Skater Goals Against Averages And Save Percentages

by J.P. on Dec 9, 2008 6:42 PM EST Comment 24 comments

On Sunday, Mirtle took a look at how a goalie's play at five-on-five impacts a skater's plus-minus and questions whether or not there's enough evidence to establish that the relationship is a two-way street (i.e. skaters' poor defense leading to higher quality shots and thus poorer numbers for the goaltender). Tyler at mc79hockey goes a little deeper into the analysis, and the two come to the conclusion that it's too early to read too much into the numbers, as they're largely somewhat random - a big save or two here or a soft goal there and the list changes dramatically.

Nevertheless, it's interesting to look at the numbers, especially when you're just comparing players on the same team. So, without further ado, the five-on-five goals against averages, save percentages and (gasp!) plus-minuses for every Cap skater who has played in at least seven games (stats through Saturday night):

So what do these numbers tell us? For one, Sami Lepisto had been, through Saturday at least, the beneficiary of the best five-on-five goaltending on the team... luckily for him. On the flipside of that coin, Tom Poti, Tyler Sloan and Brooks Laich have good plus/minuses, despite not getting much of a bailout from their netminders, not to mention (in the blueliners' cases) being on the ice for a ton of defensive zone draws, which no doubt lead to better scoring opportunities and worse goalie numbers.

To that last point, though, it's interesting that Michael Nylander and Boyd Gordon have taken 212 and 211 even strength faceoffs on the season, respectively, but Nylander's goalie numbers are far worse despite probably taking as few defensive zone draws as Gordon has taken offensive zone faceoffs.

Another point - and this gets back to the question of randomness versus correlation and causation - goalies are performing very well behind John Erskine. Is he the reason for this (perhaps he's keeping the crease clear and shots to the outside) or is it dumb luck? I'm not sure what the answer is, but I do know that I'd rather have him in the lineup than Bryan Helmer or Sean Collins.

What do you think? Does anything there jump out at you?

Update: Further to DMG's comment, Rage sent along the following graph that plots 5-on-5 GAA against Quality of Competition:

Basically, Karl Alzner = teh awesome (though the sample size is small) - he is playing against the toughest competition of anyone on the team at five-on-five andhas the third-lowest GAA.

3 recs | Comment 24 comments | Add comment |

Story-email Email Printer Print

Comments

What jumps out at me in the inverse correlation between the save percentage and GAA numbers and quality of competition. For example. Lepisto, Jurcina, Erskine, and Brashear are all near the top and those are guys who generally play against the opposition’s weakest players, while Poti, Gordon, and Schultz are all near the bottom and are guys who play against the opponent’s best scorers.

Honestly, it does give me a little hope Lepisto as a guy who can maybe be a contributor as a sixth defenseman and powerplay specialist.

Great numbers for Alzner too, especially since he hasn’t really been given the luxury of playing 12 minutes a night against the other team’s worst forwards.

by DMG on Dec 9, 2008 6:52 PM EST reply actions 0 recs

I was thinking of putting together a chart like that but I couldn’t figure a source/way to measure quality of competition.

Rage, you the man!

by DMG on Dec 9, 2008 8:14 PM EST up reply actions 0 recs

You know what they say about great minds.

by RickyRage on Dec 10, 2008 10:00 AM EST up reply actions 0 recs

Nice chart

I think it makes it clear that Alzner needs to stay up as much as possible if/when the team gets past this injury bug. The guy can play at the NHL level, will continue to improve with experience, and I’d rather see him on the ice than Mo or Juice, both of whom are likely gone next year.

by grapejoos on Dec 10, 2008 10:52 AM EST up reply actions 0 recs

I disagree that the chart directly indicates anything, really. I just wanted to see how strongly correlated the QComp and GAA were… and it is pretty strongly correlated.

Except for those Lepisto/Fehr/Alzner who have played 7/16/7 games each. The only people on that list who have played fewer games than Fehr are Lepisto, Alzner, Semin and Fedorov.

