Lehtonen vs Hedberg
by Hawerchuk on Oct 12, 2009 8:28 PM EDT
in Statistics
Email
PrintThe Atlanta Thrashers have two goaltenders who have essentially split duties for the last two years, Kari Lehtonen and Johan Hedberg. By every metric, Lehtonen is the better goalie of the two: they've faced roughly the same quality of shots, but Lehtonen has a better save percentage and a better goals-against-average, not just overall, but every single year.
What's even stranger is that Lehtonen consistently faces more shots than Hedberg per 60 minutes:
| Hedberg | Lehtonen | Diff | |
| 06-07 | 28.4 | 31.6 | 11.5 |
| 07-08 | 31.9 | 34.6 | 8.2 |
| 08-09 | 30.6 | 34.3 | 12.0 |
| 30.6 | 33.2 | 8.5 |
Where does this difference come from? It's not rebounds - they each give up approximately 9% of their shots on rebounds. So what else could it be?
It's also not that Lehtonen is somehow getting starts against better teams. This table shows the shots/60 for opposing teams, prorated to how much ice time each goalie had against each team:
| Hedberg | Expected | Diff | Lehtonen | Expected | Diff | |
| 06-07 | 28.4 | 29.4 | -3.6 | 31.6 | 29.9 | 6.0 |
| 07-08 | 31.9 | 30.1 | 6.3 | 34.6 | 29.7 | 16.4 |
| 08-09 | 30.6 | 30.4 | 0.5 | 34.3 | 30.8 | 11.3 |
| 30.6 | 30.0 | 2.0 | 33.2 | 30.1 | 10.5 |
So if it's not rebounds and it's not due to a better group of opponents, what is it? It's certainly not luck - while the standard deviation in shots/60 was slightly higher for the group of teams Lehtonen faced, the odds are much, much less than 0.01% that it happened by chance. It's not a home/road issue either - Lehtonen was much more likely to play at home (54%) than Hedberg (42%), so if anything this makes Lehtonen's high shot total even less likely.
So what about the impact of Atlanta players who were in the lineup when each goalie was playing? At 5-on-5 over the last two seasons (including missed shots), it's very stark:
| PCT TOI | PCT TOI | Shot/60 | Shot/60 | ||
| Player | Lehtonen | Hedberg | Lehtonen | Hedberg | Delta |
| TOBIAS ENSTROM | 33.0 | 33.9 | 47.4 | 38.6 | 8.8 |
| NICLAS HAVELID | 30.2 | 30.8 | 49.3 | 41.6 | 7.6 |
| ILYA KOVALCHUK | 30.2 | 30.9 | 46.0 | 44.2 | 1.9 |
| GARNET EXELBY | 26.5 | 29.8 | 45.9 | 43.2 | 2.7 |
| VYACHESLAV KOZLOV | 27.0 | 26.2 | 41.1 | 41.2 | -0.1 |
| TODD WHITE | 25.7 | 26.0 | 45.3 | 43.9 | 1.4 |
| ERIC PERRIN | 23.5 | 24.8 | 42.7 | 43.8 | -1.0 |
| CHRIS THORBURN | 19.5 | 18.6 | 46.8 | 38.5 | 8.2 |
| BRYAN LITTLE | 17.5 | 22.9 | 42.8 | 43.7 | -0.9 |
| KEN KLEE | 15.9 | 17.4 | 40.3 | 41.8 | -1.5 |
| COLBY ARMSTRONG | 17.2 | 15.8 | 42.9 | 39.6 | 3.3 |
| RON HAINSEY | 15.6 | 15.4 | 44.1 | 44.0 | 0.1 |
| BOBBY HOLIK | 14.6 | 14.5 | 45.5 | 38.2 | 7.3 |
| ERIC BOULTON | 15.8 | 14.2 | 48.4 | 43.1 | 5.3 |
| ALEXEI ZHITNIK | 14.5 | 14.2 | 43.7 | 47.3 | -3.6 |
| JIM SLATER | 14.2 | 16.6 | 49.0 | 39.8 | 9.1 |
| MARTY REASONER | 11.8 | 11.8 | 41.9 | 41.4 | 0.5 |
| BORIS VALABIK | 10.2 | 10.9 | 44.9 | 46.4 | -1.5 |
| MARIAN HOSSA | 10.0 | 12.7 | 43.3 | 41.8 | 1.4 |
| PASCAL DUPUIS | 9.8 | 10.6 | 42.7 | 38.2 | 4.5 |
| STEVE MCCARTHY | 12.5 | 9.4 | 44.1 | 46.7 | -2.6 |
| MARK RECCHI | 12.7 | 7.5 | 45.8 | 49.2 | -3.5 |
