Monday, May 28, 2007

Brodeur's Save Percentage

The main evidence that Martin Brodeur is a team creation is his save percentage record. Save percentage measures how good a goalie is at stopping the puck. Studies have shown that save percentage is not affected by how many or how few shots a goalie faces. The only hidden factor affecting save percentage is the quality of shots faced. As I have demonstrated repeatedly, shot quality measures have all overwhelming shown that Brodeur is hardly disadvantaged in that area.

Let's look at his record in terms of league ranking. I only included goalies who played in at least half the games in that season, i.e. starting goalies only.

Brodeur's Save Percentage Ranks:

1994: 4th
1995: 17th
1996: 8th
1997: 2nd
1998: 5th
1999: 17th
2000: 12th
2001: 16th
2002: 19th
2003: 13th
2004: 10th
2006: 10th
2007: 2nd

Not only has he never led, he has only been in the top 5 among starting goalies four times, and in the top 10 7 times in 13 full seasons. Nearly half the time, Brodeur has finished outside of the top 10 in save percentage, including 4 times when he ranked 16th or lower.

Looking at those numbers, you wouldn't immediately guess that he has led the league in wins in 8 of the last 9 seasons. How did that happen? Do the wins overrule save percentage and shown it to be flawed? Of course not. Brodeur plays more games than anyone else, so he gets more opportunities to win. His team has also played a consistently successful style that contributes greatly to his performance.

Brodeur is a solid goalie, and at times, such as through much of this year, he has been great. But his career as a whole is clearly not elite. These stats also seem to point to the conclusion that 2007 was a bit of a fluke. It will be interesting to see whether he can repeat this year's success or whether he will revert to the 10th-15th ranked performance he delivered for the 6 seasons prior.

Tuesday, May 22, 2007

Hasek vs. Brodeur in the Playoffs

I know I've already argued this in considerable depth here, but watching Dominik Hasek make another deep playoff run makes me again wonder why Martin Brodeur is almost universally considered a better playoff goalie than the Dominator.

Hasek has a better career playoff save percentage and a better career playoff winning percentage than Brodeur. Only once in his playoff career has his team lost to an opponent with fewer regular season points (although the Ducks could soon make it twice). Brodeur has lost 5 times to weaker opponents, and he has been outplayed by average opposing goalies like Emery, Esche, Irbe, Ward, and Rhodes. Brodeur has been on three Cup-winning teams, granted, but in 2001 he was the main reason that New Jersey lost the Cup, getting totally outplayed in the Finals by Roy. Brodeur has never done anything like what Hasek did in 1999 in terms of being a difference maker on a mediocre team.

And those Cup wins? Cup wins are partly just a numbers game. Be on enough good teams, and eventually you'll win. From the time that Brodeur entered the league (the same time that Hasek became an NHL starting goalie) until the introduction of the shootout messed up the standings there were 52 100-point teams in 11 seasons. Nine of those teams won the Stanley Cup, so 100 point teams had about a 1 in 6 chance of winning it all. The odds were much worse for teams with less than 100 points: of all the teams that finished with less than 100 points in that time period, just 2 of them won the Stanley Cup (and one of them was New Jersey in the lockout season in 1995).

Brodeur has played on 10 100-point teams, so he would have been expected to have won about 2 Cups by now. He's won 3. This is the second 100-point team that Hasek has played on in his career, so based on that he shouldn't really have been expected to win anything. He won a Cup in 2002.

The number of rings doesn't matter - how well the goalie played does. Both Brodeur and Hasek have more rings than they would have been expected to win, given their respective teammates. Despite those weaker teams, Hasek also has the better rate stats (winning percentage, save percentage, etc.). The only conclusion, therefore, must be that Hasek's playoff record is better than Brodeur's.

Sunday, May 20, 2007

Patrick Roy and the Luck of the Draw

Between 1980 and 1993, the Montreal Canadiens were a consistently excellent team, winning 7 division titles. In that span, they were one of the top 6 teams in the league in points every season except for 1984, 1986, and 1993. What happened in those three seasons? In 1984, the Habs went to the Conference Finals. In 1986 and 1993, they won the Stanley Cup. The division winners weren't so lucky - six of the seven teams were eliminated in the the first or second playoff round.

This shows how the playoffs are often a crapshoot. The three worst Montreal teams in a 14 year period combined for 2 Stanley Cups and a Conference Final appearance. The seven division champion teams only once made it past the second round. How did the weak teams manage to succeed where the stronger ones failed?

One of the key reasons is playoff matchups. Obviously, teams playing stronger opponents are more likely to lose. For example, in the six years between their Cup wins, Montreal lost four times to 100 point teams, all of whom were better than they were.

In 1986 and 1993, they quite simply avoided the top teams. The best team Montreal played in the 1986 playoffs was Calgary, an 89 point team during the regular season. They beat 86 point Boston, 78 point New York Rangers, and 84 point Hartford. Sure, the Habs were only a 40-33-7 team, but they were as good as anybody they played in the playoffs.

In 1993, it was a similar story. The Habs were a solid team that put together a 48-30-6 record throughout the season, good for third place in the division. After a tough first round series against 104 point Quebec, where Montreal won three games by just a single goal, 2 of them in overtime, the Habs beat Buffalo (86 pts), the Islanders (87 pts), and Los Angeles (88 pts), all of which were clearly inferior teams.

The 1986 and 1993 Montreal Canadiens had tough, experienced lineups and all-star goaltending from Patrick Roy. But probably the most important reason that both teams won the Cup was because they had very favourable matchups. None of the 8 teams they played had more regular seasons wins than Montreal did, and none of them were elite teams, or even division winners.

Even on the Avalanche, Roy benefitted from a good "luck of the draw" in the playoffs, although that was partly because his own team was so good that they usually played lower seeds. In his four Cup runs combined, Roy only had to face three division winning teams, and the only one of them that won more games during the season than his own did was the 1996 Detroit Red Wings.

