Monday, January 14, 2008

Martin Brodeur and the Upside Down Career Curve

When NHL play resumed after the 2005 lockout, Martin Brodeur re-entered the fray as the two-time defending Vezina Trophy winner. However, he longer had Scott Stevens and Scott Niedermayer playing in front of him, and the new rules were supposed to open up the game and circumvent the defensive tactics executed to perfection by his Devils teammates. Many saw this to be a test of how good he actually was.

For the first three months, Brodeur failed the test. On New Year's Eve, 2005, Brodeur was lit up for 6 goals on 22 shots by the lowly Toronto Maple Leafs. At that point in the season, his stat line was a mediocre 14-14-3, 2.94, .896. Based on that small sample, it looked like those who were claiming he was a team creation had been right all along.

But whether it was a New Year's resolution, something his goalie coach spotted in his game, or perhaps some other reason, the Martin Brodeur that went between the pipes in 2006 was very different than the 2005 version.

On January 3, Brodeur made 16 saves to shut out Florida, beginning a 9 game personal win streak, 3 of them by shutout, where he posted a .934 save percentage. This was followed by a strong finish to the year, a Vezina Trophy winning season in 2006-07 where Brodeur had his highest finish in save percentage since 1996-97, and after a rocky start to the current campaign he seems to have regained his form over the last two months to appear back up on the leaderboards.

So what happened? Has Brodeur's post-lockout performance been the proof that his backers have been waiting for? Can they claim that he was unjustifiably denigrated for years because of his poor defence, based on the evidence from the last 2 1/2 seasons? This post is certainly not going to stop them from claiming that, of course, but the reality is that the Martin Brodeur of 2006-08 is not the same goalie as he was before the lockout. Since New Year's Day 2006, Brodeur's performance has improved by every goaltending measure that I am aware of.

The result is that Brodeur's career path has been the exact opposite of most goalies. Here it is, broken down into three sections:

Up to 1997-98: .915 season save %, .931 playoffs save %
1998-99 to Dec. 31/05: .909 season save %, .917 playoffs save %
Jan. 1/06 to today: .921 season save %, .919 playoffs save %

Ordinarily, goalies would be peaking somewhere during the age 26-31 range, but those were actually some of the worst seasons of Brodeur's career so far, even during some of the most successful seasons for the New Jersey Devils and when overall league scoring was very low. Brodeur has spent roughly half his career as one of the top 5-10 goalies in the league, and has spent the other half around league average, i.e. somewhere in the top 15-20, despite playing for the same consistently strong team. This is a peculiar dichotomy, and belies the common characterization of Brodeur as a "consistent goalie".

Are there any underlying factors at work here? Could it be something related to power play differential, shot quality difference, or something else related to the new NHL rules? Let's break it down further to look at even-strength and penalty kill save percentages. Overall save percentages are often skewed by how many penalties a team takes, and Brodeur's teams have nearly always been the best in the league at avoiding calls. Detailed special team save percentage data from is only available from the 2000-01 season, but that at least allows us to compare the "dead-puck" Devils to today's version and see what has changed.

2000-01: Brodeur .919, League .916 (Leader .932)
2001-02: Brodeur .917, League .917 (Leader .934)
2002-03: Brodeur .921, League .920 (Leader .940)
2003-04: Brodeur .924, League .924 (Leader .941)

Penalty Kill:
2000-01: Brodeur .839, League .863 (Leader .902)
2001-02: Brodeur .849, League .863 (Leader .922)
2002-03: Brodeur .866, League .868 (Leader .913)
2003-04: Brodeur .878, League .868 (Leader .916)

What do we see here? We see basically a league average goalie. This despite New Jersey allowing the easiest shot quality against in the league, as calculated by Alan Ryder at Hockey Analytics. Brodeur won two Vezinas in this period, but they were certainly undeserving choices.

Moving on to the post-lockout years:

2005-06: Brodeur .922, League .916 (Leader .941)
2006-07: Brodeur .927, League .916 (Leader .933)
2007-08: Brodeur .934, League .925 (Leader .948)

Penalty Kill:
2005-06: Brodeur .869, League .863 (Leader .899)
2006-07: Brodeur .904, League .863 (Leader .921)
2007-08: Brodeur .905, League .885 (Leader .917)

Here we see significant improvement, as Brodeur outperformed the league average significantly in both categories in 2006-07, as well as so far in 2007-08. Although he has never been the league leader in any of these categories, he is certainly much closer now than he was in the early '00s. The most drastic improvement is on the penalty kill.

Here is a graph that shows the above numbers visually.

Brodeur still faces a relatively low shot total against, even post-lockout. This year he is facing 26.4 shots per 60 minutes of play, which is just two more shots per game than he did in 2003-04 when the Devils were at their trapping best. Save percentage is independent of shots faced, so that does not seem to be a factor in his improvement.

Additionally, shot quality results from Hockey Analytics and Hockey Numbers indicate that New Jersey was among the league's best in terms of shot quality against in both 2005-06 and 2006-07. It is only this season that they have fallen off. They weren't as smothering as they were at their peak under Pat Burns in 2002-04, but the system was still functioning very effectively. Many have pointed out the inexperienced defence, but recent work by hockey bloggers indicates that it is the forward group that has a larger impact on the number and quality of shots for or against, and New Jersey has some very strong two-way forwards, led by Patrik Elias and Brian Gionta.

By bucking the normal career trend by improving as he ages, as well as having multiple seasons of average play camouflaged behind an excellent defence, Brodeur is leading many to conclude that we were missing something all along, and that he was the main reason for all the wins and shutouts, not the defence in front of him. That conclusion is wrong.

Even while facing similar numbers and quality of shots as before, Brodeur is simply stopping more of them. He is better now than he was with Stevens and Niedermayer. Nevertheless, his performance over the last two seasons does not change the fact that for 6 1/2 seasons he was very average behind one of the best defensive teams of all-time. It certainly does help his career standing, and if he continues his late career resurgence for several more years he could clearly separate himself from the crowded pack of solid long-time starters from the '90's and '00's (Belfour, Joseph, etc.). Brodeur is doing an excellent late career sales job, but he is going to have to sustain, if not even increase, his current pace for a number of seasons before he can be legitimately considered even one of the top 10 goalies of all-time.