Sunday, July 30, 2006

The Sabremetric Problem of Terry Sawchuk

One sabremetrics and hockey problem is ranking the best goalies of all time. The method that I have shown is Pnep's (Roman Nepomnyaschev) Hall of Fame monitor. He has taken it through various incarnations including a more simplified version reported by Daryl Shilling. One constant is that Terry Sawchuk is rated in the mid to lower half of the top 10.

Terry Sawchuk is one of the greatest goaltenders of all time. In the 1980's and early 1990's before the emergence of Patrick Roy and Dominik Hasek, he was often rated as the best goalie of all time. He led the NHL in career shutouts and career wins (though Roy has beaten that record).

Terry Sawchuk was born and raised in the Winnipeg area. As a kid in the Winnipeg minor system he was signed into the Detroit system. The Red Wings moved Sawchuk to Ontario where he played for the Galt Red Wings and Windsor Spitfires in the Ontario Junior League and then moved into minor pro leagues with Windsor in the IHL, Omaha in the USHL and Indianapolis in the AHL before getting a chance when Harry Lumley, the incumbent Red Wing goalie, got hurt.

Sawchuk played seven games, but he played well enough that Lumley was traded to Chicago. Sawchuk responded with a Calder trophy and a first team all star selection putting up a 1.99 GAA. In his second year, Sawchuk won the Vezina trophy (which was the top GAA at that time) and again made first team all star. This year, he lead the Red Wings to the Stanley Cup with eight straight wins posting a 0.63 GAA. Sawchuk followed that up in 1953 with a third straight first all star nomination and a second Vezina win. In his next two years, Sawchuk and the Wings won the Stanley Cup two more times. Sawchuk won another Vezina trophy and made second team all star both years. Sawchuk was a high strung nervous goalie and scared Red Wing management, in that he didn't seem capable a mentally remaining at the NHL level, so in a blockbuster deal they traded hm to Boston. He was traded with Marcel Bonin, Lorne Davis and Vic Stasiuk for Gilles Boisvert, Real Chevrefils, Norm Corcoran, Warren Godfrey and Ed Sanford. Boston got the best player getting Sawchuk and Detroit getting a package of mostly prospects who did not work out.

Sawchuk never really adapted to Boston. He played there for a year and a half before temporarily retiring from the NHL due to exhaustion. He was convinced to return to the NHL when he was traded back to Detroit. He was traded for Johnny Bucyk and cash. This was a trade where both teams got star players.

Back in Detroit, Sawchuk played several solid seasons where he twice more made the second team all star. Since Montreal was the dominant team, their goaltender Jacques Plante put up the best numbers and won the biggest awards. Sawchuk played very well on a not so powerful Wings team. In a 1964 Intra-League draft (which is an early waiver draft), Sawchuk was selected by the Toronto Maple Leafs.

In Toronto, Sawchuk shared time in goal with Johnny Bower. He won another Vezina trophy in 1965 and another Stanley Cup in 1967. In the 1967 expansion draft, the agin Sawchuk was left exposed (after just having won the Stanley Cup). He was selected by Los Angeles.

After a year in LA, Sawchuk was traded back to Detroit for Jimmy Peters. He played only thirteen games there before being traded the next year to the New York Rangers with Sandy Snow for Larry Jeffrey.

In New York, Sawchuk played only eight games. Late in the year, he had a fight with teammate Ron Stewart. In the fight, Sawchuk suffered internal injuries including a lacerated liver and blood clots. Sawchuk later died when one of the blood clots netered his heart. The exact circumstances of his death were never clearly made available to the public.

Terry Sawchuk retired with a record 103 shutouts. He retired with 447 wins which was a record at the time. His prime years (his first 5 full years in Detroit) where he never had a GAA above 2.00 and he won three Vezina trophies are arguably the best prime any goalie had ever had (at least until Dominik Hasek won back to back Hart Trophies). Sawchuk has a very good claim as a top goalie. Probably the top all time before Roy and Hasek.

His detractors might argue that he was highstrung and this took away from his value (though it shouldn't show up in a sabremetric analysis). The other top goalies of the time Jacques Plante and Glenn Hall were also very high strung and the more modern top goalies Patrick Roy and Dominik Hasek are too. It is hard to argue that it hurt one goalie more than any other.

Jacques Plante gets rated ahead of Sawchuk in this system because, despite a slightly shorter career, he won 6 Vezinas in 7 years. The Montreal defence was the best in the NHL and gave him a good chance at the best GAA (which won the trophy). The reliance upon award victories to gather points in the hall of fame monitor system gives Plante the edge for his extra awards. Glenn Hall also played a slightly shorter career and was roughly a contemporary of Sawchuk. He had seven first team all star selections which garner him more HOF monitor points. It is believed that after Sawchuk broke down in Boston, he was downgraded in the minds of several voters and they refused to vote for him for awards until he rehabilitated his image in Toronto many years later. Sawchuk was not a favorite of many of the voters. Their bias kept him from winning awards and gave Glenn Hall the chance to win them instead. This is the problem with a Hall of Fame monitor that rewards players for being voted an award winner. It is dependant upon the voters not making mistakes. In the case of Sawchuk, they shortchanged him.