The point I guess is that it’s a small sample for both young d-men. We’ll see how it plays out.

by RickyRage on Dec 10, 2008 12:50 PM EST up reply actions 0 recs

question

since my own indifference has been regaled: what is your take on the statistical interpretation we layfolks should consider when we look at these stats?

by bigonetimer on Dec 10, 2008 1:04 PM EST up reply actions 0 recs

I think this says, to a large degree, that your competition is the primary driver of the GAA. You can’t make the causal statement that your own play DRIVES the GAA.

The second factor in the equation, your own play, obviously has an effect. But it is much smaller than the quality of competition. I.e. the qual of competition gets your GAA against near the regression line; your own play determines if you fall above or below it.

by RickyRage on Dec 10, 2008 3:15 PM EST up reply actions 0 recs

and the league-wide regression line is

by bigonetimer on Dec 10, 2008 6:11 PM EST up reply actions 0 recs

that is, graphically, where does the regression line lie on your chart?

by bigonetimer on Dec 10, 2008 6:24 PM EST up reply actions 0 recs

The Brashear number is the least surprising in this sense: He takes very, very few defensive-zone faceoffs.

I think this is a better number for forwards than for D. I mean, if the Caps had been healthy all year (::and a good, hearty laugh was shared by all::), Poti+partner would have been on for far more than 33.3% of D-zone faceoffs, and their numbers would be correspondingly worse. (And Poti’s number is.) But when you jumble up the list with all those AHL D, it’s hard to guess who else took the most 5-on-5 D-zone faceoffs.

In a related story, while this is an excellent post, you really need to stop talking about data and analysis. Instead, please be a mindless homer. (Just beating others in the commentariat to it…)

Modern Art Notes

by TylerG on Dec 9, 2008 7:02 PM EST reply actions 0 recs

GAA is more telling than SV%

IMO, there is very little correlation between a goalie’s save percentage and who is on the ice at the time a save is made. I’ll grant that some teams are better than others at clearing the crease of traffic, giving their netminders a better view of pucks, mopping up rebounds, etc.

by bigonetimer on Dec 9, 2008 7:58 PM EST reply actions 0 recs

If you’re gonna bring that stuff in the house, you’ve gotta bring data! Look at the data in this post! And you bring an “IMO?” ;-)

Modern Art Notes

by TylerG on Dec 9, 2008 8:23 PM EST up reply actions 0 recs

yup, it's an opinion

based on 15+ years of playing experience. I can tell you from that perspective that there is very little individual players on the ice can do to help a goalie actually make a save once the shot is fired, other than clear traffic. To me, a more significant stat would be shot quality and number against.

Team systems, IMO, have a huge impact on an individual’s SV. Throw a loafer like Kozlov on the Wild and I guarantee you his percentages improve. A wide open, transition game like we play will yield more opposing odd man rushes, and more quality chances against, thus higher GAA and lower SV. Minnesota? Not so many.

by bigonetimer on Dec 9, 2008 8:52 PM EST reply actions 0 recs

based on 15+ years of playing experience. I can tell you from that perspective that there is very little individual players on the ice can do to help a goalie actually make a save once the shot is fired, other than clear traffic.

Yes, but the players on the ice have a lot of control over where the shot is being taken from and that shot quality you mention is partially built into the save percentage statistics for the players.

by DMG on Dec 9, 2008 8:58 PM EST reply actions 0 recs

so Sami Lepisto has more "control"

than Jeff Schultz?

Surely you jest.

by bigonetimer on Dec 9, 2008 9:07 PM EST up reply actions 0 recs

fyi

I’m not belittling these particular stats, nor the people who have a lot of time on their hands to crunch the figures. I’m just sayin’ that they’re not truly conclusive of anything.

Gimme a goddamn goalie who doesn’t give up a weak-ass, short side wraparound because he’s not paying attention with 8 minutes to play, and I’ll give you a boat load of beauitiful defensive stats for an entire team.

by bigonetimer on Dec 9, 2008 9:13 PM EST up reply actions 0 recs

they are very conclusive. MORE DONALD BRASHEAR. :-)

by macvechkin on Dec 10, 2008 2:53 AM EST up reply actions 0 recs

I don’t think that any more than I think Erskine does a better job than Poti of forcing his opponents to take bad shots.