Patrick Roy was a great goaltender, but two of the main reasons he won 4 Stanley Cups were the quality of the teams he played on, and great luck in avoiding the best teams in those playoff runs. Winning with two of the weaker Hab squads has also been instrumental in creating Roy's reputation as a clutch, game-stealing goalie that wins with less than impressive teammates. However, that's not really true - the Montreal Canadiens of the late 1980s and early 1990s were excellent teams, they just happened to underachieve during the regular seasons of the years they won the Cup. Many have the perception that Roy did not play on great teams during his time in Montreal - nothing could be farther from the truth. He did not win the Cup with the best teams there (although he came close in 1989) but the Habs were consistently excellent with Roy between the pipes, which certainly helped him in the win column as well with his performance statistics.

Friday, May 18, 2007

Why Luongo Never Made The Playoffs in Florida

In 2001-02, Jose Theodore had a career year, posting a .931 save percentage in 67 games played. The Montreal Canadiens were a weak team, ranking 19th in the league in scoring (2.52 goals per game) and 27th in shots allowed (31.7). Montreal made it to the playoffs anyway, edging out Washington by 2 points for the 8th seed, mostly because of Theodore's outstanding play. At the end of the season, Theodore took home the Vezina Trophy winner as the league's top goalie and also won the Hart Trophy for league MVP. His season was an example of how great goaltending can make a big difference for a mediocre team.

In 2003-04, Roberto Luongo had a similarly fantastic year, with an identical .931 save percentage in 73 games played for the Florida Panthers. The Panthers ranked 23rd in scoring (2.29 goals per game), and 29th in shots allowed (34.0). They finished with 75 points, 16 short of the playoffs. Despite his excellent season, Luongo finished 3rd in Vezina Trophy voting, and 6th in MVP voting. Luongo played as well as Theodore did in more games and on an even weaker team. Yet he did not get the same recognition. In the view of many observers, if he was really that good he should have found a way to get his team into that 8th spot, as Theodore had done.

The problem is that it is doubtful that any goaltender, even Dominik Hasek at his absolute peak, would have been able to drag the Panthers teams of the early millennium into the playoffs. Theodore's Canadiens were bad, but Luongo's Panthers were horrible. From 2001 - 2004, the Panthers never scored more than 200 goals in a season, and in Luongo's last four seasons there they ranked 29th, 30th, 30th, and 29th in the league in shots allowed. With little offensive support and facing a barrage of shots, Luongo needed to be almost perfect for his team to win.

To make the playoffs, the Panthers would have needed around 90-95 points. That implies that they would have to score more goals than they allowed. The problem was that they scored so few that it was almost impossible for the goaltender to keep the goals against low enough. Between 2002 and 2004, the Panthers scored 544 goals, and allowed 8439 shots. To have an even goal differential, the Panthers needed their goalies to collectively save 93.6% of the shots, which is just below Dominik Hasek's all-time save percentage record of .937. Taking into account the minutes played by his backups, Roberto Luongo would have needed to play better than any goalie ever has in the history of the league, and even that would probably have left Florida with around 80-85 points, not enough for a playoff spot.

The only season in which Luongo realistically had even an outside chance to make the playoffs was last year. The Panthers had an improved offence, scoring 240 goals (21st in the league). They still were among the worst in the league in shots against, but if Luongo had repeated his dazzling play from 2003-04 the Panthers would have at least been in the playoff hunt. Unfortunately Luongo, maybe distracted by contract negotiations, had the worst year of his career (and still had a .914 save percentage and 35 wins on a poor team, which just shows the level of his talent).

So if Luongo never made it with the Panthers, why was he in the playoffs this season with Vancouver? The Canucks were almost as offensively inept as the recent Panthers teams, ranking 22nd in the league in goals for. However, they were above average in terms of shot prevention (and probably shot quality allowed), ranking 14th in the league. Facing fewer shots dropped Luongo's GAA by about half a goal from where it usually was in Florida, and despite Vancouver's punchless offence Luongo's play was enough to push the Canucks over the edge and into the postseason.

The worst teams in the league will miss the playoffs almost regardless of how well their goalies play. They don't score many goals and give up too many difficult shots, so that even outstanding goaltenders won't be able to save enough pucks to keep them in the game most nights. Roberto Luongo had virtually no chance to make the playoffs in Florida. Unfortunately there are many fans and observers that hold his lack of playoff participation against him (and there are some that still do). The absurd result of that is that many people have viewed this season as a breakout year for the man who may have been the best goalie in the NHL since 2002-03.

Thursday, May 17, 2007

Missed Shots

With the availability of NHL play-by-play records, many new statistics are being tracked. One such statistic is missed shots against a goaltender, which can be found at behindthenet.ca.

I have heard it argued that goalies who are solid positionally force shooters to miss the net more often. Anecdotally from my experience, this argument seems very reasonable, since goalies often can put themselves in blocking positions where they cover almost the entire net, making it very difficult for shooters to pick the corners. This, goes the argument, is one of the reasons why some goalies face fewer shots, such as Martin Brodeur. However, this position isn't supported by the facts.

According to Behind the Net, Brodeur ranked 45th in the league in the number of missed shots per 60 minutes at even-strength. He might have made a shooters miss a bit more often on the penalty kill, where he ranked 13th. But when his team was on the power play, he ranked 51st.

So Brodeur doesn't come off very well based on this metric. But are there some goalies who do routinely make shooters miss more shots?

Looking at some of the goalies who rated highly in terms of misses, there did not seem to be much correlation with talent. Among the goalies who made shooters miss often at both even-strength and on the penalty kill were Mathieu Garon, Dan Cloutier, Andrew Raycroft, and David Aebischer, four of the worst goalies in the league. At the other end, among goalies who rarely made shooters miss, could be found Nikolai Khabibulin, Chris Mason, Ilya Bryzgalov, and Tomas Vokoun.

