Defending Brodeur

Eric T.
May 30 2012 04:31PM

 

 

With Martin Brodeur headed to the Stanley Cup Finals yet again, the discussion about how to evaluate his career has been renewed recently.

Those on the statistical analysis end of the spectrum tend to point to his good-but-not-great save percentages and argue that Brodeur was never a dominant performer, but his longevity is remarkable.

The Case Against

The argument goes something like this: over Dominik Hasek's career, his average save percentage was 0.020 higher than the league average. Patrick Roy beat the league average by 0.015. Brodeur beat it by only 0.007, which makes him very good for a long time but not great -- he has never led the league in save percentage or even finished second, and he's been in the top 10 only six of 19 seasons.

Moreover, he did that while facing an extremely easy distribution of shots. This isn't the standard anecdote about shots being forced to the outside, which is often claimed and rarely appears to be true when recorded shot locations are analyzed; it is an observation about how little time the Devils spent on the penalty kill throughout his career.

There have been 77 goalies with at least 200 appearances since the NHL started separately tracking penalty kill shots faced in 1997-98, and only Pekka Rinne and Byron Dafoe faced a lower percentage of their shots on the PK than Brodeur, which surely inflates his save percentage a bit. At even strength, he has beaten the league average save percentage by only 0.005 over his career, which is fine, but not outstanding.

That's the argument against Brodeur. But the headline here says I'm going to defend him. Why? Because I think the arguments on the other side have mostly been terrible, and I think there's a much better case to be made.

The Case in Favor

I see a lot of people saying that what made Brodeur so outstanding was his consistency, that the team knew he was reliable. This is the same argument that was being made to defend Marc-Andre Fleury a month or two ago, and I've already shown that it's nonsense -- no goalie appears to be any streakier than any other goalie, no matter what hot or cold stretch you might happen to remember. Yes, Brodeur was very good in the 1999-00 playoffs, but how reliable was he in 1998-99?

(On a side note, I don't understand why people think consistency is a desirable quality in a goalie anyway. Brodeur's career save percentage is .913; in his three Cup years his playoff performance was .927, .927, and .934 -- I would argue that a team would win more Cups with a goalie who is sometimes great for 25 games and sometimes terrible than with a goalie who is always the same.)

Another common argument is that Brodeur's clutch play is what makes him great. People will point to his Cup wins and maybe link to a single great save on Youtube and say that proves the point. Well, the excellent blog Brodeur is a Fraud (perhaps you can guess which side of the argument he falls on) broke down playoff results by score and period and found that Brodeur's posted a .919 save percentage in the first two periods of the game, .913 when tied in the third period, and .916 in overtime. Other goalies' save percentages went up over that span (perhaps because fewer penalties are called, or perhaps because of skater fatigue), but Brodeur's performance declined in those most critical clutch moments.

A third argument relies on his puck-handling skills, and I agree that this is a factor. It is probably not as large as most people think, but I found that in recent years he has faced about one shot per game less than his backups, which might make his puck-handling worth as much as 0.005 on his save percentage. I have since found that Brodeur is a Fraud performed an identical analysis over the previous years, and also found a shot-per-game difference.

So let's start with that as the argument in Brodeur's favor: his puck control suppresses shots to such an extent that his save percentage underrates his performance by perhaps as much as 0.005. If we just add that on as a correction factor, we now have an adjusted performance of ~0.010 to 0.012 above league average for his career, which is still not quite where Hasek and Roy are, but it is getting closer.


Photo by slgckgc via Wikimedia Commons

The second argument that I consider convincing relates to home scorer effects. It has been widely observed that some scorers tend to count more saved shots than others, but the corrections are usually small and so this effect often gets overlooked (at least by me). Triumph of the excellent blog Driving Play recently brought this up specifically in response to a discussion of Brodeur's save percentage.

