Thursday, January 28, 2010

Shots Against

This is just a short post on a little idea I had a while back and have been meaning to get around to doing. It's on the topic of the relationship between save percentage and shots against. One way to put a team in the best possible context is to look at other teams in the league that are similar. For example, if it is easier to play goalie on a team that allows a lot of shots against then it is likely that other teams with a similar shot profile would have the same advantage.

I looked at the situational save percentage numbers for the five teams that were closest in shots against per 60 minutes to the New Jersey Devils and the Florida Panthers from 2003-04 to 2008-09. New Jersey and Florida are two teams that are often brought up when various theories are advanced about whether the number of shots per game a goalie faces has an effect on their save percentage. I also reran the numbers using a few possible shot bias/goalie shot effect adjustments (adding one shot and two shots to New Jersey's totals and subtracting one from Florida's).

The result? There doesn't seem to be any significant effects from different levels of shots. If either of the teams are outliers then they also stand out against teams with similar shot prevention.

League average: .918 EV SV%, .866 PK SV%, .912 PP SV%

TeamAdjustmentEV SV%PK SV%PP SV%
New Jersey0.918.868.906
New Jersey+1.916.865.918
New Jersey+2.919.865.915
Florida0.916.867.913
Florida-1.916.865.911

14 comments:

Agent Orange said...

"Team Adjustment EV SV% PK SV% PP SV%
New Jersey 0 .918 .868 .906
New Jersey +1 .916 .865 .918
New Jersey +2 .919 .865 .915
Florida 0 .916 .867 .913
Florida -1 .916 .865 .911"

Maybe I missed something but what are you doing with New Jerseys numbers here?

It looks like you are just adding shots but why are the save%'s dropping when a shot is added?

Anonymous said...

I was just about to say the same thing... Then again, all of the "data analysis" on this site is skewed or biased to favor Luongo over Brodeur (hence the Panthers). We'll see who Canada's GM's prefer... if its Brodeur, I guess we will just have to ponder how they could have overlooked this analysis!

The fact is, as the author of this blog knows, ever since the lockout all that Brodeur has done has improved his save percentage and Luongo's has just gone down (but apparently switching teams has nothing to do with it). I guess, based on their save percentages (which the author deems the most, if not only, valid stat) means since 2004 Brodeur has just gotten better and Luongo has just gotten worse.

It is funny however because the whole point is that save percentage doesn't matter. In the later years when Brodeur's save percentage numbers have been higher, he hasn't made it past the 2nd round of the playoffs. Neither has Louongo. When it comes to hockey, it a team sport, and the most important stats are Wins. I don't care how many shots you save, if you let in that one goal that allows your team to lose, you blew it. So when you are in a double OT game facing elimination in the playoffs, it doesn't matter what saves you made all series and what your save percenage is, it matters that you don't allow your team to get eliminated from the playoffs by raising your arm to try to get the ref's to call a penalty while you let a weak 59 foot wrister from the blue line go by you. (Source: http://www.youtube.com/watch?v=LBLbVUPOoh8 )

Anonymous said...

Different 'Anonymous' here. This is not that relevant to this post outside of the fact that I think Oz has faced some shots in some games as bad as Florida lately, but I am going to make it anyway. I would like to know why you keep selling Chris Osgood short.

Dude, call it whatever you will but he would have taken the Conn Smythe if Detroit had won the cup this year. I want to know why sports writers lavish the praise on Roloson, Giguere, and Ward, for their cup runs, but say nothing about Osgood.

The 2006 Edomnton Oilers, and the 2009 Detroit Wings, were basically the same team. Very good teams, but struggled with mediocre goal-tending all season. In the playoffs they suddenly got the good goal-tending they had needed all year, and they each nearly won the cup with it. Edmonton would have won the cup if Rollie was healthy for the finals, and Detroit would have if Datsy and Hossa were healthy for the finals.

Osgood stole games one, two, and five for Detroit, and was strong in all the others. He almost stole the Stanley cup from a Penguins team, that was better than Detroit. Nobody with a brain would not call Oz one of the top five goalies of the decade.

Anonymous said...

What is great about your analysis is that you are trying to prove that the number of shots has no effect on save percentage, and what your analysis actually proves is that Brodeur is not a fraud. If you perform a statistical analysis using only 2 teams, you clearly do not understand statistics. Using 5 years of data for 2 teams is much less accurate that using 1 season of all 30 teams. Which I did. For this year and last year. In each, the linear trend is clear, and almost identical. The lowest shot teams in the league have an expected .90 save percentage and the highest have an expected .915 - in both analyses.

Now for the kicker - your analysis shows that a low shot team of NJ has the same save percentage as a high shot total team Florida (who happens to be the highest in both seasons I used). So what you are seeing is that New Jersey is an outlier. Florida's save percentage should be .915 which it is, and New Jersey's should be lower because they allow significantly fewer shots. So NJ has a higher save percentage than they should, according to their shot total. So only one thing can account for their team exceeding their expected save percentage: the great Martin Brodeur - thanks for showing his true value.