The correlation between save percentage and miss rate was slightly negative, meaning that as save percentage went up, shooters were less likely to miss the net, not more likely, which goes against the conventional wisdom. One possible explanation for this is that good goalies are more likely to save shots that are going slightly wide or high, shots that then get recorded as saves whereas if the goalie had let them go they would be a missed shot on goal.

I then compared starting goalies to their backups to see if they forced more misses. Of all shots attempts at net that weren't blocked (i.e. saves, goals, and misses), 29.1% directed at the starting goalies ended up missing the net. The stat for backups was 28.9%, almost exactly the same. It was also similar on the penalty kill, 28.7% for starters, 27.7% for their backups.

I also looked at the correlation between starters and backups at even strength. It was .098, indicating almost no relationship at all. On the penalty kill, however, it was .268, indicating that team penalty kill tactics have a small influence on the rate of shots missed.

Missed shots therefore seem to be of little use in terms of evaluating goalies, at least based on this year's results. The missed shot rate does not seem to be correlated with goalie ability at all. If anything, shooters seem to miss the net less often against good goalies.

Wednesday, May 16, 2007

The Importance of Asking "What If?"

Alan Ryder wrote an article for Globe Sports contrasting the Luongo acquisition with the Raycroft acquisition.

He analyzed how both Toronto and Vancouver did this year, and concluded, quite reasonably, that Luongo had a major effect on Vancouver making the playoffs, whereas Raycroft gave the Leafs goaltending that was just as mediocre as what they had in 2005-06. He also threw in a jab at Leaf fans, by pointing out that if the Leafs had acquired Luongo instead of Raycroft, they would have been as much as 20 points better in the standings.

What was the response to this article in the comments? By and large, the commenters bashed the article, mostly because it dealt with "what-if" scenarios (which it did, briefly at the very end in a passing comment). Many Brodeur fans respond in a similar manner when people, such as myself, claim that Brodeur simply would not have close to as many wins or shutouts on a league average team.

That argument boils down to something like this: We don't know for sure what would have happened, so let's not even think about any alternate possibilities. Let's all just measure every goalie by wins and shutouts and blindly agree with everything the mediots say. Let's not use our brains to think about possible scenarios or try to separate out what was a team effect and what was a goalie effect. Cam Ward won a Cup, he's better than Rick DiPietro and Henrik Lundqvist. Roberto Luongo never made the playoffs in Florida, he's not worthy to carry Chris Osgood's jock. And Martin Brodeur is the greatest goalie in the history of the world.

Hypotheticals are critical to sports analysis. In scientific experiments, there are things called control variables, and scientists try to keep them constant to measure the effect of changes in the variables they are investigating. In sports, there are no control variables. You can't tell the teams to switch goalies and play the game over again. If you want to compare a player of today with someone who played 40 years ago, you have to compare different eras, different rules, different opponents, different equipment, different playing styles and techniques, and different numbers of teams and levels of talent dilution, among other things. Even in comparing, say, Luongo vs. Brodeur, they have different goal support, different shots against numbers, different quality of shots against, different teammates, different opponents, different team tactics, different special teams situations, different conferences, different travel schedules. Because of all these adjustments that have to be made, everything is really just a "what-if" consideration, or there cannot be any useful ranking or comparison at all other than just looking at the stat sheet and sorting everyone by points or wins.

You cannot properly evaluate anyone without asking "what if". Baseball stat guys use a stat called WARP, which measures how many extra wins a player contributes to the team, compared to if that player had been replaced by a minor leaguer or "replacement-level" player. That is a "what if" scenario (basically, what is the Yankees record if they replace A-Rod with a Triple-A third baseman?), but clearly based on statistical evaluation.

It is the same with goalies. If a goalie plays on the worst team in the league, his performance must be viewed in the context of how other goalies would have been expected to have performed on that same team. No, we will never know for sure how they would have performed, but we can adjust for their actual performance on their own teams, and then apply some correcting factors to fairly reasonably surmise where their expected performance level would have been. For example, we do not know 100% for sure that Martin Brodeur would never have made the playoffs if he played for Florida between 2001 and 2006, but based on his performance record, Florida's defensive and offensive performance, available shot quality information, and formulas for predicting team records given the number of goals scored and allowed, we can be about 99.99% sure that he would not have.

By not asking "what-if" and taking all goalie stats at face value, fans end up penalizing players for being on the wrong team, or conversely, crediting them for playing on a good team. That is poor analysis, and is why good goalies can be considered legends of the game mostly because of their team situations (Brodeur, Fuhr, Billy Smith, etc.).

Monday, May 14, 2007

Playoffs 2nd Round: The Better Goalies All Lose

In the second round of this playoffs, one could make the case that the better goalies all lost. Lundqvist lost to Miller, Luongo lost to Giguere, Brodeur lost to Emery, and Nabokov lost to Hasek. The last one is probably the only one that is really debatable; Nabokov was probably slightly better during the regular season, although Hasek has been quite good in these playoffs.

Ottawa won tonight to go up 3-0 over Buffalo, meaning that Ray Emery will likely play in the Stanley Cup Finals. He is this year's Cam Ward, an average goalie who simply plays for a team that is getting hot at the right time. Over in the West, we'll see if either Hasek or Giguere can be a difference maker, or if it will just come down to whichever team has the best 18 skaters in front of them.

So in the second round, as in the first one, goaltending was not a major factor in the majority of the series, regardless of what the commentators say. Teams do not necessarily need good goaltending to win; the combined inputs of scoring, defence, and goaltending determine the final result. Sometimes that means a team needs great goaltending, a la Vancouver, and sometimes it just needs a Ray Emery or a Cam Ward. The mix can vary, but there are many different ways to the final goal of being a Stanley Cup champion, and it will be interesting to see which team and goaltender put everything together in the best combined fashion to bring the trophy home.