The average save percentage at home this year was .9082, whereas on the road it was .9045. So any combination of scorer bias and home ice advantage would seem to give the home goalie a 0.0037 edge over his road performance. For Brodeur's career, he has had a home save percentage 0.0065 lower than his road save percentage -- 0.9092 versus 0.9157. This doesn't appear to be the result of playing worse at home -- his GAA at home was 2.17 and on the road it was 2.28.

Assuming the official scorers counted the goals correctly, it would seem that he had a typical home-ice advantage, but was getting substantially fewer of his saved shots counted than other goalies get at their home rink. With a typical scorer, his home save percentage would probably have been about 0.010 better than it actually was (turning a 0.0065 deficit into a 0.0037 boost at home).

Adjusting his home save percentage up by 0.010 to account for the apparent scorer issues would boost his overall save percentage by 0.005. Couple that with the puck-handling adjustment and now we're getting his adjusted performance up into the range of ~0.015 to 0.017 above league average. At that point, he's right in the conversation with Roy and Hasek.

It may be that I've overestimated the adjustments here (though I think these numbers are defensible), but the point is this: it's true that Brodeur's talent isn't fully captured by save percentage, but "I remember him making some huge saves" isn't what's missing.

Recently by Eric T.

 

2654ef2681c88bc3252431ec45e30590
Eric T. writes for NHL Numbers and Broad Street Hockey. His work generally focuses on analytical investigations and covers all phases of the game. You can find him on Twitter as @BSH_EricT.
Avatar
#1 Derek Zona
May 30 2012, 06:41PM
Trash it!
0
trashes
Props
1
props

This is damned good work, Eric. The known scorer bias strikes again.

Avatar
#3 Geoff Detweiler
May 30 2012, 04:55PM
Trash it!
0
trashes
Props
0
props

Brodeur's clutch playoff performance, including Game 7 meltdown against Carolina in 2009: http://www.youtube.com/watch?v=0Aiin4D31xY

But seriously: Great article. The home scorer bias is a damning piece of evidence.

Avatar
#4 PopsTwitTar
May 30 2012, 08:12PM
Trash it!
1
trashes
Props
0
props

I may be missing something, but are you only adjusting Brodeur's #s for scorer bias? In other words, wouldn't Roy and Hasek also get some minor bump in their own numbers?

Avatar
#5 PopsTwitTar
May 30 2012, 09:05PM
Trash it!
0
trashes
Props
0
props

@Eric T.

I wonder if anyone has looked at that more deeply. I know Gabe has his few MSG-focused posts. Maybe someone can come up with a "per rink adjustment factor"

Avatar
#7 Cam Charron
May 30 2012, 10:44PM
Trash it!
0
trashes
Props
0
props

I noticed a similar effect when Triumph pointed it out in the comments. My numbers differ slightly from yours but show the same picture. MB30 has a lower GAA and save percentage at home.

Avatar
#8 Ben Wendorf
May 30 2012, 11:01PM
Trash it!
0
trashes
Props
0
props

I have looked at the numbers a bit, too, and found New Jersey and especially Minnesota (aka Niklas Backstrom's Wild teams) to be chronic undercounters for shots-against. Here's the link: http://www.arcticicehockey.com/2011/6/29/2228337/shot-counting-in-pictures

It's probably the one thing that sticks in the back of my mind when I think of Brodeur. Now, should Henrik Lundqvist continue his trajectory and break all those counting records, maybe people will stop talking about it. On to Osgood...

Avatar
#9 Cam Charron
May 31 2012, 08:34AM
Trash it!
0
trashes
Props
0
props
Ben Wendorf wrote:

I have looked at the numbers a bit, too, and found New Jersey and especially Minnesota (aka Niklas Backstrom's Wild teams) to be chronic undercounters for shots-against. Here's the link: http://www.arcticicehockey.com/2011/6/29/2228337/shot-counting-in-pictures

It's probably the one thing that sticks in the back of my mind when I think of Brodeur. Now, should Henrik Lundqvist continue his trajectory and break all those counting records, maybe people will stop talking about it. On to Osgood...

Very good stuff there Ben ^^^.

Comments are closed for this article.