Sawchuk is a goalie consistently rated in the top 10 all time, but I think he should be ranked higher than these systems show. He has a more dominant prime and better career numbers than contemporaries ranked above him.

An ideal player ranking system would not use award wins to give points. It would give points based on a player's contribution to his team. The player with the most points in the system should then be chosen as the award winner. The system could be used to predict award winners, instead of award wins being used as an input. Such a system would require a better understanding of exactly how much a goaltender contributes ot the wins and losses of his team. There is no easy statistical way to show this (wins, GAA, saves percentage, shutouts don't clearly show it). Better goaltender statistics are needed. Probably we need to better know the circumstances of shots. How far are they from the goal? How many people are in front of the net? How fast was the shot? Was it a rebound? Etc. Without that it might be hard to come up with a direct value of a goalie. We need to know the quantity (which is easy to show statistically) and quality (which is hard to determine) of shots a goalie faces. Relying upon voters means relying on other (sometimes biased) opinions. These biases keep Terry Sawchuk down. He probably should have made more all star teams if the voters liked him better.

Comments: - new "HHOF Monitor"
Great, great stuff. I've done some looking into his career and it's amazing how famous he was even deep into his career when clearly he wasn't at his best.

The first player chosen in the 1967 Expansion draft, even though he was long in the tooth. The headlines in several papers I've read (June 7, 1967, several articles in the following week in most Canadian dailies) have him as the major story. Of course he'd just won the Stanley Cup with Toronto but there were many interesting stories to come out of the expansion draft (the Boston Bruins were so angry it boiled over for quite some time).

Sawchuk was imo a top level goalie even during the early 60s Detroit years (his SP in 63-64 was .917 and the Wings finished only 4th).

I'm very interested in reading the rest of your articles on this subject.
I know pnep's system, like any HoF monitor, is about ranking players based on certain criteria that are seemingly important to HoF voters. It's not a pure ranking of talent, but rather how certain players score in certain areas. That's why a guy like Brodeur will be in, even though his true talent is far lesser than his HoF ranking. This is also why the HHOF voters have put in guys like Gilles...they look at certain things, like playoff success, very strongly.
I'm not sure I buy that pnep's HOF monitor can differentiate between those who are likely to go to the hall of fame and those who should go. I am aware of that argument and am trying to figure out how best to address it in a future post.

Your two examples address this point. Clarke Gillies does not score well in pnep's system. He doesn't belong any way you slice things - unless we start giving HOF monitor points for being friendly with HOF committee members I don't see any way to fix it.

I agree with you that Martin Brodeur is not as good as some believe he is. For example, he definitely didn't deserve a Vezina noimination this season (but as usual he got one). That said, he belongs in the hall of fame. He is a perennial all star goalie who has won several awards (though probably only deserved about half of them) who looks very likely to be the all time winningest goalie in hockey history when he retires. That is a hall of famer. He might be more of an Ed Belfour level hall of famer than a Patrick Roy level one, but he belongs there.

I think it is a mistake to build a "distorted" player ranking system like this HOF monitor is purported to be because there are stupid HOF inductees who are mostly selected due to the old boys network (ie Gillies, Duff) and deserving players who never fit into the old boys network who find it hard to get inducted (ie. Ciccarelli). We cannot realistically measure cronyism in any statistical manner. That is a harder thing to do than measure the value of a hockey player (and we cannot do that either). I think a HOF monitor type player ranking system can only attempt to measure how good a player was with any success. Stupid cronyism moves like inducting Clark Gillies and not Glenn Anderson (if winning Stanley Cups really is so important to induction) cannot be build into a HOF monitor and shouldn't even be attempted because it takes something valuable (attempting to rate players) and at the last minute runs it through a lens that is intended to blur the results with the hope that it is blurred the way the old boys network in the HOF committee might blur things.
AS time goes on and equipment gets better and better players like Brodeur and Roy will outscore those who came before them. the criteria for the Vezina has changed. It is now a popularity vote and some players like Terry would not get that vote because of his attitude towards writers. He never did win the MVP or the conn Smyth in 1967 when he clearly was the best.
Every weeκend i usеd to visit
this sitе, bеcausе i want еnjoуmеnt, since this thiѕ web page conatiоns truly good funny
datа too.
Feel free to visit my homepage
I was аble to find good information from your blog
my webpage: samsung galaxy note 2
An interesting discussion is definitely worth comment.
I do believe that you ought to write more about this topic, it may not be
a taboo matter but generally folks don't talk about such subjects. To the next! Kind regards!!

my web site; zetaclear nail fungus relief
It's wonderful that you are getting ideas from this article as well as from our dialogue made at this time.

Here is my webpage ... USA Viagra Online
I have to thank you for the efforts you've put in writing this blog
Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?