If Brodeur Was a Forward

Comparing goalies is challenging, because of different eras, different teammates, different opponents, different equipment, different styles, and the difficulty of separating individual performance from team performance. This makes it tougher to rank goalies than skaters. Ask 10 different experts on who the top 5 goalies of all-time are, and there will be substantial variation in the choices and rankings. Ask the same experts about the top 5 skaters, and they'll all have Gretzky, Orr, Lemieux, Howe, and maybe Richard, probably in that order.

Since it seems to be a lot easier to rank players, I thought of the question: If Martin Brodeur was a skater, who would he be? Which forward or defenceman is very similar to Brodeur in terms of career accomplishments? I think I have come up with a pretty close fit: Joe Sakic.

Both have been consistent performers. Sakic has scored over a point per game almost every season, and 6 times went over 100 points. Marty Brodeur has posted 10 consecutive 35 win seasons.

Both have piled up the numbers. Sakic will likely rank top 5 all-time in points scored by the time he retires, and probably in the top 10 in playoff points as well. Brodeur is 2nd all-time in wins and 3rd in shutouts.

Both are well-respected in hockey. Sakic is seen as a great leader, the only captain the Avalanche have ever had. Brodeur is seen as classy and a nice guy.

They have both led consistently good teams. The Avalanche have been a good team since 1995. The Devils have been near the top of the league since 1994.

Both are known as clutch players. Sakic has won 2 Cups, a Conn Smythe Trophy (for his 18 goal playoff season in 1996), and has a better goal scoring record in the playoffs than the regular season. Brodeur has 3 Cups and a better playoff save percentage and GAA than during the regular season.

Both were overshadowed by other players early in their careers. Sakic played in the shadow of other great centers like Gretzky and Lemieux, as well as Yzerman. Similarly, throughout the 1990s most of the goalie recognition went to Roy and Hasek, rather than Brodeur.

This led to late recognition of their excellence. Sakic never made the NHL First or Second All-Star teams until his MVP season in 2001, and then he made it to the First Team again in 2002 and 2004. Brodeur never won any Vezinas until 2003, then won two in a row. They both did however play in many All-Star Games throughout their careers: Sakic in 12, Brodeur in 9.

Both have great international records. Sakic was the MVP of the 2002 Olympics, and captain of the 2006 team. Brodeur was the most outstanding goalie of the 2004 World Cup, as well as an Olympic gold medallist.

The two players appear to be a pretty similar in terms of team success, reputation, and individual awards. The point is, though, that while Joe Sakic is greatly respected and is considered one of the top players of his generation, nobody in their right minds would think of Joe Sakic as a candidate for the greatest ever. He is probably top 50, maybe top 30 at best. The greatest of all-time are Gretzky, Orr, and Lemieux. The reason for that is obvious: their individual trophy hauls are just staggering. Gretzky has 9 MVPs, 10 Art Rosses, and 2 Conn Smythes. Lemieux has 3 MVPs, 6 Art Rosses, 2 Conn Smythes, and a Calder. Bobby Orr has 2 Art Rosses, a Calder, 2 Conn Smythes, 3 Hart Trophies, and 8 Norris Trophies. Even Gordie Howe, a guy who is known mainly for his longevity, won a good deal of hardware, including 6 scoring titles and 6 Hart Trophies.

In contrast to skaters, goalies are usually judged based on their wins, Cups, and shutouts. All of these are very team dependent. They should instead be judged on their trophies won, just like players are. On that basis, the best might be someone like Dominik Hasek, with 6 Vezina Trophies, 2 Hart Trophies, and 6 1st All-Star Teams. Or maybe Ken Dryden: 5 Vezinas, a Conn Smythe, a Calder, and 5 First Team All-Stars in a short career. Other candidates are Jacques Plante, who had 7 Vezinas, 1 Hart Trophy, and was 3 times a First Team All-Star, and Patrick Roy, with his 3 Vezinas, 4 First Team All-Star selections, and 3 Conn Smythe trophies.

When it comes to major individual awards, both Sakic and Brodeur fall a little short. Sakic has a Conn Smythe, a Hart, a Pearson, and 3 First All-Star nods. Brodeur has a Calder, 2 Vezinas, and 2 First All-Star nominations.

Just like Joe Sakic, Marty Brodeur has had some notable career accomplishments. Just like Joe Sakic, Brodeur is considered one of the best at his position during the time period in which he played. And just as Joe Sakic is not even considered in discussions about the greatest of all-time, neither should one consider Martin Brodeur. The reason is that their performance statistics pale in comparison with the very best, and they do not have the trophies and awards that indicate dominant performance. The main difference between the two in terms of their all-time ranking in the game is that goalies are much more likely to be judged based on the success of their teams than skaters are.

Thursday, May 10, 2007

Ryan Miller and the Grant Fuhr Defence

"In the regular season the high flying Oilers often left the goaltenders to fend for themselves. This devalued (Grant) Fuhr's career numbers, but those who saw him play in the playoffs know he was one of the top goalies of his era. He has 5 Stanley Cup rings to prove it. He was the goalie who might have let in a lot of goals because of the Oilers philosophy of simply outscoring the opposition instead of shutting them down, but he was also the goalie who would always make the big save when the team really needed it."

Grant Fuhr profile at SabresAlumni.com

"Miller plays a very Grant Fuhr-esque game...plays for a high scoring team that plays a high risk game resulting in a lot of quality scoring chances against, but almost always makes the big saves to win the game."

A Sabres fan, posting on HFboards

Ryan Miller is another goalie that inspires controversy. He was elected as an All-Star Game starter this year, although many disagreed with the choice. Some name him as the reason for Buffalo's success, while others say that he is very overrated. Even Buffalo fans seem to fluctuate back and forth, although in these playoffs he is winning most of them over by stepping up his play.

