Wednesday, February 28, 2007

Why Billy Smith Is Not a Hall-of-Famer

Billy Smith is one of the most overrated goalies in hockey history. He played a total of 680 regular season games, but his entire reputation comes from 93 playoff games played between 1980 and 1984. In that period, the New York Islanders reached the Stanley Cup Finals five years in a row, winning four times before finally losing to the Edmonton Oilers. This gave Smith a reputation as a "winner" and a "money goalie", and booked his place in the Hall of Fame.

But how much did the Islanders, a team good enough to win four Stanley Cups in a row, rely on their goaltender? Evidence suggests not very much. The Islanders were consistently one of the top scoring teams in the league. In 1981, they led the league in scoring. In 1982, they finished 2nd to Edmonton. After a down year in 1983 (15th), they finished 3rd in 1984. From 1981 to 1984, the Islanders averaged 4.4 goals per game. In the playoffs, they were even better, averaged 4.7 goals per game over their 4 Cup runs.

During their Cup winning playoff seasons, the Islanders were ridiculously dominant. They went 15-6 in 1980, 15-3 in 1981, 15-4 in 1982, and 15-5 in 1983. In those four years, they only faced elimination once, winning the decisive game 5 against the Penguins in 1982. They also only ever trailed in a series once, after dropping the opener against the Rangers that same playoff year (which they followed up by rattling off 12 wins in 13 games to bring home the Cup).

The Islanders also had an incredibly low percentage of one-goal games. Only 20 of 99 playoff games were decided by a single goal, and in those games the Islanders went 13-7. That seems pretty good, but it is actually much worse than their overall record of 72-27. The record was mostly fueled by the 1979-80 team, the weakest of the Stanley Cup winning squads. From 1981-85, the Islanders had only 13 games decided by one goal in a stretch of 78 playoff matches, winning 7 and losing 6, indicating that Smith didn't exactly steal the close ones. And over 80% of the time, he could have let in another goal without even changing the final result.

So, evidence suggests that the Islanders didn't need top flight goaltending, since their team was so powerful they would usually win anyway. That reflects on Smith's value, but he doesn't really address his actual level of performance, which is what we turn to next.

Most of the games Smith played were before save percentage was an official stat, so we will have to rely on the traditional stats for the most part. Back then, however, the Islanders tended to platoon goaltenders during the regular season, which allows for a good comparison.

Billy Smith was an Islander from when they entered the league as an expansion team in 1972-73 until his retirement in 1989. Over that period, he played mainly with three goalies: Chico Resch, Rollie Melanson, and Kelly Hrudey. All three significantly outplayed Billy Smith in the regular season, but for some reason Smith was given the playoff starts that enabled him to build his reputation as a winner.

Between 1975 and 1981, when Smith and Resch shared the load, Billy Smith had a lower GAA in a season only once. Only twice did he play more games than Resch, and only twice did he have a higher winning percentage. For the period, Resch recorded a 2.55 GAA and a .662 winning percentage, much better than Smith's 2.81 and .621. Resch also posted 25 shutouts to Smith's 15. Nevertheless, Smith played 30 playoff games, nearly as many as Resch's 34. In those games, Smith's GAA was 2.87, worse than Resch's 2.49, but his record was 16-11 while Resch's was only 17-17. Maybe this apparent run of luck by Smith was what started to earn him a reputation as a clutch performer. A couple of shaky games by Resch in 1980 opened the door for Smith, and he played in 20 of 21 games as the Islanders won their first Cup.

Resch was traded during the 1980-81 season, and Smith's new playing partner became the 21 year-old Rollie Melanson. Despite his youth, Melanson posted excellent numbers over the next 3 seasons, recording a cumulative .906 save percentage, better than Smith's .901, and also besting Smith in goals against average and winning percentage. Nevertheless, Smith got the starts come playoff time, and rode along with the Islanders juggernaut to Cup after Cup.

Smith was awarded the Conn Smythe trophy as playoff MVP in 1983. He played in 17 games that year, going 13-3 with a 2.68 GAA. The Islanders destroyed their opposition, scoring 4.7 goals per game, going 15-5 and never facing elimination. They only played in one game that was decided by a single goal, a game that they lost. Mike Bossy led the team in scoring with 17 goals in 19 games (exactly the same as his Conn Smythe peformance in 1982), Denis Potvin added 8 goals and 12 assists from the blue line, and Brent Sutter had an excellent playoffs with 10 goals and 21 points to finish 2nd on the team in scoring. Melanson played in 5 games and posted a lower GAA than Smith. Yet somehow, the voters decided that the MVP of this scoring machine was the goalie, Billy Smith.

