Wednesday, January 31, 2007

Defining the Devils

One of the posters on HFBoards wrote that the defining sentence for the New Jersey Devils over the last 10 years was “shotsaveBrodeur!”.

Given that New Jersey spent maybe 1% of the time watching Brodeur making saves, and the rest of the time trapping and stifling the opposition, a more appropriate definition would have been, “turnedoverintheneutralzone!”

Tuesday, January 30, 2007

Goalie Quiz

Which of these two goalies is having the better season?

Goalie A: 30-14, 1.99 GAA, .929 save %, 9 shutouts
Goalie B: 15-15, 2.76 GAA, .903 save %, 0 shutouts

The answer: After adjusting for the team defence in front of them, the level of their performances is almost identical, as measured by shot-quality neutral save percentage.

The first is Brodeur, the second is Khabibulin. Marty faces the easiest shots of any goalie in the league. The Bulin Wall faces the hardest.

An average starting goalie facing Brodeur's shots could be expected to have a save percentage around .925. The exact same goalie facing Khabibulin's shots would be expected to have a save percentage of about .898. Even though the goalie is exactly the same, he would be over 33% more likely to give up a goal on any given shot playing for Chicago than New Jersey.

This, once again, shows why goaltending performance needs to be viewed in context. For some statistics (wins, GAA, shutouts), the team has probably more of an impact on the stats than the goalie himself. Even save percentage is very team dependent.

Goalies playing for bottom of the league teams can and do have outstanding seasons, even if they don’t make the playoffs. This year, Khabibulin, Rick DiPietro, Manny Legace, and Olaf Kolzig are all having excellent seasons, even though if the season ended today they would all be at home be watching the playoffs on TV.

Saturday, January 27, 2007

The Value of a Win

Here are two somewhat opposing viewpoints on the value of a win.

"I take a lot of pride in winning hockey games," Martin Brodeur said. "Stats are good, but wins really show the success you have. When I look at my years, I always look at wins." (From Hockeygoalies.org)

“All I gotta do is let one more in and I’ve got the win.” Marty Turco, 2007 NHL All-Star Game. (Watch the YouTube clip here).

The Playoffs and Reputation

Exhibit A in how team defence and playoff success define goaltenders:

Dan Cloutier vs. Martin Brodeur

2002-03:
Dan Cloutier: .908 save %, .908 shot-quality neutral save %
Marty Brodeur: .914 save %, .906 shot-quality neutral save %

2003-04:
Dan Cloutier: .914 save %, .910 shot-quality neutral save %
Marty Brodeur: .917 save %, .909 shot-quality neutral save %

(Stats courtesy hockeyanalytics).

Their performances over 2 NHL regular seasons were essentially the same.

The difference:

This goal was forgotten, as the Devils won the Cup.

This goal made Cloutier’s reputation as a choker, and he was run out of town in disgrace.

Oh, and another major difference?

Vezina Trophies: Brodeur 2, Cloutier 0.

Martin Brodeur – Vezina Trophy Winner

Martin Brodeur won his first Vezina trophy in 2003. There was generally described in the media as well-deserved and “long overdue”. In 2004, he won again for the second straight year.

Alan Ryder of Hockey Analytics did an interesting and in-depth study on goaltending performance from those two seasons. His conclusion: "He [Brodeur] still plays behind the league's best defense and he still doesn't deserve the Vezina."

In 2002-03, Florida gave up the 2nd most shots in the league, and according to Ryder also allowed the most dangerous scoring chances of any team. Luongo stopped 91.8% of the shots he faced to rank among the best goalies in the league.

In startling contrast, New Jersey gave up the fewest shots and least dangerous chances. Despite facing fewer and easier shots, Brodeur's save percentage (91.4%) was worse than Luongo's. After being adjusted for difficulty, his performance was ranked by Ryder as essentially league-average. Nevertheless, this average performance was good enough for recognition as the most outstanding goaltender of 2002-03.

