Tuesday, September 04, 2007

Terry Sawchuk From A Goals Saved Perspective

Terry Sawchuk is a Hall of Famer who is often listed among the greatest hockey players all time, however, when one looks at his career from a sabermetrics perspective this dominance is not so obvious. Sawchuk had a long career where he won three Vezina Trophies (at the time this award was for best GAA in the NHL) and played in eleven all star games. He retired from the NHL the career leader in wins and shutouts (he still holds the career shutout lead). That success in counting stats may have led some to overrate his career. He is second in career games played for a goalie (behind Patrick Roy) and his counting stats reflect his longevity.

It is hard to rate goaltenders sabermetrically, but for this analysis I will use the goals saved method developed by Peter Albert (the hockey outsider). This method attempts to calculate the number of extra goals a goaltender saves relative to the average goalie in a season if he played on the average NHL team. It is very successful picking winners of the Vezina Trophy, but not so good comparing different seasons (due to different performances of "average" goaltenders over time). Nevertheless, if Sawchuk is indeed on of the all time greats, one would expect he should be one of the leaders in goals saved over the seasons of his career and due to the length of his career one of the all time leaders.

Here is Terry Sawchuk's career in terms of goals saved:

Terry Sawchuk's Career in terms of Goals Saved
Season Sawchuk's Goals Saved Goals Saved League Leader League Leader's Total
1949/50Sufficent Stats Not Kept
1950/51 Sufficient Stats Not Kept
1951/52 38.4 Terry Sawchuk 38.4
1952/53 29.6 Terry Sawchuk 29.6
1953/54 22.2 Harry Lumley 34.6
1954/55 31.8 Harry Lumley 40.0
1955/56 -11.4 Jacques Plante 32.4
1956/57 11.2 Glenn Hall 45.8
1957/58 -9.3 Jacques Plante 29.3
1958/59 -18.0 Jacques Plante 53.8
1959/60 1.5 Johnny Bower 32.3
1960/61 -13.0 Glenn Hall 36.6
1961/62 -31.9 Jacques Plante 53.1
1962/63 9.6 Glenn Hall 21.1
1963/64 0.8 Glenn Hall 38.2
1964/65 8.4 Glenn Hall 22.4
1965/66 6.3 Johnny Bower 34.2
1966/67 11.7 Ed Giacomin 29.8
1967/68 -7.1 Gump Worsley 23.5
1968/69 3.5 Glenn Hall 18.4
1969/70-2.0 Tony Esposito 32.5

We note that sufficient statistics were not kept in the first two years of his career to calculate Sawchuk's goals saved. There are insufficient records to calculate saves percentages from that time. These records exist from 1951/52 onwards. These statistics were not properly kept by the NHL. From 1955-67, these statistics were calculated by Edward Yuen, who went through thousands of newspapers and calculated these stats by hand. He was very careful to maintain their accuracy. The remaining years of Sawchuk's career were calculated by Thomas Awad of the hockey analysis group. There may be a few errors in Awad's calculations at times, but they are a very good reference. The NHL did not officially keep records of saves percentage until the 1982/83 season.

Goals saved is not a statistic that most of us are used to seeing, so it takes some thought to make sense of the table. It is intended to be the number of goals this goalie would have prevented if he played on the average team that season in relations to the average goalie that season. Since shot quality data does not exist this far back, it is another unknown. A goalie facing low quality shots (due to a good defence) will tend to have a higher goals saved value than he deserves (and vice versa if he faces high quality shots due to a poor defence). Since each year is adjusted relative to the averages of that season, the reference of the league leader is goals saved is useful to be aware of.

The most obvious problem is that there exists no data for Terry Sawchuk's first two seasons. His 1949/50 season was short. He only played seven NHL games after being called up from the AHL. Likely, he did not record any great goals saved totals on such limited work. However, his 1950/51 season was a very good one. He led the NHL in wins and shutouts and made the first team all star (which is the closest equivalent to winning the Vezina Trophy today) and won the Calder Trophy as top rookie. This season is lost in the analysis.