Melanson was traded in 1984, as another top prospect, Kelly Hrudey, was waiting in the wings. Hrudey arrived at the tail end of the Islanders' run of dominance, and once again Smith was soundly outplayed by his teammate. Hrudey's save percentage between 1984 and 1989 was .889, Smith's was .879. Hrudey posted more shutouts, a lower GAA, and a better winning percentage. He also took over in the playoffs from Smith in 1985, and Smith rode the pine during the playoffs for the rest of his career.

Another career accomplishment of Smith's is that he has the honor of being the first goalie to be voted the Vezina winner, after the rules were changed in 1982 from automatically awarding the trophy to the goalie with the lowest goals against. That season, Pete Peeters and Philadelphia had off-years, and Montreal (the Jennings trophy winners) split their playing time among three different goalies, leaving the door wide open. Richard Brodeur in Vancouver was almost certainly the best goalie that year, posting a winning record and a 3.35 GAA for a 77 point Vancouver team that scored only 290 goals (4th worst in the league), playing in the same division as Gretzky's Oilers. Brodeur's backup Glen Hanlon, an NHL starter for most of his career, had a .389 winning percentage and a 3.95 GAA in 28 games behind the same defence. But Smith led the league in wins and GAA, and as it so often does, that managed to swung the vote his way.

In summary, Billy Smith had an unremarkable career. He was almost always the second best goalie on his team, and his regular season stats are not outstanding. His reputation is based entirely on his playoff performance, played entirely behind one of the greatest teams of all-time. The Islanders of the early '80s were a dominant offence and defence that won Stanley Cups virtually unchallenged. They weren't reliant on their goaltending at all - their wins came from their prolific offence, and most of the time Smith could have given up another goal or two and still got the W. If Resch, Melanson or Hrudey had been in net instead of Smith, it is completely reasonable to expect that the Islanders would have won all of their Cups. In addition, his Vezina and Conn Smythe trophies were both questionable choices and probably undeserved. Billy Smith is one of the best examples of how a team can make their goalie look good; he is in the Hall of Fame today more because of what Trottier, Bossy, Potvin, and the Sutters did than anything he did on his own.

Thursday, February 22, 2007

The Difference-Makers

Which goalie has been the biggest difference maker this year? There are many ways to look at this question, like value over average, team record vs. Pythagorean projection with an average goalie in the net, etc.

However, I am going to look at it in a different way - I am going to compare the records of goalies when they have a quality game (defined as 2 goals against or less) to when they have a poor game (defined as 4 goals against or more), and see which goalies have the greatest discrepancy. That is, which goalies are almost guaranteed to win when they have a good game and almost guaranteed to lose when they have a poor game. The goalies playing for those teams tend to have individual contributions that are very significant in determining the final result.

This categorization is of course very strongly team-influenced. A goalie can be good without being a big difference-maker. It is more difficult for a goalie on a good team to be a huge difference-maker, since the team often makes up for his weak games. Take, for example, J.S. Giguere of the Anaheim Ducks. Giguere has had an excellent season, racking up a .922 save percentage, although he is doing it behind a good defensive team. When he posts a quality game (2 goals against or less), his record is 20-4-3, which is good but not exceptional when compared to other goalies around the league. When Giguere lets in 4 goals or more, the Ducks are an amazing 5-2-2 for a .667 winning percentage. Clearly, the Ducks are an outstanding team and Giguere's performance, while excellent, is not pivotal to their success.

My expectation was that Roberto Luongo would rank very highly based on this measure, since he is an outstanding goalie on a low-scoring team. Others who are similar are Kiprusoff and Brodeur. However, those three goalies all won a reasonable amount of games even when they played poorly. They were certainly large difference-makers, they just weren't the biggest.

The NHL goalie who has been the biggest difference-maker this year to his team has been Andrew Raycroft of the Toronto Maple Leafs. Raycroft has had an up and down season, but his splits are like night and day: when Raycroft has a quality game, the Leafs are 23-0-2. When Raycroft has a poor game, the Leafs are 1-11-2. His play doesn't seem to depend on the number of shots against; Raycroft gets about 29 or 30 shots a game no matter how well or poorly he plays. The correlation coefficient between goals against and shots against for Raycroft is a mere 0.27, indicating very little effect. One reason is that Raycroft has a high tendency to play atrocious games that make it very difficult for his team to win, as he has let in 5 or more goals 8 times in 49 starts and has been pulled 4 times. The Leafs have an excellent record when Raycroft plays well, in part because the Leafs have scored 4 goals 16 times in his 48 starts, seven times winning 4-2 and four times winning 4-1. Part of it is that both goalie and team have been hot at the same times, but the Leafs can score goals so as long as they hold it together on the defensive end they give themselves a good chance to win.

Here is a case where my subjective viewpoint collides head-on with a statistical perspective. My anecdotal viewpoint, before this study, was that Raycroft more or less plays as well as everyone else around him, and people tend to blame or credit him for the result simply because he plays the most important position. However, I think I will have to defer to the stats here. Raycroft was quite a bit more streaky than I would have expected, both in terms of good games and bad games, and it doesn't look like team defence was as big a factor as his personal performance. Therefore, it appears that Raycroft is, like or hate it for Leaf fans, extremely important to the success of the Blue and White.