I also don’t think anything I said leads to that conclusion. As I pointed out, there’s an inverse relationship between the numbers and the quality of competition the guys are playing against – to assume that Lepisto is a better defensive player based on these numbers would be to assume that he and Schultz played under the exact same circumstances, which is obviously not the case. However, if you had some experimental setting and you had all variables controlled except the players you were evaluating, I would think save percentage would give you some information as to who’s better.

But really, to suggest that I meant to suggest Lepisto is a better defensive player because of the higher save percentage when he’s on the ice compared to Schultz doesn’t make any less sense than using the metric you said you’d favor, GAA, in which Lepisto is also markedly better than Schultz.. Lepisto’s GAA is 19% better than Schultz’s while his save percentage is only 6% better, so if anything the GAA is much suggestive of Lepisto being a better defensive player.

The bottom line is that Schultz is a better player and faces better competition. If the difference in competition faced is greater than the skill difference of the two player, it’d make sense that the guy who’s a worse defensive player but facing weaker competition would have better numbers.

by DMG on Dec 9, 2008 9:29 PM EST up reply actions 0 recs

you make an interesting point

regarding quality of competition faced by individuals impacting these stats. And generally speaking, yes, our best defensive pairs will play the opposing top lines. But it’s not a static universe out there: lines change on the fly, coaches are angling to get better matchups in transition—you know that, because you actually play, if I recall an allusion to that fact in a prior post. Case in point: the GW goal Sunday. Eric Staal v. Sami Baby, an opposing first liner vs a weak D man (and man, that GV was weak).

The fact is, the results of a quantitative study like this one are not necessarily supported by qualitative evidence. My comment about GAA should not be taken as “favor” either. I thought it was notable that Feds, our best defensive C, had the lowest GAA in that group. Clark, Koz and Flash not as, ah, defensive-minded as they should be. There are some interesting data points. They’re fun to play with. But I don’t think GMGM or BB will be making any personnel decisions based on them.

by bigonetimer on Dec 10, 2008 11:40 AM EST up reply actions 0 recs

What we can deduce is that when The King enters the building permanently, (you know, when we have cap room), he’s shaping up to be a top pairing D-man like we all expected. In that you can now move Shultz down a pair and assuming all things remain the same would help out the entire defense by letting them play against their um, “natural” level of comp.

by wittcap79 on Dec 10, 2008 7:34 AM EST reply actions 0 recs

And shame on J.P. for not approving my Facebook request!

by wittcap79 on Dec 10, 2008 7:35 AM EST reply actions 0 recs

Wha?! I log onto Facebook approximately once a month. Next time I do, you’ll be approved.

Japers' Rink: Hockey blogging from the most powerful city in the world

by J.P. on Dec 10, 2008 3:51 PM EST up reply actions 0 recs

Also

Is anyone else shocked and awed by Lepisto’s placement on this list/graph? He’s decent on the PP but a horror show behind his own net.

by grapejoos on Dec 10, 2008 10:54 AM EST reply actions 0 recs

Yes, but note – this does not include the ‘Canes game, which was Sami’s nightmare game.

Japers' Rink: Hockey blogging from the most powerful city in the world

by J.P. on Dec 10, 2008 3:52 PM EST up reply actions 0 recs

Join the community and dive in!

Search

FanPosts

Community blog posts and discussion.

Recent FanPosts

New FanPost All FanPosts

FanShots

Quick hits of video, photos, quotes, chats, links and lists that you find around the web.

Southeast Standings

	GP	W	L	OTL	PT
Washington Japers' Rink	82	50	24	8	108
Carolina Canes Country	82	45	30	7	97
Florida Litter Box Cats	82	41	30	11	93
Atlanta Bird Watchers Anonymous	82	35	41	6	76
Tampa Bay Raw Charge	82	24	40	18	66

(updated 4.13.2009 at 12:21 PM EDT)

Japers' Rink: An SB Nation Community

Login Bar

Network Bar

Hockey blogging from the most powerful city in the world