The Puck Stops Here

One issue I have looked at this summer in sabermetrics and hockey posts is the adjustment of scoring from different eras. I have listed the top 10 adjusted goal scoring seasons and the top 10 adjusted assist scoring seasons according to a normalization method devised by the hockey outsider (Peter Albert). This normalization method is the best method I have seen to date to solve this problem, but it has problems with false normalization of assists in the 1920's when there were significantly less assists than today and lacks an adjustment for the quality of opposition the player plays against.

Here are the top 10 adjusted point scoring seasons:

Top 10 Adjusted Point Scoring Seasons of All Time

Name	Team	Year	Games Played	Adjusted Points	Actual Points
Cy Denneny	Ott	1924/25	28	179	45
Wayne Gretzky	Edm	1985/86	80	172	215
Wayne Gretzky	Edm	1984/85	80	170	208
Mario Lemieux	Pit	1988/89	76	169	199
Wayne Gretzky	Edm	1983/84	74	166	205
Duke Keats	Edm - WCHL	1921/22	25	165	55
Howie Morenz	Mon	1927/28	43	163	51
Wayne Gretzky	Edm	1982/83	80	161	196
Wayne Gretzky	Edm	1981/82	80	159	212
Wayne Gretzky	Edm	1986/87	79	159	183

This list shows the dominance of Wayne Gretzky. Six of the ten highest adjusted point scoring seasons are his. That shows his incredible dominance, but he may be somewhat helped that he played in the 21 team era and not the six team era. There are no original six team era players anywhere near the top 10. The highest ranking in the hockey outsider list is the 1952/53 95 point season Gordie Howe had that translates to 129 adjusted points and 33rd place. The original six (which is the era of the highest level of opposition calibre) is significantly underrepresented by this normalization system.

Also appearing in the top ten are Mario Lemieux's 199 point season in 1988/89. This year is deserving. There are also three years that are carryovers from the problems of false normalization that significantly increase 1920's adjusted assist totals. Three players make the list because of that problem. They are Duke Keats, Howie Morenz and Cy Denneny. Denneny is the most embarrassing of the three because he is ranked number one. There is no legitimate way to argue that Cy Denneny's 1924/25 season was the best point scoring season of all time. In fact, Denneny did not even win the NHL scoring title that season. He placed second to Babe Dye of Toronto. Dye scored 44 points and Denneny scored 42. It seems ridiculous to claim that the best scoring season of all time was not the best scoring season in the year where it occurred. The mistake occurs because Denneny scored 15 assists in an era when there were very few assists. This led the NHL. Babe Dye scored on 6 assists (which was good enough for 7th place in the entire league) but had more goals. Assists are overvalued by the normalization method in the early years of hockey because there were so few of them. If a player got a lot of assists (relative to the rest of the league) his point totals are normalized upwards well beyond the level they deserve. Denneny got 15 assists, but the normalization method considers this the equivlaent of 128 assists. It isn't. Fifteen assists can at most influence 15 goals. That isn't in the same ballpark as the number of goals and hence games that 128 assists would influence.

The adjustment of single season totals is interesting, but the false normalization of assists has led to some suspect results.