The second biggest difference maker is Olaf Kolzig. He plays for a mediocre Washington team, and has had an excellent season when that is taken into account, posting a .909 save percentage and helping keep Washington on the fringes of the playoff race despite facing over 33 shots a game. When Kolzig allows 2 goals or less, Washington is 14-1-3; when he lets in 4 or more, the Caps are 1-13-2.

In third spot is Dwayne Roloson of the Edmonton Oilers. Edmonton is 20-3-2 when he puts together a quality game, and 1-12-1 when he plays poorly.

Rounding out the top five are Manny Fernandez and Tim Thomas. Luongo comes in seventh, Brodeur eighth, and Kiprusoff eleventh. Some of the goalies that rank low based on this measure include Miller, Emery, DiPietro, Fleury and Giguere.

I think this brief study supports the belief that goaltending is more important for mediocre and bad teams than for good teams. The teams with the greatest differences were all teams that are struggling to try to make the playoffs. Two of them, Raycroft and Thomas, aren't having particularly good years, but interestingly enough streaky goalies seem to have extra value on bad teams. When they are bad, the team is likely to lose anyway, but when they are good they give the team an excellent chance to win.

Therefore, before rushing to credit goalies on dominant teams for their success, make sure they truly are making a difference in the final results.

Tuesday, February 20, 2007

The Value of a Shutout

Goalies love shutouts. They guarantee victory and usually result in first star billing. They mean that the goalie has done his job perfectly through 60 full minutes of play. Goalies who post lots of shutouts attract attention; one reason that Martin Brodeur's current season is being considered very special is that he has posted 11 shutouts, nearly twice as many as anyone else.

But what is the value of a shutout? First of all, shutouts are very team dependent. They become more difficult with each additional shot, so goalies that face more shots will generally record fewer shutouts. Shot difficulty is also a factor - the number of odd-man rush chances, power plays against, and close-in shots have a strong impact on the likelihood of a shutout. These problems are of course common to all goalie stats, so I will just point them out as a caveat.

The main issue with shutouts is that very few games end up 1-0, so most of the time the goalie could have given up another goal without changing the game in any way. For example, Ryan Miller of Buffalo has been infamous this season for conceding late, meaningless goals with his team well in front, costing him the shutout but not impacting the result in any way.

How much does a goalie lose by giving up 1 goal? I took Brodeur, Luongo and Kiprusoff, considered by many to be the best goalies in the game today, and broke down their performances based on the number of goals allowed in each game. The results indicated that the difference between a shutout and a one-goal game is slight. Brodeur won 1-0 twice, Luongo and Kiprusoff once each. That means in 16 of the combined 20 shutouts they recorded, they could have given up another goal in the last minute of play (or "pulled a Miller") without changing the game's result.

Brodeur's winning percentage when giving up 1 goal was 91%, Luongo's was 90%, and Kiprusoff's was 95%. These are all close to 100%, and indicate how unfair it is that shutouts are counted and celebrated, while solid, one-goal-against performances are overlooked and forgotten. The winning percentages were similarly high with 2 goals against, Brodeur at 81%, Luongo at 82%, and Kiprusoff at 81%. The break point appears to be with the third goal, as all the goalies had a losing record when giving up 3 goals or more. Kiprusoff was particularly pronounced: when giving up 3 goals he was 4-12-0 for a .250 winning percentage. Luongo was at .423, Brodeur .375. All of them were very unlikely to win when giving up more than that.

This indicates that rather than crediting goalies only if they get a shutout, a better alternative would be to count how many times they give up 2 goals or less. The results indicate that goalies who give up 2 goals or less will win 85-90% of the time. I will check this stat for the entire NHL, but the assumption seems valid, especially since all three of the goalies sampled play for relatively low-scoring teams.

Therefore, rather than awarding 1 point for a shutout and 0 for everything else, I propose creating a stat called a "quality game" (inspired by baseball's quality start statistic), which refers to any game where the goalie played the entire game and allowed 2 goals or less. Of the three, Martin Brodeur ranks first in quality games with 35, Luongo second with 32, and Kiprusoff third with 29. All three had about a 90% winning percentage when they played a quality game. I feel this stat gives a better reflection of their relative levels of play than using shutouts alone (Brodeur 11, Kiprusoff 6, Luongo 3).

I plan on doing more research with the quality games stat, to find out which goalies consistently give up 1 or 2 goals per game, which goalies are often bailed out by their teammates, which goalies are having bad luck in close games, and which goalies have a few abnormally good or bad performances that are skewing their overall records.

Tuesday, February 13, 2007

Whither Consistency?

