Friday, August 26, 2005
The Sabermetric Challenge of Ray Bourque
Ray Bourque was a very good defender. He is the highest scoring defender of all time. He has made the First Team All Star 13 times in his career (which is more times than any other defender made the First and Second Teams put together). Bourque also adds 6 more Second Team selections. From those numbers alone, one might conclude that Bourque is the best defender of all time. However, it is a universal (or almost universal) belief that Bobby Orr is the best defender of all time.
In hockey sabermetrics systems it is a common problem to rate the top defencemen of all time. I compared two sabermetric systems, Daryl Shilling's Hockey Project Monitor and Pnep's Hall of Fame Monitor. Daryl Shilling concludes that Bobby Orr is the best defender of all time, while Pnep concludes Ray Bourque.
With all of Bourque's acheivements, he never reached the extreme dominance over the league that Bobby Orr did. Orr won the Norris Trophy as best defenceman 8 years in a row. He was the top scorer in the NHL twice. He was the NHL MVP twice. He was the Playoff MVP twice (both times he won the cup). By comparison, Bourque won 5 Norris Trophies over a period of 8 years. He never seriously contended for the NHL scoring title. He never won any MVP's in the regular season or playoffs, though he was twice Hart Trophy runner up. Bourque's career lasted much longer than Orr's (in fact he played almost 1000 more career games 1612 vs. 657), so this allows Bourque to exceed Orr in career numbers. How does one reach the sabermetric conclusion that Orr is ahead of Bourque given Bourque's much better career numbers? Shilling does this by considering dominance in his Hockey Project Monitor. Pnep does not consider dominance. This shows the importance of dominance of a player being considered as well as his career acheivements, instead of only looking at career acheievments.
There are other defenders that I would argue were significantly more dominant than Bourque, but played in earlier times. They are Doug Harvey who won 7 Norris Trophies in 8 years in the 1950's and 60's and Eddie Shore who had a career in the 1920's and 1930's predating the existance of the Norris Trophy, but he won the Hart Trophy four times. Both of them played in eras where offense from defencemen is less common then it has become in modern times (though they were some of the higher scoring defenders of their eras), so they do not have significant career numbers. It is a tough sabermetric problem to design a system where Shilling ranks Shore sixth in his top 10 defenders and Harvey not at all. Pnep ranks Harvey third in his top ten and Shore not at all. I think that both of them should probably be rated above Bourque (with Bourque ranked as the 4th best defender of all time). Clearly, these two systems reach very different conclusions. In future posts, I will discuss why this is. The problem is that it is hard to show dominance statistically when that dominance is defensive and not offensive. It is one of the problems in hockey sabermetrics that has been a fatal flaw so far. I hope that someday, it gets solved, but until it is, any results of any hockey sabermetric system must be taken with a grain of salt.
In hockey sabermetrics systems it is a common problem to rate the top defencemen of all time. I compared two sabermetric systems, Daryl Shilling's Hockey Project Monitor and Pnep's Hall of Fame Monitor. Daryl Shilling concludes that Bobby Orr is the best defender of all time, while Pnep concludes Ray Bourque.
With all of Bourque's acheivements, he never reached the extreme dominance over the league that Bobby Orr did. Orr won the Norris Trophy as best defenceman 8 years in a row. He was the top scorer in the NHL twice. He was the NHL MVP twice. He was the Playoff MVP twice (both times he won the cup). By comparison, Bourque won 5 Norris Trophies over a period of 8 years. He never seriously contended for the NHL scoring title. He never won any MVP's in the regular season or playoffs, though he was twice Hart Trophy runner up. Bourque's career lasted much longer than Orr's (in fact he played almost 1000 more career games 1612 vs. 657), so this allows Bourque to exceed Orr in career numbers. How does one reach the sabermetric conclusion that Orr is ahead of Bourque given Bourque's much better career numbers? Shilling does this by considering dominance in his Hockey Project Monitor. Pnep does not consider dominance. This shows the importance of dominance of a player being considered as well as his career acheivements, instead of only looking at career acheievments.
There are other defenders that I would argue were significantly more dominant than Bourque, but played in earlier times. They are Doug Harvey who won 7 Norris Trophies in 8 years in the 1950's and 60's and Eddie Shore who had a career in the 1920's and 1930's predating the existance of the Norris Trophy, but he won the Hart Trophy four times. Both of them played in eras where offense from defencemen is less common then it has become in modern times (though they were some of the higher scoring defenders of their eras), so they do not have significant career numbers. It is a tough sabermetric problem to design a system where Shilling ranks Shore sixth in his top 10 defenders and Harvey not at all. Pnep ranks Harvey third in his top ten and Shore not at all. I think that both of them should probably be rated above Bourque (with Bourque ranked as the 4th best defender of all time). Clearly, these two systems reach very different conclusions. In future posts, I will discuss why this is. The problem is that it is hard to show dominance statistically when that dominance is defensive and not offensive. It is one of the problems in hockey sabermetrics that has been a fatal flaw so far. I hope that someday, it gets solved, but until it is, any results of any hockey sabermetric system must be taken with a grain of salt.
Comments:
<< Home
Very nice blog! I like the work you're doing on hockey and sabermetrics.
I think it's absolutely crucial to distinguish between Career Value and Peak Value (I'm borrowing these terms from Bill James) when discussing which players were the best.
In terms of Career Value (ie playing well over a long period of time), Bourque was probably the best defenseman ever. Who else has been a top five defenseman every year, for 22 seasons? (Larry Robinson, Doug Harvey, Tim Horton and others rank high in this regard, but not as high as Bourque). It's a simple fact that Bourque was more productive over the course of his career than Orr. Orr played 9 seasons; Bourque provided 13 _additional_ years of Norris-calibre hockey, and outscored him by nearly 700 points.
In terms of Peak Value, Bobby Orr is by far the best defenseman of all-time. Nobody in his prime was close to Orr's level of dominace. I'd probably rank Bourque behind Shore and Harvey (for sure) and perhaps Potvin and a few others in Peak Value.
(Another good example: Mike Gartner has higher Career Value than Mike Bossy, but who would you take in their prime?)
In summary, it's reasonable to say that Orr was far more dominant than Bourque, but Bourque had more longevity and consistency. I think the problem arises when one tries to arbitrarily combine these two different measures of player greatness into a single formula--it may not be possible to weigh Career and Peak fairly.
Post a Comment
I think it's absolutely crucial to distinguish between Career Value and Peak Value (I'm borrowing these terms from Bill James) when discussing which players were the best.
In terms of Career Value (ie playing well over a long period of time), Bourque was probably the best defenseman ever. Who else has been a top five defenseman every year, for 22 seasons? (Larry Robinson, Doug Harvey, Tim Horton and others rank high in this regard, but not as high as Bourque). It's a simple fact that Bourque was more productive over the course of his career than Orr. Orr played 9 seasons; Bourque provided 13 _additional_ years of Norris-calibre hockey, and outscored him by nearly 700 points.
In terms of Peak Value, Bobby Orr is by far the best defenseman of all-time. Nobody in his prime was close to Orr's level of dominace. I'd probably rank Bourque behind Shore and Harvey (for sure) and perhaps Potvin and a few others in Peak Value.
(Another good example: Mike Gartner has higher Career Value than Mike Bossy, but who would you take in their prime?)
In summary, it's reasonable to say that Orr was far more dominant than Bourque, but Bourque had more longevity and consistency. I think the problem arises when one tries to arbitrarily combine these two different measures of player greatness into a single formula--it may not be possible to weigh Career and Peak fairly.
<< Home