| BRAD LARSEN | 7.5 | 6.7 | 44.2 | 38.5 | 5.7 |
| ERIK CHRISTENSEN | 10.8 | 6.7 | 44.5 | 45.4 | -0.9 |
| ZACH BOGOSIAN | 12.4 | 6.6 | 46.4 | 42.6 | 3.8 |
| COLIN STUART | 8.5 | 6.5 | 45.9 | 39.4 | 6.5 |
| NATHAN OYSTRICK | 6.4 | 12.3 | 40.1 | 39.8 | 0.3 |
| MATHIEUSCHNEIDER | 11.8 | 5.8 | 42.3 | 41.4 | 0.8 |
| RICHPEVERLEY | 8.8 | 5.6 | 51.1 | 40.9 | 10.2 |
| MARKPOPOVIC | 4.9 | 7.1 | 38.0 | 39.1 | -1.1 |
For reasons that are unclear to me, many of Atlanta's player give up dramatically higher shot rates when Lehtonen is in goal. The top D pair of Enstrom and Havelid is particularly odd - how could they possibly give up so many more shots when Lehtonen's in goal given that the caliber of opponents is the same as when Hedberg's in goal? And it's not as though this effort (or lack thereof) is matched on the offensive end - Atlanta had 34.6 shots on goal per 60 regardless of who was in goal. A hockey mystery remains unsolved...
***update***
Some have suggested that the defensemen are aware that Hedberg is a weaker goalie and they consequently try to block more shots - that's not borne out by the stats: 12.4 blocks/game when Hedberg's in vs 12.6 for Lehtonen.
***update #2***
Commenter Likens has suggested that the difference in shot totals may have something to do with the score when each of these goalies are playing. Because Hedberg is demonstrably worse than Lehtonen, he ends up playing with his team behind more often. When teams are behind, they allow fewer shots, mostly because their opponents are playing more defensively.
It's true that Hedberg played behind more than Lehtonen over the last two seasons - he was down 1-3 goals 29.8% of the time versus 27.3% for Hedberg. This explains some of the difference. But what caught my eye was that when the game was tied, Lehtonen saw 47.0 shots/60, while Hedberg saw 43.3 shots/60. That's a big difference, and again, not something we can write off as luck.
27 comments
|
0 recs |
Do you like this story?
Comments
Great post
This is top-notch stuff. I see a bit of this brewing up in Chicago as well. A small sample size, but a stark difference between the amount of shots that go on Cristobal Huet in comparison to Antti Niemi. As the season progresses, the mystery that was in Atlanta could spread..
I vaguely remember...
… something like this on BWA about a year ago. This is the great unanswered question for the Thrashers, really.
Actually, ok, it’s like one out of 637 questions, but that’s beside the point.
Anyway, Thrashers fans have always noticed this, and we think that it comes down to something that isn’t statistically related. Half the fans believe that they have greater confidence in Kari’s abilities to stop the puck, so they lighten up on D and focus on offense. The other half looks at past statements by Kovalchuk pretty much admitting that the team wants to play better for Hedberg, and thinks that maybe it’s a subconscious thing. Maybe the team is tired of his unreliability and the games he might have cost them by being out, and they just slack off. Hedberg is highly respected in the locker room and in the organization as a whole, so that might play into it too.