David Johnson over at Hockey Analysis is getting all giddy with Raycroft's recent performance. Because of 7 good games in a row, he appears to think that Raycroft has turned the corner:

"So, if this play by Raycroft become the norm and not just an anomoly (sic) then one has to ask if the rest of the Leafs have what it takes to be Stanley Cup contenders."

Suffice it to say that his arguments in the rest of the post can only be described as extremely optimistic, and contain hypotheticals that could apply to almost any team in the league. However, we are concerned here with the assertion that a recent hot streak is evidence that a goalie has elevated his play, and that he can be expected to play at a high level for the rest of the year.

Every goalie has hot streaks during the season. It is the nature of the position, where the difference between a mediocre .900 save percentage and an excellent .920 is about half a save a game. A couple weeks of lucky bounces, improved team defensive play and natural statistical clustering can make any goalie temporarily into a superstar. Almost every starting goalie in the league will have put together a hot streak much better than his actual record.

Want proof? I grabbed two random bad goalies, Tim Thomas and Jose Theodore. Thomas is admittedly more streaky than most because of his ridiculous throwback style, but despite his mediocre seasonal stats (.902, 3.24, 22-17-3), he put together a pretty good 9 game run in there (.933, 2.13, 7-2-0). Jose Theodore's been even worse (.892, 3.30, 10-12-1), but he had a 9 game run that was pretty decent as well (.924, 2.42, 5-3-0).

Raycroft's best 9 games? .931, 2.06, 7-2-0. Almost identical to Thomas. I just had a thought: If the Bruins could just get this kind of consistency in goal for the rest of the year, maybe they could challenge for the Cup...

I don't think we need to read anything into this at all. Raycroft's probably the same goalie he's ever been over 46 games this year, which is a much better sample size for analysis than a couple of weeks. The Leafs have been playing better as of late, and that has definitely had an impact on his play as well. Every NHL starter has the capability of putting together a few hot games in a row; they are, after all, among the 30 or 40 best goalies in the world. It is easy to say that if only a goaltender could show some consistency, he would really be good. That is precisely the difference between the best goalies and the average ones: They put together consistently excellent performances. In this season alone, Roberto Luongo has put together 3 9-game streaks comparable to what Raycroft has done, which is a clear indication that he is playing at his true level of ability, rather than just getting lucky.

In summary, don't extrapolate hot streaks, and don't expect something that a goalie is never going to give you. All goalies have good games and bad games, and all goalies put together hot streaks. Good goalies have good games more consistently, that's why they are good. Bad goalies don't just suddenly become consistently good (unless the teams in front of them are excellent, but even then they are not really "good", they just appear to be so). And all the evidence still suggests that Raycroft has quite a way to go before he can even be considered "good", much less Stanley Cup caliber.

Thursday, February 8, 2007

How Trevor Kidd and Felix Potvin Narrowly Missed the Hall of Fame - Part 2

Read Part 1 on the careers of Kidd, Brodeur and Potvin here.

In the 1994-95 season, all three goalies posted respectable win totals on decent teams. Trevor Kidd finished 8th in the NHL with a .909 save percentage, Felix Potvin one spot behind at .907. Playing again behind the Devils’ stingy defence, Martin Brodeur faced his career low in shots against – a mere 22.7 shots against per game. However he only stopped 90.2% of his shots, barely better than his backup Terreri and well behind Kidd and Potvin. This was good enough for 15th in the league. But that was quickly forgotten in the playoffs, as the Devils shifted their defensive game into overdrive and went on an amazing run. In 20 playoff games, Brodeur had a .932 save percentage as the New Jersey Devils won the Stanley Cup.

This was the break point, where Brodeur’s career took off into perceived superstardom (Hockey Stars Presents even ranked him as the #1 netminder in pro hockey in November 1995, despite Dominik Hasek’s two consecutive years of dominating the NHL with .930 save percentage). However, let us compare the numbers of all three of them in their careers up to this point in time:

Career statistics (after the 1994-95 season):

Trevor Kidd: 76 games, 2.86 GAA, 36-22-12, .899 save %, 26.0 shots/gm, .895 teammate save %

Martin Brodeur: 91 games, 2.45 GAA, 48-23-14, .909 save %, 24.5 shots/gm, .905 teammate save %

Felix Potvin: 154 games, 2.76 GAA, 74-52-24, .909 save %, 29.5 shots/gm, .901 teammate save %

Brodeur had the best traditional stats, but he faced the least shots. He also likely faced the easiest since Mike Vernon and Damian Rhodes were both better than Chris Terreri, yet the Devils’ backups registered a much higher save percentage than the other goalies on the Leafs or Flames. Despite this, Potvin was level with Brodeur in save percentage. Kidd was lagging a little behind the other two, but his team appears to have had the weakest defence and he was coming off of a strong 1994-95 season. It is not at all unreasonable to think that Kidd or Potvin could have done what Brodeur did or better, if they only had been given the opportunity.