One of the main arguments for the pro-Miller side is that he permits the Sabres to use their "run-and-gun" offence. The Sabres take chances offensively, they say, so Miller gets called on to make big saves the other way to keep them in the game. This could be called the "Grant Fuhr defence" - that an excellent offensive team still needs good goaltending to win, and even if their goalie's stats aren't great, the shots he faces are tougher and the fact that they win is proof that he made the big saves his team needed.

The problem with the Grant Fuhr defence is that it is often a false perception. Fans see a team that scores goals, and assume that it must be trading off defence for increased offence. This is certainly not always the case. This year, for example, the Detroit Red Wings took the most shots and allowed the least shots in the league. One dimensional teams do not usually win anything; good players and good teams are often just as good defensively as they are offensively.

As the President's Trophy winners, Buffalo is certainly a good team, so just how good is their defence? Evidence suggests that they are actually very good. According to Ken Krzywicki, they were third in the league in terms of shot quality against in 2005-06. According to Chris Boersma, they ranked first in the NHL in terms of lowest shot quality against at even strength this season. The Sabres give up an average number of shots, but limit most of them to the perimeter. The unsung heroes are their defencemen, led by Henrik Tallinder and Brian Campbell. They are effective at moving the puck and limiting chances, and all are plus players at even strength. The key to Sabre goal prevention is their blue line corps; despite the Sabres' offensive strength, Buffalo is actually a relatively easy place to play goal.

Perhaps a more serious problem with the Grant Fuhr defence is that it is totally illogical. How is it tougher to play for a dominant offensive team? It is much tougher to play against that team. Let us say there are two teams, one that has Wayne Gretzky, Mark Messier, Jari Kurri, Paul Coffey, and Glenn Anderson. The other is an average team, say the mid-'80's Winnipeg Jets. The game is a back-and-forth, wide open contest. Both goalies face 30 shots. Which goalie do you think faced harder shots? The goalie facing the Hall of Famers, or the goalie facing the Winnipeg Jets? Sure, the Oilers probably would give up some good scoring chances, but as one of the greatest offensive teams ever they would have more dangerous scoring chances and have better shooters taking the shots. The playoff stats for Grant Fuhr from 1986-1989 from hockeygoalies.org seemed to prove this point: with Fuhr in net, Edmonton outshot their opponents by an average of 2 shots per game, and scored on 13% of their shots in the usually more tightly played playoffs, at a time when the regular season scoring rate was around 12%. If the Oilers were winning by only one goal, it would not be because the Edmonton goalie was making the big, clutch saves, it would be because the Winnipeg goalie was outplaying him.

It is the same with the Sabres. How many teams in the league can skate and score with the Sabres, the highest scoring team in the league? If the game is played wide open, Buffalo is usually going to create more dangerous chances than their opponents. They also have great shooters, as the Sabres scored on 12.3% of their shots this year, the best rate in the NHL. They may allow some scoring chances against, but with their skill and scoring depth they can more than easily make up for a few goals against. It does not seem at all reasonable to credit the goalie for success when a team's biggest strength is offence. They are like pitchers on a team of sluggers, and their win totals are similarly inflated. The Grant Fuhr defence appears to be just a biased argument designed to justify crediting a goalie for playing on a winning team.

Ryan Miller had 40 wins this season. He faced 30.7 shots a game, about average in the offensive Eastern Conference, but as previously mentioned those shots he faced were easier to deal with than average. His save percentage was a middling .911. Miller is not yet an elite goalie. His overall performance was no better than average, and the wins were mostly because of the scoring exploits of his teammates. In fact, he might have been better last season. However, with the Sabres' excellent regular season and second straight deep playoff run, he is developing a reputation as a winner.

Game 6 of the Buffalo-Islanders series was a typical example of the Grant Fuhr phenomenon at work. Buffalo led 4-1 after two periods, but gave up 2 late goals as the Islanders looked to fight back and tie it up. Miller made a sprawling glove save on his back in the last minute (Youtube). That is the kind of save that helps build a reputation as a clutch goalie. Of course, if he did not have the luxury of a team in front of him that scored 4 goals, it would have likely been a much different story. Such is the enviable position of playing for an offensive powerhouse. I like to think that somewhere Grant Fuhr was watching and smiling.

Conference Finals Game 1 update: The Sabres defence looked pretty shaky last night. That likely had a lot to do with the strength of the Senators, but it will be interesting to see if their performance will improve, or if Ottawa will continue to outplay Buffalo. Miller was decent, but if the Senators maintain an edge in play I don't think he will be able to make the difference.

Wednesday, May 9, 2007

Cujo the Choker

Curtis Joseph is a playoff choker, a loser, someone you can't count on when the games get really important. He has some great playoff performances, but they all were in the first round. When the games really start to get tough, then Cujo just disappears. He just doesn't have that clutch ability, that calmness under pressure that separates him from winners like Brodeur. Or so the conventional wisdom goes.

Let's look at how Cujo did in the first round:

Won 8 of 12 series, 39-29 record, .922 save %

That's pretty good. In fact, it's strikingly similar to Brodeur's line:

Won 8 of 12 series, 40-26 record, .927 save %

But in the second round, the tables turn.

Curtis Joseph: Won 2 of 8 series, 20-29 record, .914 save %
Martin Brodeur: Won 5 of 8 series, 20-22 record, .916 save %

What happened? Did Cujo choke? Not at all. His save percentage is almost the same as Brodeur's. His win/loss record isn't too far off either. Let's look at the teams he was playing on:

Average Cujo team, 1st round: 91 points
Average Cujo team, 2nd round: 93 points

So on average, Curtis Joseph was playing on teams that were just a little bit above .500. What about his opponents?

Average Cujo opponent, 1st round: 92 points
Average Cujo opponent, 2nd round: 101 points

This explains the first and second round splits. In round one, Cujo's teams have been about as good as their opponents on average. He has a winning record, including several times when he stole games against teams much better than his own. In the second round, Cujo's opponents are 100 point teams that have almost always been much better than his own team. When his team was better, Cujo went to the Conference Finals in 2 times out of 3. The rest of the time they were huge mismatches against better opponents, and Cujo and his teammates lost out every single time.