In 2003-04, Roberto Luongo had one of the best goaltending seasons ever, despite Florida being the worst team in the NHL by far in shots against. Ryder measures his performance as being worth 35 points in the standings to Florida, which nevertheless was so bad that the team missed the playoffs. Luongo’s save percentage, after being adjusted for shot difficulty, was .931.

New Jersey allowed the fewest shots and the least dangerous scoring chances in the league, despite their slightly above league average goaltending. Brodeur’s shot-quality neutral save percentage was .909, which failed to crack the top 15 in the NHL.

Luongo was blamed for his defence’s shortcomings (and Florida's lack of offence), while Brodeur got all the credit for his team’s excellence. The Vezina, yet again, went to Martin Brodeur.

Friday, January 26, 2007

A Special Opportunity

“You don’t need a reason like an NHL All-Star Game to tune in to see Marty Brodeur play. Any chance you have you see this guy in action – especially in person – take it.” Good call, FoxSports.

First of all, the All-Star Game is not real hockey, so it is difficult to say anything very serious about the glorified, contact-less shinny that they play. One thing it does show is how very dependent goalies are on their defence. The best goalies in the league get killed repeatedly in these games, simply because nobody is backchecking or trying in the defensive zone. And nobody got lit up on Wednesday night worse than Marty Brodeur.

I did take advantage of the very special chance to watch Brodeur in action, and observed him getting scored on 6 times on 16 shots in one period of play. This was especially amusing considering the stream of accolades that the announcers were pouring down upon him even as the Western shooters were running up the score. It was of course the All-Star Game, and a few of those goals were pretty much unstoppable, but a few of them were downright ugly, including the Yanic Perreault deflection between Brodeur’s legs and a pair of long-range Rolston slapshots, one of which knocked Brodeur’s broadcasting microphone off on its way past him. Brodeur's excuse for his performance: He was just trying to avoid injury.

Brodeur did make made one impressive looking save. He stacked the pads on Cheechoo and made a glove save. A butterfly goalie like Luongo or Kiprusoff would have come across in the butterfly slide and made a routine save off of the chest or shoulder, but not Marty. Brodeur’s unique, highlight-reel friendly style is another reason why many people who never watch the Devils play consider him to be the best in the league.

But it's just the All-Star Game, and it doesn't mean a whole lot. After all, Patrick Roy holds the record for the most goals allowed in his All-Star Game career. Although based on this year's results, it looks like maybe Brodeur will soon be taking that record away from him as well.

If you haven’t seen them, you can watch all the goals (and Brodeur’s “big save”) on NHL.com.

Brett Hull on Martin Brodeur

Brett Hull on Martin Brodeur (NBC): “I guarantee he’s averaged less than 20 shots (against) a game.”

Yet again, another Brodeur basher who doesn’t know his numbers. Brodeur has actually faced an average of 25 shots against per game in his career. So there, Golden Brett!

And even though that is less than pretty much all of his contemporaries, and nearly 20% below NHL average, I’m sure Brodeur must just be making shooters miss through the power of his amazing reputation.

And even though several statistical studies have indicated that low total of shots also consisted of less dangerous scoring chances than average, it doesn't matter because Brodeur knows how to make the big, important saves because he is a such a “clutch” player.

And even though Brodeur is mathematically nearly twice as likely to get a shutout facing 25 shots than facing 31 shots (like say, Curtis Joseph’s career average), all his shutouts must show that Brodeur can just win games singlehandedly.

OK, so Hull's numbers were a bit off, but his overall point stands – New Jersey is certainly a very comfortable place to play goal.

Van Schwabenssen for Vezina?

One way to try to evaluate the effect of New Jersey’s defensive team play is to look at Brodeur’s backups and see how they performed. So I looked at the Devils’ backup goalies over the last 5 seasons (including this one). Here is the breakdown:

M. Brodeur: 340 games, 2.17 GAA, .916 save%, 37 shutouts, 0.11 shutouts/game

Backups: 44 games, 2.17 GAA, .913 save%, 4 shutouts, 0.09 shutouts/game

Remember, these are the 5 years in which Brodeur won 2 Vezinas (probably 3 after this year), a Stanley Cup and an Olympic gold medal, and was virtually unanimously acclaimed as the best goalie in the world. Yet his backups had an identical GAA, nearly the same save percentage, and nearly the same shutout rate, playing behind the same defence.