Reporter: There`s a "stamp out the Beatles movement" underway in Detroit. What are you going to do about it?
Paul McCartney: We`re going to start a campaign to stamp out Detroit.
Thanks for putting some effort into this. The question has plauged many of us (especially Falconer) for a long while. I think it would be very fascinating to see if the differences are as stark while Pavelec is in as the Thrash #1 netminder. That could go a long way in answering the debate hildymac alludes to above—is the difference because of respect for Moose or faith in Kari.
It is a head scratcher. The one thing Enstrom-Havelid-Holik group have in common is that they were regularly matched against the other team’s best forwards. But how does a goalie have so much influence over shots. ’
As Hildymac alluded to, but the only thing I can think of is that perhaps they trust Lehtonen more and therefore don’t block as many shots when Lehtonen is in net? If they don’t trust Hedberg the D might go crazy in trying to prevent NHL SOG. Is there a big difference in Opposition Shots Blocked between these two netminders?
The other theory is the “screw you Lehtonen” theory that some Thrashers skaters don’t like him and don’t try as hard when Kari starts. I find that hard to accept.
All things Thrashers + stats: www.birdwatchersanonymous.com
Well, it’s not blocked shots either. (See update.) If it was because of reduced effort by the D, you’d think heads would roll.
Thanks for checking the blocked shots. One more potential explanation can be crossed off the list.
All things Thrashers + stats: www.birdwatchersanonymous.com
by The Falconer on Oct 13, 2009 1:30 PM EDT up reply actions
ive always thought from my own observations that Hedberg is a far superior puck handler to Lehtonen. And Hedberg is very aggressive at cutting off the dump in behind the net to assist his defenseman. Lehtonen not so much. Marty Brodeur is widely acknowledged as one the best goaltenders in this regard and if memory serves the Devils are consistently one of the stingiest teams in the NHL at shots allowed. The team’s style of play is often given credit for this, perhaps some of that credit is misplaced? Not sure what statistical information is available to see if that “skill” proves out to support a reduction in shots allowed.
No data on that. But I’d imagine the impact is pretty small. Hedberg’s save percentage is 14 points lower than Lehtonen’s, which isn’t made up for by additional goals allowed due to the higher shot volume.
i wasnt suggesting that it had an impact on his save percentage. merely suggesting that by cutting off the dump in and leaving the puck in a favorable position for the defense they can more easily clear the zone and therefore lead to fewer shots against. Isnt puck handling the reason Brodeur is sometimes referred to as an extra defenseman? And also the reason the league put in the “no handle” zone behind the net? to prevent goalies from playing the puck back there and in theory leading to more offensive opportunities? obviously I have no statistical data to back this up. just thinking out loud.
My hypothesis:
Because Lehtonen is the superior goaltender in terms of save percentage, it can be assumed that his team plays with the lead more often when he is in net. Conversely, it can be assumed that Atlanta plays more often from behind with Hedberg in net, who has a much lower save percentage.
As goal state biases shot ratio in favor of the trailing team, this might account for why the Thrashers surrender more shots when Lehtonen plays, and fewer when Hedberg plays.
What do their respective SA/60 figures look like when goal state is controlled for (for example, when the score is tied)?
It’s unlikely that this effect, if it exists at all, accounts for all of the difference. However, I suspect that it may account for part of it.
Absolutely agree (I was just about to post the same theory, really). Also add in the fact that when Hedberg enters the game as a non-starter he is usually entering a game where the Thrashers are trailing (badly). And a quick glance at game logs over the past two years shows that 1. Hedberg gives up a lot of early goals 2.even when he wins as a starter, an inordinate number of the games involve comebacks in the third period, ie, the Thrashers have been trailing for most of the game and their opponent has been defending a lead for much of the game.
by Big Picture Guy on Oct 13, 2009 4:18 PM EDT up reply actions
I can’t speak to game stats (i.e. leads) but I do know that Hedberg benefits from a much stronger offense when he starts. Despite having a terrible SV% his Won-Lost record is better-than-expected because they Thrashers frequently post big fat numbers in the Goals For column when he is in net.