How about Martin Brodeur? Is he also facing tougher competition in the second round? Not really.

Average Brodeur team, 1st round: 100 points
Average Brodeur opponent, 1st round: 88 points

Average Brodeur team, 2nd round: 99 points
Average Brodeur opponent, 2nd round: 93 points

Thus the secret for Brodeur's playoff success is revealed: Play on the better team. Brodeur's teams have been better than Joseph's, and his opponents have been weaker. In that context, his second-round performance is not more impressive than Joseph's, it is actually worse.

Curtis Joseph has probably been better than Martin Brodeur in the first AND second rounds of the NHL playoffs throughout his career. The average team that Cujo lost to in the second round had 107 points. In his entire career, he only once lost to a weaker opponent in the second round, and that was despite a .928 save percentage in the series, because his Red Wings could not score on Miikka Kiprusoff (back-to-back 1-0 losses in games 5 and 6). Martin Brodeur has a sub .500 win/loss record and only a slightly higher save percentage than Joseph in the second round, despite playing on better teams that outshot and outplayed their opponents.

Curtis Joseph is not a choker, a loser, or any such thing. He has been an excellent goalie throughout his career, including in the second round of the playoffs. His career is yet another cautionary tale about the folly of rating goalies based on team accomplishments.

Tuesday, May 8, 2007

The 10 Types of Overrated

I have identified ten types of players who tend to be overrated by fans and media.

1. Players in large media markets or with high media exposure
2. Players playing on dominant teams
3. Players who have clutch reputations
4. Players who are made to look good by teammates
5. Players who make highlight reel plays
6. Players who are aging and losing their once outstanding skills
7. Players who are good at doing things that are overvalued by the fans
8. Players who are media favourites.
9. Players who play important positions and are inappropriately credited for team results
10. Players who have large counting stats.

Nearly every professional athlete will have several of these factors apply to them, but there are usually some other factors that tend to balance them out. There are, however, a few athletes that seem to meet all the criteria for a "perfect storm" of overratedness, players that are almost universally considered to be better than they actually perform. One of those athletes is Martin Brodeur.

1. Players in large media markets or with high media exposure

Brodeur plays in New Jersey, which is close to New York. He is well-known to fans in New York and Philadelphia, since he has played in the same division for his entire career. In addition, his international play for Team Canada has resulted in tremendous media exposure around the world.

2. Players playing on dominant teams

Brodeur has played on an excellent team for nearly his entire career.

3. Players who have clutch reputations

Brodeur is considered a clutch performer, and on first glance at his statistics it appears that Brodeur has indeed raised his game in the playoffs. However, a closer look shows that his success is not really so impressive, once you factor in New Jersey's overall talent, strong defensive play, underrated offensive abilities, and team discipline with respect to taking penalties. In addition, studies of Brodeur's save percentage in various clutch situations do not indicate any special clutch ability.

4. Players who are made to look good by teammates

Regardless of what one thinks of Brodeur's current situation, this is difficult to debate for his early career, when he would face an average of just 23-24 shots per game behind a very good defence.

5. Players who make highlight reel plays

Brodeur is a goalie that often makes big saves. He relies on his athleticism to react to pucks rather than simply blocking them, and is more likely to move around and make impressive looking saves than many of his counterparts, even those with similar levels of efficiency. His scoring a goal in the playoffs was also the kind of signature highlight that fans remember long afterwards.

6. Players who are aging and losing their once outstanding skills

Brodeur had a strong year this year, but in general his best seasons came earlier in his career, between 1995 and 1999.

7. Players who are good at doing things that are overvalued by fans

Many point to Brodeur's rebound control and puckhandling skills as reasons why he is the best goalie in the league. While he is undeniably good at both of those skills, they are both secondary to stopping the puck in the job description of a goaltender and many onlookers overestimate their impact.

8. Players who are media favourites

Listen to hockey broadcasters, analysts and discussion panels and you will hear overwhelming praise for the well-spoken and affable Brodeur.

9. Players who play important positions and are inappropriately credited for team results

This is true for all hockey goalies. Brodeur just happens to have played on a consistently successful team, which magnifies the effect.

10. Players who have large counting stats.

Brodeur will likely retire as the number one ranked goalie all-time in wins and shutouts. Both of these stats are very strongly team-influenced.

Why You Can't Say Anything Bad about Brodeur

Sportswriter Jim Kelly recently dared to suggest that Brodeur was at fault for the game 4 loss against Ottawa.

Here are some of the responses:

"Brodeur is the only player in NJ keeping them alive"
"the number of shutouts Brodeur has (92) is higher than your IQ"
"If it wasn't for Brodeur, the Devils would have went out in the first round against T-Bay"
"People have been saying how solid Ray Emery has been, but I still say Brodeur is making more saves, and bigger saves"
"You can't sit there as a legitimate hockey fan and say anything bad about Brodeur. The guy is one of the best of all time and always will be. Argueably the best of his generation. While you sit there and write about the couple softer goals that went in you forget to mention the dozens of incredible saves he's made look easy in this series."
"Brodeur is by far the best goalie in the league"
"It's not hard to imagine them (Ottawa) scoring 6-7 goals most games against a lesser but still competitive goalie"
"This must be a joke...broduer has how many cups ? How many wins ? How many vezina's ? How many shutouts ? How many olympic gold medals ? How many....? This opinion is not even worth the screen its written on."

All hail the power of a great reputation! A goalie has a below average playoffs, and lets in two soft goals in a 3-2 loss, and yet dozens of fans come out and bash a sportswriter who suggests that just maybe it might have been the goalie's fault.

One comment (that I have seen reproduced many times in different forums and media) shows particularly clearly how goaltending is entirely based on reputation and past performance, rather than what is going on right now:

"And for those of you on the luongo train. Get a grip. MAYBE once he makes it to the playoffs more than once, and MAYBE if he gets past the second round, and MAYBE if he wins a vezina, and MAYBE if he EVER wins ONE cup then MAYBE he will be able to be menitoned in the same breath as marty...until then give brodeur his due."