The major difference between Kidd and the others was playoff performance, as Brodeur and Potvin both had sparkling starts to their playoff careers:

Career playoff stats after the 1994-95 season:

Kidd: 434 minutes, 3-4, 1 shutout, 3.59 GAA, .856 save%, 25.0 shots/gm

Brodeur: 2425 minutes, 24-14, 4 shutouts, 1.86 GAA, .931 save%, 26.7 shots/gm

Potvin: 2856 minutes, 23-23, 5 shutouts, 2.69 GAA, .909 save%, 29.6 shots/gm

At this point, the careers of Potvin and Brodeur began to diverge. The reason for this was not the relative play of the two goaltenders. In fact, Potvin arguably continued to outperform Brodeur for the next 3 seasons. The reason was the deterioration of the Toronto Maple Leafs and the continued defensive excellence of the New Jersey Devils.

Over the next 3 seasons, the team effects grew greater and greater. Kidd played only 2 playoff games in that span, and Potvin managed only 6. The stats for the period (1995-96 to 1997-98) show Potvin’s heroics under fire, as well as Kidd also putting up decent numbers on a poor team:

1995-96 to 1997-98:

Trevor Kidd: 149 games, 2.60 GAA, 57-65-17, 10 shutouts, .908 save%, 26.0 shots/gm, .898 teammate %

Martin Brodeur: 214 games, 2.05 GAA, 114-61-33, 26 shutouts, .920 save%, 24.7 shots/gm, .904 teammate save %

Felix Potvin: 210 games, 2.92 GAA, 83-95-25, 7 shutouts, .911 save%, 31.6 shots/gm, .891 teammate save %

By 1999, Felix Potvin had been traded from Toronto to the New York Islanders. He would bounce around to the Canucks, Kings, and eventually Bruins, before retiring after the 2004 season. He had a few flashes of brilliance left, particularly in L.A., where he posted a .919 save percentage in 23 games in 2000-01, and played in 71 games with a .907 save percentage in 2001-02. However, this was mostly because of the team effect. Felix faced just 26 shots per game over his L.A. career, and his backup, Jamie Storr, had excellent numbers, including a .922 save percentage in 2001-02. Only at the end of his career did Potvin finally experience the pleasure of playing behind a quality defence. His prime years had been wasted by the Toronto Maple Leafs.

In 1997-98, Trevor Kidd’s career peaked, and he had probably one the most overlooked seasons in recent memory. Playing in 47 games for the Carolina Hurricanes, he posted a .922 save percentage, the 2nd best in the NHL behind only Dominik Hasek. He was so good that Carolina traded his backup, Sean Burke (who was a starter in nearly every other season in his career) to Vancouver. For the season, Kidd’s backups, mostly NHL veterans like Burke, had a mere .882 save percentage, indicating just how poor Carolina’s defence was.

Unfortunately for Kidd, Arturs Irbe landed in Carolina the next season, and he went on a season-long hot streak to win the starting job. Trevor Kidd moved on to Florida, but was unlucky enough to get there around the same time as the emergence of Roberto Luongo. As a backup, Kidd posted three straight years of save percentages around .895 behind an awful defence. He was probably playing at a league-average level or better those years, as he had been in the years leading up to them. Luongo was at .915 or better all three seasons, which is evidence of his largely unrecognized brilliance. Kidd had one final NHL stop in Toronto, and today is playing in Europe.

Martin Brodeur is still in the NHL. In the intervening years, he posted relatively unimpressive save percentages, considering New Jersey’s continued defensive prowess. After winning the Cup in 1995, the Devils and Brodeur had a string of postseason disappointments in the late '90's, before winning again in 2000. Brodeur got his name on the Cup one more time in 2003, and was a member of the Olympic-gold-medal-winning Team Canada of 2002. This year, he is having perhaps his best season, and is considered to be the best goalie in the league by the majority of hockey pundits, general managers, and players.

Looking at their career paths and their numbers, it is quite plausible that either Trevor Kidd or Felix Potvin, if they had been selected by the New Jersey Devils in 1990, would have had a Hall of Fame career. They both had more distinguished junior careers than Brodeur, and both of them had the physical abilities to be a top-level NHL goaltender. Both of them were above average goalies in the NHL for an extended period of time. And in contrast to the often weak teams they played on, New Jersey’s strong defensive play created the perfect environment for its goalies to put up outstanding numbers and experience regular season and postseason success. From 1994 to 2004, the Devils won 469 games, scored an average of 2.9 goals per game, and consistently ranked among the top defensive teams in the league. Some of this success is attributable to goaltending, but certainly not the majority of it. A decent NHL starting goalie such as Kidd or Potvin would probably have posted over 300 wins and 50 shutouts in a decade of play in East Rutherford, and there’s a good chance they would have won at least one Stanley Cup.