In the seasons for which we have data, Sawchuk twice led the NHL in goals saved. These were the first two years where data exists. Very possibly, he would have led the season before the data exists, be we cannot be certain. After the 1954/55 season, Sawchuk stopped recording top values in goals saved. He became a much more average goaltender instead of being the best (or at least arguably the best). There is still a great deal of value in being an average goalie for a long period of time. This is especially true in a six team NHL that made little use of backup goalies. The average NHL goalie is the third or fourth best in the world most years.

During Sawchuk's career, Glenn Hall led the NHL in goals saved six times and Jacques Plante led four times. These are larger totals than Terry Sawchuk could have had even if we assume that he likely led the league in 1950/51 (the year before stats exist). This is certainly evidence that Sawchuk was not as good as those two rivals.

Over his career, Terry Sawchuk recorded 82.3 goals saved in the years for which we can verify stats. By way of comparison, Jacques Plante had 350.8 goals saved, Glenn Hall had 296.2 goals saved and Johnny Bower (who twice led the league) had 222.6 goals saved. These are all values significantly higher than Sawchuk (no matter what reasonable assumption we give him for his one and a bit seasons before stats exist). It is reasonable to conclude from this analysis that Sawchuk was the fourth best NHL goalie throughout much of his career (behind Plante, Hall and Bower). He was likely the best goalie until those three established themselves (though rivalled by Harry Lumley). From that point on, Sawchuk was roughly an average NHL goalie (which in a league with only six regular goalie jobs is still quite a mark of success).

However, a strong argument can be made that Sawchuk faced tougher shot quality than any of those three. Sawchuk's drop in his goals saved corresponded with his trade to Boston. Boston in the 50's was not a particularly good team and likely allowed high quality shots. By the time Sawchuk returned to Detroit, their defence was not as good as it had been. Red Kelly (their top defenceman in the past) was traded to Toronto in 1960 and was being developed from a defenceman into a centreman.

Certainly Jacques Plante played behind a better defence in Montreal (Doug Harvey, Tom Johnson, Jean-Guy Talbot), Johnny Bower played behind a better defence in Toronto (Carl Brewer, Tim Horton, Bob Baun) and Glenn Hall played behind a better defence in Chicago as well (Pierre Pilotte, Elmer Vasko). All three were Stanley Cup winners, while Detroit no longer made the playoffs every season. When Sawchuk moved on to Toronto in 1964/65 his goals saved recovers (though at this point he was no longer as good as Johnny Bower and became Bower's backup).

It is clear that Terry Sawchuk was a good goalie to last as long as he did and be at worst an average goalie on a weaker team throughout the majority of his career, after a several year run as an elite goalie. In a six team NHL being an average goalie is likely being the fourth best goalie in the world most seasons and that is an impressive record. There is reason to believe that were shot quality data available, it would benefit Sawchuk's standing in the league throughout much of his career. His goals saved ranking tends to track with the expected shot quality that he faced. I think it is clear from this analysis that Terry Sawchuk is not the best goalie ever. There are contemporaries (and people in later time periods) who consistently rank above him and it is hard to imagine that further data would change this, but Sawchuk is a top goalie. He was one of the three or four best in the NHL (and the world) for a long time. He was the best for a while. His placing in a list of the all time greats in hockey is not unreasonable. There is reason to argue if he should be top 20 or top 40 or top 60, and the data available does not answer this well. Those who saw Sawchuk's full career consider him one of the all time greats (almost unanimously) and it is hard to argue with that.

Because goals saved are referenced to average in a league, it is much easier for a goalie in a large NHL (say 30 teams) to be above average than it was to be in a six team NHL. In a 30 team NHL, a goalie might have a year where he is very good, but not not a serious Vezina candidate (such as Rick DiPietro or Marty Turco) this past season and rank well above average. In a six team NHL, with little to no backup goaltending, this would rank them below average. Sawchuk is a victim of this comparison to average since he played in a small league and he played much of his career behind a weaker than average (for a six team league) defence. How much of Plante, Hall or Bower's superiority is due to better defences may never be possible to tell, but I think the reality is closer than these numbers show.

Comments: Post a Comment

<< Home

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