I don't know if there can be found in any other sport the equivalent of this logic that only team accomplishments in the playoffs from the past matter, and a younger athlete simply cannot be considered as good as a much older one with more team accomplishments, regardless of current performance. Maybe occasionally a football quarterback, but that's probably it.

Roberto Luongo is the most talented goalie in the NHL. He is in the prime of his career, he just completed his sixth consecutive excellent season in the league, and he was the best goalie in this year's playoffs. To not consider him at least the equal of the 35-year old Martin Brodeur is completely unreasonable. Hockey fans should either for consistency's sake start claiming that Sidney Crosby needs to win a Cup to be considered better than Martin St. Louis, or preferably stop treating goalies as such a special case and just do the same thing they do for all other players and evaluate them based on how they are playing right now.

Monday, May 7, 2007

Brodeur's Playoffs 2007 - Analysis

Here is a breakdown of how Brodeur performed in the 2007 playoffs, using my own statistical analysis, as well as borrowing on the work done by Chris Boersma at Hockey Numbers:

vs. Tampa Bay:


Brodeur was mediocre against Tampa Bay. He stopped 91.7% of the shots he faced, but he faced easier than average shots (expected save percentage of .923), meaning I have his series shot-quality neutral save percentage at .909. Most of his value, however, came from game 5, where he posted a 31 save shutout. In the first three games of the series, he was terrible (over 1 goal below average all in all three games). In the other two games he was about average. So I have him as costing his team one game, winning his team one game, and playing slightly below average in the rest of them.

For the series, I have Tampa Bay as being expected to score 12.9 goals. Hockey Numbers calculated it as 11.5. The Lightning actually scored 14, indicating that Brodeur struggled.

vs. Ottawa:

Martin Brodeur was about average against Ottawa. He had two very good games, one awful game, one mediocre game, and two average games. The result was a .915 save percentage against nearly average shot difficulty, for a shot-quality neutral save percentage of .917.

Overall, I had Ottawa as being expected to score 14.4 goals, pretty close to Hockey Numbers' 14.3. Excluding empty netters, Ottawa scored 14 goals, or pretty much what was expected.

Just as in the Tampa series, Brodeur won his team one game (game 2) and lost his team one game (game 1). He was good in the game 3 loss as well, but Emery outplayed him in games 1, 4, and 5.

Overall:

In the playoffs, Brodeur had a below average save percentage (.916 with a playoff average of .922). I had him expected to have a .918 save percentage, meaning that he was 0.7 goals below average and had a shot-quality neutral save percentage of .913. Hockey Numbers calculated his shot-quality adjusted efficiency at .891, and rated him at 2.2 goals below average.

Brodeur faced more shots than his opponent in both series, but in both series he faced a greater percentage of easy chances while his opponent had to make more tough saves. From the play-by-play records, New Jersey had 42 excellent chances (>20% chance of scoring), compared to just 23 against Brodeur. Brodeur faced more easy saves (<5% chance of going in) than his opponents as well, facing 125 gimmes to 110 for his playoff rivals. The average shot against Brodeur was from farther away than the average shot against either Emery or Holmqvist. Hockey Numbers has the Devils ranked #2 in shot quality against, which supports my numbers. New Jersey gave up a lot of shots, but did a very good job at taking away the most dangerous scoring chances. This also matches my subjective viewpoint from watching the games.

Brodeur's biggest problem in the playoffs was actually those routine shots. Hockey Numbers tracks the save percentage against easy shots, those with less than 5% chance of going in. The playoff average save percentage against such shots is 98%. Brodeur stopped just 95% of them, worst of all playoff goalies.

I have Brodeur as being outplayed by the opposing goalie in 5 out of 11 games. Hockey Numbers says he was outplayed in 7 of them. Both of us have Brodeur letting in more goals than an average goalie would have in 7 out of 11 games (64%).

All things considered, Brodeur was weak in the playoffs. He was not the reason his team lost, but goaltending was not a strength for New Jersey as many had expected. He had a few excellent games (game 5 against Tampa, games 2 and 3 against Ottawa), but they were balanced by some clear off-games (such as in the opening games in both series). On most nights his play was slightly below average, and he also let in more soft goals than any other playoff goaltender. All in all, he did not even come close to living up to his reputation as Martin Brodeur, clutch goalie extraordinaire.

Sunday, May 6, 2007

Somebody Agrees With Me

Here is a comment from James Mirtle's hockey blog:

At 9:43 AM, May 04, 2007, mike said...

"Only a man who is convinced that fans show up at hockey games to see goalies play the puck behind the net while being protected from any contact would say this about his opposing netminder when he and his team are down 3-1 in games in a best-of-seven series. Only a man who feels that hockey's problems can be fixed by allowing ever larger pieces of equipment to be worn by goalies would say something this stupid about Ray Emery's play against the NJ Devils. Only the most overrated hockey player of the last 25 years could sound this stupid--criticizing his opponent while that opponent is performing at a much higher level than he is--after giving up two more unbelievably soft goals in a crucial playoff game.

Is there any netminder in the NHL who has given up as many softies in the playoffs as this guy? I don't care how many Cups he has won, if the Devils had Mike Richter, Ed Belfour, Mike Vernon, or Dominik Hasek that team would have won five Cups in a row. Once again he is being outplayed in the playoffs by someone widely perceived as an inferior goalie to him...when does the hype job or the whining ever stop with this athlete?"

Friday, May 4, 2007

Brodeur's Playoffs

I haven't posted much about the current playoffs, mostly because I don't believe in judging players solely by a playoff series or two, but also because I obviously haven't been as shocked as some at Brodeur's struggles.