Potvin retired with 266 wins and 36 shutouts, Kidd with 140 wins and 19 shutouts. Neither won a Stanley Cup, or even played in a Cup final. Since they never had the required team success, neither will be considered for the Hall of Fame. But if the Calgary or New Jersey scouting staffs had differed slightly in their prospect rankings 17 years ago, Brodeur may very well have been sitting on the sidelines today, watching the hockey world celebrate Felix Potvin or Trevor Kidd as one of the best goalies in the world.

Wednesday, February 7, 2007

How Trevor Kidd and Felix Potvin Narrowly Missed the Hall of Fame

Part 1:

In the 1990 NHL entry draft, two teams were looking to draft their goalie of the future. Both the New Jersey Devils and Calgary Flames had carefully scrutinized Trevor Kidd, Martin Brodeur, and Felix Potvin. Their scouting staffs had watched them play and sifted through their statistics, trying to figure out who would be the best at the NHL level. New Jersey sat with the 11th overall pick, Calgary at #20.

On Calgary’s draft board, the scouting staff had Trevor Kidd, the reigning Canadian Major Junior Goaltender of the Year, as the top goaltending prospect. Wanting Kidd, and fearing another team would grab him first, Calgary proposed a swap of first round picks with New Jersey, throwing in a couple of later round choices into the deal.

Lamoriello had sent his scouting staff to look into Kidd and Brodeur as well. The reports he got back from his scouts indicated that Brodeur was the better bet. Gambling that none of the teams in between them would draft a goaltender, Lamoriello agreed to the Calgary trade proposal. Kidd went 11th, and Brodeur fell to the Devils at #20. Potvin went third, taken 31st overall by the Toronto Maple Leafs.

Usually in draft retrospectives, sportswriters look at what each player accomplished over their career and give good grades to teams that drafted players who turned out to be good. By that logic, Calgary blundered horribly by trading up to get Kidd, while New Jersey got a steal in Brodeur. I’m going to approach it from a different angle – my argument is that the three goalies were essentially fungible. Their junior and early pro careers were comparable, and it was primarily the team situation that determined their successes and failures. It is not unreasonable that either Kidd or Potvin, if they had gone to New Jersey, could have emulated Brodeur’s success and even potentially had a Hall of Fame career.

Coming out of juniors, all three goalies were top prospects. They were all very athletic, although with slight variations in style: Kidd relied more on his size, Potvin favoured the butterfly, while Brodeur had a hybrid style. Potvin and Brodeur played in the QMJHL, while Kidd played in the WHL. As 18 and 19 year olds, all 3 had goals against averages of around 4 on middle-of-the-road teams, but were impressive enough to be drafted highly by NHL teams.

In 1990-91, Brodeur had a decent season (.886 save percentage in the high-scoring QMJHL) for St. Hyacinthe. Over in Chicoutimi, Potvin had an outstanding season. He registered a .910 save percentage and a 2.70 GAA for Chicoutimi, which finished 1st overall in the league. The team was a powerhouse, and Potvin likely faced easier shots than average, but his save percentage was very high for his league and his stats were much better than the other goalies on the team. Kidd was traded from Brandon to Spokane, and having finally been given the opportunity to play behind a strong team he helped the Chiefs win the 1991 Memorial Cup.

The next season Brodeur was still in junior. Trevor Kidd spent the year touring with the Canadian National Team. Potvin made the jump to the AHL. He posted an excellent .908 save percentage, much better than his playing partner and future NHLer Damian Rhodes’ .889. All three goalies were also called up during the season to experience a few games at the NHL level.

For the 1992-93 season, the Leafs were ready to give Felix Potvin a serious look. Grant Fuhr was the starter, but he was aging and coming off a poor 1991-92 season. Over the course of the season, Potvin took Fuhr’s job, posting a .910 save percentage, which was the 2nd best in the NHL. Potvin was a Calder Trophy finalist, and played in 21 playoff games as the Leafs went to the Western Conference Finals. With that kind of performance, and playing in the large media market of Toronto, Potvin was quickly becoming an NHL star.

Meanwhile Brodeur and Kidd were both trying to make their mark in the minor leagues. Brodeur played in the AHL with the Utica Devils, alongside fellow 1990 draftee Corey Schwab. The two goalies split the games, and had nearly identical statistics (.884 save % for Brodeur, .883 for Schwab). Trevor Kidd played for the IHL’s Salt Lake Golden Eagles, posting an .887 save percentage while backing up Andrei Trefilov.

Heading into the 1993-94 season, Potvin was recognized as one of the top goalies in the game. He was the undisputed #1 goalie on the Leafs, one publication ranked him as the #3 goalie in hockey, and he would be chosen to play in the mid-season All-Star Game. Kidd and Brodeur both caught on as their team’s backup, Kidd behind Mike Vernon and Brodeur behind Chris Terreri.