Anonymous said...

yep this analysis is garbage, and biased in its presentation to say the least. maybe we are better of calling philip the fraud

overpass said...

Maybe I missed something but what are you doing with New Jerseys numbers here?

It looks like you are just adding shots but why are the save%'s dropping when a shot is added?


This is how I read it. From the post:

I looked at the situational save percentage numbers for the five teams that were closest in shots against per 60 minutes to the New Jersey Devils and the Florida Panthers from 2003-04 to 2008-09.

Those aren't the actual numbers for New Jersey and Florida. They are the numbers for other teams with similar numbers of shots faced.

New Jersey = teams that face a similar # of shots/game to New Jersey. New Jersey+1 = teams that face about 1 more shot/game than New Jersey.

The data presentation is a bit unclear but I'm pretty sure my interpretation is correct.

Anonymous (haakon84?):Now for the kicker - your analysis shows that a low shot team of NJ has the same save percentage as a high shot total team Florida

I don't think you interpreted the analysis correctly. Rather, the analysis shows that teams that face similar shots to New Jersey (low shots) have the same save percentage as teams that face similar shots to Florida (high shots).

Your game-by-game results do not appear to hold on a seasonal level.

The Contrarian Goaltender said...

Sorry for the confusion. I can't imagine how someone could interpret my presentation in this post to be biased, but obviously I did not explain things well enough. Overpass has it right, I'm looking at the five closest teams. For the adjusted figures, I added or subtracted from the Devils' or Panthers' actual total and then took the five closest teams to the revised number.

It's possible that New Jersey or Florida are outliers in the way playing there affects goalie save percentages. But if either of them happen to be outliers that is because of scorer bias or team defensive play or something else other than just the shot totals.

In the aggregate, there is little to no relationship between shots against and save percentage in the NHL, and that has been confirmed by every study that I am aware of.

The Contrarian Goaltender said...

I don't care how many shots you save, if you let in that one goal that allows your team to lose, you blew it.

If that's your position, then you and I are probably going to disagree on almost everything in goaltending analysis, it's as simple as that.

Winning is a team objective. It makes sense to evaluate teams. It is unfair to evaluate individuals based on the results of others.

The Contrarian Goaltender said...

The 2006 Edomnton Oilers, and the 2009 Detroit Wings, were basically the same team...Osgood stole games one, two, and five for Detroit, and was strong in all the others...Nobody with a brain would not call Oz one of the top five goalies of the decade.

Wow, and here I thought I was the contrarian.

Here are two simple reasons why I don't give Osgood much respect.

1. Over the last decade (1999-00 to 2009-10), here are the records of Osgood's teams with him in net and without:

Osgood in net: 230-135-59
Another goalie: 230-113-39

That's not entirely fair to Osgood since he got most of the starts in St. Louis and New York and his record suffered, but if you adjust his teammates' numbers seasonally based on Osgood's decisions you still get a .612 winning percentage with Osgood and a .631 winning percentage without. Great goalies should be helping their team win, and by helping their team win I don't mean sitting on the bench watching a better goalie play.

2. Over the last three years, while Osgood has been hailed by many as a so-called great clutch goalie, he has an .897 save percentage on 2,708 shots.

I find it hard to believe he's not trying. If he's not trying he might be costing his employer millions of dollars, since if the playoffs started today the Red Wings would be left at home. If that happened then a big reason for it would be that they are below .500 in games started by Chris Osgood.

I also find it hard to believe he's unusually clutch, since his record didn't show anything in that regard prior to 2008.

In short, take your pick. Osgood's 2008 and 2009 playoff runs were because of:

A) An average goalie developed a crazy clutch ability late in his career and doesn't try at all in the regular season.

B) A goalie got hot for a short period of time on a strong defensive team.

Option B matches not only my interpretation of the numbers, but my subjective perception of watching the games. I'll go with what the larger sample size says almost every single time, whether it's Osgood or anybody else.

Anonymous said...

[i]Here are two simple reasons why I don't give Osgood much respect.[/i]

Say what you will, he got Detroit one cup and almost a second. In 2008 I will give you that he was on a good team and was not tested all that much, but lots of goalies fail even then, look at Nabokov and Lalime. In 2009 though, the Wings were not as good and the team had to lean on Oz much more.

[i]Great goalies should be helping their team win, and by helping their team win I don't mean sitting on the bench watching a better goalie play.[/i]

Need I remind you, Dominik Hasek had a meltdown against the Nashville Predators in the openers of 2008, and would have cost Detroit the playoffs, if Oz did not step in?

[i]if the playoffs started today the Red Wings would be left at home.[/i]

There are many reasons for that, not the least of which were long-term injuries to many of their star players this year. Any team no matter how good, will suck with many injuries to star players.