I actually have been impressed by a few of the things Brodeur has done this playoff season, including performing well in overtime to improve his record in that department to 10-19, as well as to win 2 games where his team was outshot by 10+ shots, something he had only done 3 times in his playoff career before this season. I think he probably has some game stealing ability he hasn't previously shown, mostly because he didn't really have the opportunity.

However, on the whole he has been subpar. His save percentage is .919, which appears good, until you consider that the average save percentage in the playoffs so far has been .922, and many of the goals he has let in have been softies.

Because of his reputation, though, Brodeur sure gets cut a lot of slack by the media. His routine saves are described as masterpieces, and every time he plays a good game they write the same "Brodeur is back" article that has so far invariably been followed by another up-and-down performance.

New Jersey is still a solid defensive team, they still play the same defensive system and collapse down in front of the net to take away the most dangerous scoring chances, but they are giving up more shots than in years past. In terms of league ranking, this year and last year were the two worst years for New Jersey shot prevention since Brodeur's rookie season, and that has continued in the playoffs.

How's this for an amazing stat: Before the lockout, in Martin Brodeur's entire career the New Jersey Devils had only ever been outshot twice over an entire playoff series. That covers 25 series over 11 seasons of hockey, and both times were against President's Trophy winning teams. It's been different since the "New NHL" started, though. Last year both the Hurricanes and Rangers outshot the Devils. This year the Lightning matched them, and Ottawa has a significant edge so far. Brodeur's pre-lockout playoff record: 84-60. Post-lockout: 10-9. Coincidence? I think not.

With the shot totals no longer in his favour, Brodeur's margin for error is now greatly diminished, and some of those extra half-chances are ending up in the back of his net. We should not be tremendously surprised by a bad game here and there, or a soft goal here and there, because they can be found throughout Brodeur's entire playoff history, it is just that in the past New Jersey was usually good enough to make up for it. This year, it appears that Ottawa is too good for the Devils, as even when Brodeur is on his game, as he was in game 3, his team just wasn't good enough to win. It almost always comes down to the team in the end, rather than the goaltender, and usually Brodeur has been on the side of the better team. This time he isn't, so it looks like he is going to lose, it's as simple as that.

Thursday, May 3, 2007

Simpson's Paradox

In my last post, I discussed the impact of special teams on goaltending. Here is an interesting example of that effect at work.

One of the main arguments for Brodeur as Vezina Trophy winner is that he leads Luongo and the other nominees in wins, goals against average, and save percentage. Wins are very team dependent, and goals against average is basically save percentage times the number of shots against, so these two are not very good measures. If Brodeur, however, ranks ahead of Luongo in save percentage, that is a good argument to use on his behalf. During the 2006-07 season, he did just that, with a .922 save percentage to Luongo's .921.

A closer look shows that this was not because Brodeur played better, but rather was caused by a large discrepancy in special team play between the Canucks and Devils. Here are the save percentage splits for Brodeur and Luongo for each game situation (even strength / penalty kill / power play):

Martin Brodeur: .927 / .904 / .888
Roberto Luongo: .928 / .906 / .910

Luongo ranks ahead of Brodeur in every game situation. How can he possibly end up behind Marty in overall save percentage? The reason is that Vancouver took more penalties, meaning Luongo had to face 199 more shots on the penalty kill than Brodeur. If Luongo had faced Brodeur's shots, and vice versa, and both of them stopped the puck at exactly the same rates, Luongo's save percentage would have been .923, and Brodeur's would have been .920.

In statistics, this is called Simpson's Paradox, where the ranking of several groups changes when they are combined. Luongo ranks ahead of Brodeur in each special team state, but then ends up behind in the overall rate. This is yet another example of hidden team effects, and how they can impact goaltenders.

Tuesday, May 1, 2007

The Disciplined Devils

The New Jersey Devils are the best team in the league at not taking penalties. For the last 5 years in a row, the New Jersey Devils have led the NHL in fewest times shorthanded. They have averaged just 283 times shorthanded per season (3.5 per game), which is nearly 10% better than the second place finishers over that same time period. Given that the average even-strength shot had about an 8% chance of going in this year, while the average power play shot was likely to be a goal 13% of the time (and 19% of the time with a two-man advantage), facing fewer power plays is clearly a big advantage to the goalies on disciplined teams.

Martin Brodeur ranked 1st among all NHL starters this season in terms of the lowest percentage of shots faced on the power play. Just 15.8% of his shots came with a man disadvantage. Some other goalies had it a lot tougher, such as for example Roberto Luongo, Cristobal Huet, and Olaf Kolzig, all of whom faced over 25% of their shots on the penalty kill.

Using each goalie's shots faced on special teams and at even strength, I calculated their expected save percentages based on the league average for starting goalies at each game situation. Brodeur ranked first with an expected save percentage of .911, compared to a league average of .908. Luongo, Huet, and Kolzig all were expected to finish around .906. Brodeur's save percentage should therefore be about .005 better than Luongo's, simply because of his special teams advantage. It was virtually the same, which indicates which goalie has had a better year.

New Jersey's penalty kill has not only been infrequent, it has been very good. In each of the last 5 years, they have ranked among the top 6 teams in the league in terms of fewest power play goals against. This has a lot to do with them not taking penalties, but they also ranked top 6 in terms of efficiency in three of those years, including this one. Could it be that Brodeur is an excellent penalty killer who just never gets a chance to shine?

Evidence suggests no. This year he was very good, with a .904 shorthanded save percentage, 5th in the league among starting goalies. The previous four years, however, his best rank was 12th and his save percentages ranged from .849 to .878. In 2002-03, Brodeur's Vezina and Stanley Cup winning season, he stopped just 86.6% of power play shots against, for 18th best among starters. Nevertheless, the Devils ranked 1st in the league in efficiency, and led the league with just 32 power play goals against all season, indicating they did an excellent job in front of their goalie. New Jersey's success on the penalty kill is primarily a function of their discipline and defensive play up front, more so than the performance of the goaltender.