To that point in his career, Terreri had been a mediocre goaltender. His career best save percentage was .893, and at the age of 30 there wasn’t much reason to expect a major improvement. The Devils had finished 4th in the division for three years in a row, each time losing in the first playoff round. But something significant happened in 1993-94. The team hired Jacques Lemaire as its new coach, and he implemented a new defensive strategy known as the neutral zone trap. The effects were immediate: Terreri’s save percentage soared to .907. Brodeur was even better, posting a .915 save percentage as the goalies split the games. The Devils took second place in the Atlantic Division with over 100 points. Brodeur became the starter for the playoffs, where the Devils went to overtime in game 7 of the Conference Finals before losing to the Rangers. Brodeur was awarded the Calder Trophy as Rookie of the Year. He had truly arrived in the NHL a most opportune time.

Trevor Kidd’s situation was more difficult. The Flames’ starting goalie was Stanley Cup winner Mike Vernon. Kidd was initially unable to unseat Vernon, although he managed to get into 31 games and post statistics that were comparable to the veteran’s. It was enough to convince Flames management that Kidd was their goalie of the future. At the end of the year Calgary traded Vernon to Detroit, and the starting job was handed over to Kidd.

By 1994, all three goalies had starting jobs on their teams, and had proven themselves at the NHL level. Potvin and Brodeur were on noticeably similar career paths – they broke in with excellent rookie seasons, and experienced initial playoff success. But 1995 was to provide the springboard for one of the goalies from promising to elite. As so often happens, it would not be the goalie with the best performance – it would be the goalie with the best teammates.

Read part 2 here.

Why Everyone Loves Martin Brodeur

I have a friend who plays hockey and is a hockey fan. He knows most of the good players in the league. He told me recently that his favourite goalie was Martin Brodeur.

For interest’s sake, I began to question why, and he said that he just enjoyed watching him play. He watches hockey occasionally, but isn’t a hardcore fan by any means, so I asked when in particular that was. He responded, “In the Olympics.”

I suspect there are many Canadian hockey fans that have a similar perspective. After all, who wants to watch the New Jersey Devils, unless one’s goal is to fall asleep? Even now, I would guess most fans see Brodeur a couple of times when he is playing against their favourite team, and the rest of the time only on Sportscenter. Brodeur’s gold-medal winning performance in 2002, probably more than anything else, including his Stanley Cups, is responsible for defining him and creating the perception of him as the best goaltender in the world. If Curtis Joseph had backstopped the team to gold five years ago (or for that matter, if Patrick Roy had agreed to play in the first place), I doubt that there would be this level of unanimity in terms of the best goalie in the game today, even given Brodeur’s current season and his dominance of the traditional team-dependent goaltending statistics.

In a coming series, I will revisit Salt Lake City and show why Martin Brodeur was very fortunate to be given that opportunity, and how it became so career-defining.

Sunday, February 4, 2007

Comparisons Revisited

One could of course make the argument that the comparisons from my last post are somewhat dated. Of course they are, but they also represent Brodeur’s supposedly best years, in which he received the greatest amount of recognition in the form of Vezinas and MVP candidacy.

In the interests of fairness, here are the comparables for last year (courtesy Hockey Analytics) and this year (from Hockey Numbers):

Last year:
Jason LaBarbera, Kari Lehtonen, Dwayne Roloson, Alex Auld, Rick DiPietro

This year:
Marty Turco, Kari Lehtonen, Nikolai Khabibulin, Henrik Lundqvist, David Aebischer

And since I did it before, here's Roberto Luongo as well:

2005-06: Manny Fernandez, Tomas Vokoun, J.S. Giguere
2006-07: Chris Mason, Cristobal Huet, Miikka Kiprusoff

(Note: I'm not cherry picking for Luongo, his performance is at the top of the league so there are only a few goalies that are comparable).

Brodeur seems to be moving up the ranks a little, and has been unquestionably excellent this year, but he is still no Luongo.

Friday, February 2, 2007

The "Incomparable" Martin Brodeur

Who is the most comparable goalie in the league to Martin Brodeur?

Hasek might seem like a reasonable choice because he has been outstanding for many years and his NHL career is roughly contemporary with Marty’s. Or maybe one could go back a few years to select Roy, the man who holds many of the records that Brodeur is currently pursuing. Others might prefer Belfour or Khabibulin or another one of the more recent Cup-winning goalies who are “winners” or “money goalies” or have some similar arbitrarily defined reputation. Some might pick Kiprusoff or Luongo because they are currently among the very best goalies in the league. Fawning analysts like TSN’s Glenn Healy might go so far as to claim that the very notion of comparing any current goalies with the legendary Martin Brodeur is grossly insulting.

