding by the end? Tough to say, but a fair bit of volatility, ended up with a cumulative save percentage equal to his 50-game number though

Games = 185, 50 Game = 0.908, Career = 0.908

Advertisements

4 thoughts on “GUEST COLUMNIST: A look at goalies who washed out of the NHL

  1. Thanks for posting this G Money! I have to say, upon reading it back, I wish I had written it out a bit clearer and not rapidly typed it out in an email, but the final conclusion is really what you’ve summarized… at 50 games, goalies don’t magically get better. If they’re poor after 50 games, they’re probably poor goalies.

    The average even-strength save percentage of all the goalies in this study at 50 games is .907 (.906 if you exclude Ersberg who left for contract reasons). That’s pretty poor and it’s not surprising. Markkanen is the outlier. I wish I had adjusted save percentage numbers for him, but he pre-dates those numbers from war-on-ice.

    Like

  2. Good stuff.
    Based on this info, I am listing a few goalies who we could keep on eye on moving forward.
    Jacob Markstrom 43 games .898, 37% Quality Start.
    Jake Allen 52 games started, .911, 52% quality start.
    Andrew Hammond 25 games started, .942, 78% quality start.
    Cam Talbot 53 games started, .931, 62% quality start
    Martin Jones 29 games started, .923, 62% quality start
    Annti Raanti, 34 games started, .912, 59% quality start
    Eddie Lack 69 games started, .917, 55% quality start

    Like

    1. Lack is included in the next study because he’s had more than 1300 shots against. The rest haven’t yet but will be interesting to follow!

      I’m quite interested in the Quality Starts you’ve been bringing up. For instance, if Hammond continues that level of consistency, is that more predictive of future success than even his impressive save percentage.

      One clarification with the save percentages you quote… all the data in the study is 5×5 save percentages, not all situations. Based off that, Allen rises to 0.917 (.924 adjusted) after just over 1000 shots for instance. Hammond is still an amazing .938 (.940 adjusted) but after only 619 shots. It’s surprising that Hammond’s 5×5 save percentage is LOWER than his all situations Sv%. He must have killed it on the PK!

      One person to look out for, which you’ll see in the next post, is Khudobin. He’s fallen down a bit since his 50 game (and his 1300 shot) mark, but he may be prime for a rebound. If so, Carolina may be in a nice spot after Ward’s contract is up next year.

      Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s