As Big Picture Guy says above, Johan gives up a lot of early goals—does that point to an over-confidnece hypothesis? Opposition team thinks “hey, look we’re already up a goal on this crappy Heberg guy and the Thrashers rank near the bottom of the standings…I guess we can let off the gas pedal now.” On one level, that probably does happen now and again in the NHL, but I have a hard time believing such a pattern would be consistent over three seasons worth of NHL data.
All things Thrashers + stats: www.birdwatchersanonymous.com
by The Falconer on Oct 13, 2009 4:57 PM EDT up reply actions
It’s not “over confidence”, teams go into a defensive shell when leading, they don’t press the issue, dump the puck etc. I am working on a project which involves going through play by play accounts manually(because the data I want can’t be scraped) and as J. Likens says “goal state biases shot ratio in favor of the trailing team”, I see this constantly.
by Big Picture Guy on Oct 13, 2009 5:21 PM EDT up reply actions
Over the last two seasons:
Lehtonen: up<=3 – 28.7%, tied – 37.0%, down<=3 – 27.3%
Hedberg: up<=3 – 29.4%, tied – 35.3%, down<=3 – 29.8%
Shots against/60:
Lehtonen: up<=3 – 53.4, tied – 47.0, down<=3 – 35.0
Hedberg: up<=3 – 51.1, tied – 43.3, down<=3 – 37.4
Although for all trailing situations, the shot rate against Lehtonen was higher than Hedberg (39.0 v 35.9) but we can throw out blowouts.
There’s no good reason the shot rate should be 3.7/60 higher against Lehtonen when the game is tied.
It doesn’t appear that goal state has much of an effect at all.
I guess that theory can be crossed off the list, too.
It changes based on how close to tied the score is
I think Tyler’s table here is basically the answer to your question.
It's not penalties
I thought that maybe players would protect Lehtonen more during scrums near the net because he is a very valuable player. That would lead to more penalties and thus more power plays and shots for the opposition in games where Lehtonen was in goal. For instance, if Lehtonen drew an average of 1 penalty more per game due to an undisciplined teammate drawing a penalty trying to protect him, that would account for the ~3 extra shots that Lehtonen sees per game.
So I added up the power play opportunities and short handed opportunities in every game last season and grouped them into which goalie played (Lehtonen, Hedberg, Pavelec). I threw out games that were split between 2 goalies. The results showed an even distribution of penalties for/against for Lehtonen and Hedberg. So, its not penalties.
by ThrashersRecaps on Oct 14, 2009 11:48 AM EDT reply actions
Blackhawks
Agree with Jermaniac about Chicago, especially after last night’s performance by Khabibulin. Maybe its me but there seems to be a bit of selective memory in Chicago about Khabibulin’s performance versus Huet’s in last year’s regular season, in that people assume Khabibulin was far superior.
Gabe, would appreciate any assessment based on your stats on this question.
Nice post, this is a question that has long baffled me as well. I’ve looked at the different variables mentioned here and I agree with the conclusions, there doesn’t seem to be anything behind them. It’s a particularly interesting problem because solving it should help us get much closer to an understanding of goalie effects on shots against.
It could possibly have something to do with how the goalies impact the flow of play, e.g. through puckhandling. What do the Corsi results show, are there more shots being taken by Atlanta while Hedberg is on the ice?
If not, then it seems like it might have little to do with the goalies themselves or Atlanta’s team defence, and might be mostly driven by the shooters on the other team. I’m wondering if the book on Lehtonen around the league is that you should throw pucks on him from everywhere. Does Lehtonen face more shots from less dangerous areas?
by TheContrarianGoaltender on Oct 16, 2009 12:44 PM EDT reply actions
Comments For This Post Are Closed
Secondary Sidebar
User Tools
FanPosts
Community blog posts and discussion.
Certain photos copyright © 2012 by Associated Press or Getty Images. Any commercial use or distribution without the express written consent of Associated Press and Getty Images is strictly prohibited.