Seems to me that a two-pronged attack of Corey Schwab and Scott Clemmensen, with maybe a little John Vanbiesbrouck sprinkled in, could more or less duplicate Brodeur’s “Hall-of-Fame” results over the last 5 years.

Corey Schwab is a particularly interesting case. Here is the breakdown of his fairly uninspiring NHL career, Devils and non-Devils:

Devils: 24 games, .929 save%, 1.54 GAA, 2 shutouts, 7-6-2 record

Non-Devils: 123 games, .896 save%, 3.13 GAA, 4 shutouts, 35-57-11 record

I’d say the neutral zone trap works, what about you?

One final observation to keep in mind when Martin Brodeur ends up breaking Terry Sawchuk’s shutout record. If a composite of all of Brodeur’s backups (say Chris Van Schwabenssen), had played the same number of minutes that Brodeur has behind the same defence, and recorded shutouts at the same rate they actually did, they would have averaged about 5 shutouts per season and 55 in total. That would place them 17th on the all-time list, tied with Clint Benedict and one ahead of Bernie Parent.

If a bunch of career backups could have plausibly thrown up over 50 shutouts in a decade behind the New Jersey defence, then I think even the most hard-core supporters should at least reconsider Brodeur’s standing in the annals of the game.

Do Good Goalies Win, or are Winning Goalies Good?

In every sport, there are average or decent players who are considered to be stars, possibly due to media hype or the hidden efforts of their outstanding teammates. Similarly, there are very athletically gifted players that despite their physical prowess deliver sub-par production, often because of poor work ethics or poor fundamentals. It is very difficult to accurately judge both types of players, since for both of them the highlight reels and the win column tend to be very powerful blinders to the truth. This is especially true for the most important positions in team sports, such as football quarterback, baseball pitcher, or hockey goaltender.

This suggestion may be controversial to some, but ask them if their boss is deserving of his position, or about that co-worker who got the promotion ahead of them, or the owner's nephew over in marketing. Who among us cannot think of a regular guy who is in a high position of responsibility, or who is getting recognition we feel is unjustified? There is no reason why the same type of thing can't happen in sports. Sports are not a complete meritocracy, and talent valuation ineffiencies do exist. Just read Moneyball.

This is especially true for hockey goaltenders. There are only 60 jobs open in the world, so there can be many reasons why a talented goalie does not make it while a less talented counterpart becomes famous. Some examples are team strength, team depth at the goaltending position, developmental abilities, injuries, and the ability of the coach to evaluate goaltenders. Just think about Dominik Hasek – despite being the best goaltender in Europe at a young age, he spent years backing up Ed Belfour and having his non-traditional style being dismissed by all the top goalie talent evaluators in North America, before finally getting a starting job and putting together some of the best individual seasons in the history of the NHL.

Goalies are only as good as their team, in the vast majority of circumstances. Only if a goalie is substantially deviant from the norm does their performance become a major input in the team’s success or failure. A good team can win with an average goalie. The most recent example is last year’s Carolina team, and the Red Wings, Senators and Flyers have also been consistently good despite often having mediocre goaltending. There is of course also the New Jersey situation which has given title to this blog.

The problem is that it is hard to see average goalies for what they are. Good goalies are defined as goalies that win, therefore goalies that win are considered good, and it becomes circular logic that argues for the basic premise that you need good goaltending to win, when, in fact, winning largely creates the perception of good goaltending. This is especially true when you evaluate goalies based on playoff performance alone.

I certainly have no anti-goalie bias in this debate, I am a goalie myself. I won a tournament once when I was playing for the best team and pretty much just had to stand there and watch. I also once lost in a tournament where I was maybe the best goalie there, and even shut out one of the other top teams, but my team was bounced in the first round. In the short run, goalies can and do steal games. In the long run, it’s the team that wins or loses, and the goalie is in most cases just along for the ride.