[i]B) A goalie got hot for a short period of time on a strong defensive team.[/i]

That, is simply not consistent with the facts. Detroit in 2009 was not nearly as good as Detroit in 2008 (you're the numbers man, look it up) in the playoffs and was dominated many times over the playoffs by Columbus and Anaheim, which played better than expected, by the Hawks, and especially Pittsburgh which I think you and all other observers would agree, decisively outplayed Detroit, especially with Datsyuk and Hos not up to full health. With just okay goaltending, Detroit would not have gone to game seven of the finals. They would have lost in one of the first three rounds in 09.

The Contrarian Goaltender said...

I think you misunderstand what I mean when I say a goalie got hot. That doesn't mean he wasn't playing well. It means that for a short period of time he played better than his actual skill level.

Every goalie in the NHL is capable of playing great games, or they wouldn't be there in the first place. Every decent goalie in the NHL is capable of playing great for a few weeks in a row.

But playing great for a few weeks does not mean a goalie is great. It just means he's on a hot streak. The difference between great and mediocre is only revealed over the long term.

For Osgood I don't even see the slightest bit of controversy in saying he got hot over the last two playoff seasons. Ozzy is not a .930 goalie, we have 16 NHL seasons' worth of data that make that abundantly clear.

It's not just Osgood that I've made this argument for. I've said the same things about guys like Marc-Andre Fleury and Cam Ward. They played well at the most important time, but that does not make them great goalies, any more than Fernando Pisani or Maxime Talbot are great players because they did well in the playoffs.

I agree that Osgood was pretty good in the 2009 playoffs. But to me that just doesn't mean very much in terms of how good of a goalie he is. Using his '09 playoffs to define Osgood is simply a mistake.

Anonymous said...

[i]But playing great for a few weeks does not mean a goalie is great. It just means he's on a hot streak. The difference between great and mediocre is only revealed over the long term.[/i]

So then would you be prepared to say that Giguere, Kipprusoff, Roloson, and Ward were on "hot streaks" during their recent cup runs also? With the possible exception of Roloson I would call all the above average goalies, and that's being generous!

For my money top five goal-tenders of the decade are Martin Brodeur (I know you'll like that one a lot ;-)), Osgood, Luongo, Kahbibulin, and maybe, perhaps the forementioned Roloson. What do you say to that?

The Contrarian Goaltender said...

So then would you be prepared to say that Giguere, Kipprusoff, Roloson, and Ward were on "hot streaks" during their recent cup runs also? With the possible exception of Roloson I would call all the above average goalies, and that's being generous!

First off, Giguere and Kiprusoff average goalies? What?

Secondly, of course they were on hot streaks. Other than the 2003 playoffs J.S. Giguere has played 477 regular season games at .913 and 31 playoff games at .907. That shows that his .945 in 2003 was a hot streak. Again, Giguere was playing superbly during that stretch, but it does not reflect his actual level of ability. Cam Ward was similarly playing way above his head.

On the other hand, Kiprusoff's save percentage actually dropped in the 2004 playoffs compared to the 2003-04 regular season. He was simply great that entire year.

As far as the top 5 goalies of the decade, of course Brodeur would make my list. He's been among the top goalies, my argument is simply that he is not far and away the leader like everyone says he is. I'd lean towards guys like Hasek, Vokoun and Giguere over Khabibulin, Roloson and Osgood, because I don't weight playoff performance as heavily as you do.

Anonymous said...

[i]First off, Giguere and Kiprusoff average goalies? What?[/i]

Kipprusoff almost won the cup once, he never actually won it. Then, his performance decreased for five straight years! Prior to this year, he was one of the bottom third goal-tenders in the league!

Giguere won the cup, and almost won it once before, but both those times, he was on really strong Mighty Ducks teams with Selanne, Macdonald, Kariya, Niedermayer, Getzlaf, Ruchin, and many other current, future, or past NHL stars.

You yourself admitted, that outside those 2 playoff runs, Giguere's statistics in the regular season, and other playoffs, were below average.

[i]I'd lean towards guys like Hasek, Vokoun and Giguere over Khabibulin, Roloson and Osgood, because I don't weight playoff performance as heavily as you do.[/i]

Hasek was very good, but his best years were in the 90s. He won the cup, and almost won it once before, but look at the excellent Sabers and Red Wings teams he got on, with hall of famer's like Brett Hull, Lindstrom, Satan, Shannahan, Peca, Izerman, and many others.

Vokoun I'm not that familiar with. He has some good numbers on Florida, but I won't call him one of the five best of the 2000's with no cup or no good runs in the playoffs.

Giguere, I already explained my opinion of.

Osgood, and Roloson, you are selling short. We already discussed Oz, but what about Rolli? Roloson has been on poor teams every year of his career, except the 06 Oilers. Go look up his save percentage for 2004, on Minny, he got .933, he should have shared the Vezina with Martin Brodeur that year, and he for sure would have won Edmonton a cup, if Ray Whitney didn't shove his knee into the pipe, in game one of the 06 finals.