Brodeur won back to back Vezinas in 2003 and 2004, and even scored a Hart Trophy nomination in 2004 as one of the three most valuable players in the entire league. Clearly the NHL general managers voting on the award were were thinking of Brodeur as a peer of Hasek and Roy (the only other two goalies to win back-to-back Vezinas since it changed to being voted on in 1982), both of whom are generally agreed upon to be among the best goalies to ever play the game.

The stats, however, tell a much different story, once you stop giving credit to Brodeur that should instead be directed to Messrs. Stevens, Niedermeyer, Madden, et al., through the use of shot-quality neutral save percentage.

Brodeur's closest comparables in shot-quality neutral save percentage in 2004 were these guys:

Brian Boucher, Dan Cloutier, Robert Esche, Martin Biron, Olaf Kolzig, and Rick DiPietro.

And in 2003:

Jose Theodore, Dan Cloutier, Marc Denis, Patrick Lalime, Nikolai Khabibulin, and Martin Biron.

Maybe that is why trade rumours are always swirling around Martin Biron – he’s as good as Martin Brodeur! NHL GMs take note, before Darcy Regier figures this out and the price goes up! (OK, not really, but that is quite an uninspiring group of goalies).

For those who might have been wondering, Luongo’s closest comparables were Patrick Roy in 2003 and Miikka Kiprusoff in 2004.

I’m just starting to begin to consider that maybe the voting went to the wrong guy. But then I remembered that Brodeur has won 3 Cups and an Olympic gold medal and my fears melted away. He must be the best.

Thursday, February 1, 2007

Brodeur in 2006-07

This year, Martin Brodeur is having probably the best year of his career. He is leading all of the traditional goaltending categories (GAA, wins, shutouts), and he is, most unusually for him, ranked near the top of the league in save percentage.

What gives? Is this the counterargument that defeats all the Brodeur-bashers? Does this show that Brodeur is really good after all? Should we start anointing him as the best ever and waive the Hall-of-Fame entry period?

The answer that Brodeur is certainly having an excellent year by his standards. However, he is still playing for the New Jersey Devils, who despite a fair amount of roster turnover and all the NHL’s rule changes still play the same patient, disciplined style that is very effective at limiting chances. Brodeur is still not the best goalie in the league. If I dare say it, he may not even be in the top 5.

The New Jersey Devils are good at preventing shots, ranked 8th in the league in shots against with 28.8 per game. They are outstanding at preventing dangerous shots. Brodeur’s shots are the easiest shots faced by any goalie. Hockey Numbers has the shot quality against for Brodeur at 0.79, meaning that the opposition's scoring chances are 21% less dangerous than average. A middle-of-the-road NHL starting goalie playing for the Devils would therefore be expected to have a save percentage of around .925, which makes Brodeur’s .928 look much less impressive.

One of the main reasons why it is so relatively easy to play goalie for the Devils is their amazing team discipline. They are ranked 7th in penalty kill efficiency, and are probably about the 10th best unit when you consider shorthanded goals as well. However, they have been shorthanded a mere 175 times to lead the NHL by a large margin, and as a result have conceded 6 fewer shorthanded goals than any other team in the league.

To put this into perspective, the Montreal Canadiens have an excellent penalty kill and are ranked 3rd in efficiency. They have also scored 13 shorthanded goals to lead the NHL. They have a net minus on the penalty kill of 25 goals, which is exactly the same as New Jersey’s. The difference is the number of opportunities.

Brodeur is pretty good on the penalty kill in terms of save percentage, with a .921 shot-quality adjusted shorthanded save percentage (8th in the NHL). But in today’s NHL, where the power play is increasingly important and the best scoring chances often come with the man advantage, facing the fewest number of them is a tremendous advantage.

An average goalie (like, say, Marc-Andre Fleury) playing Brodeur’s minutes for New Jersey and facing the same shots would be expected to post a .925 save percentage, a 2.20 goals against average, and about 5 shutouts. Given that the Devils score 2.6 goals per game they would be expected to have a winning record (hockey’s Pythagorean equation suggests about 60 points, which still would put them in first place in the Atlantic Division). Based on the average league rate of wins to points, that’s about 27 wins.

In summary compared to Brodeur:

Brodeur: .928 save %, 2.03 GAA, 30 wins, 9 shutouts
Average goalie: .925 save %, 2.20 GAA, 26 wins, 5 shutouts

Based on shot-quality neutral measures of efficiency, Brodeur is the 10th best goalie in the league. Of course he plays just about every game, so he should be rewarded for his extra effort. That puts him probably around number 5 or 6, behind Luongo, Kiprusoff, Mason, Kolzig, and maybe Huet. Giguere would have been ahead as well, except for his injury.

None of this changes the fact that Brodeur is going to win the Vezina unanimously and get nominated for the Hart Trophy, and he could very well win that as well. Let’s hope he thanks the Devils in his speech, though, as they are once again doing the heavy lifting behind the scenes.