# Question: Talent level by years/era

Discussion in 'The History of Hockey' started by Weztex, Mar 26, 2007.

1. ### WeztexRegistered User

Joined:
Feb 6, 2006
Messages:
1,437
86
Trophy Points:
107
I'm heavily working on a database that mathematicaly list every player that ever played in the NHL (kinda like pnep's HHOF monitor) and some observations got me confused.

I've always defend the fact that there's no difference between a player finishing 1st in scoring in 1960 or in 2000, and used that as a rule. However it's obvious that, for example, the players finishing 5th in 1988 (Stastny, Robitaille, Messier) are much better players that the 5th in 1998 (Palffy, Leclair, Francis). That fact got me a lot of problems especially with the defensive years of the late 90's (i.e. Naslund over Hawerchuk). Moreover, I'm not too fond of adjusted stats which I believe grossly underate early days talent.

So, I'm not sure how can anyone could help me but what I'd like is some information about the level of talent of the NHL by era or years. For example, in 1945 the NHL was maybe 30% weaker in 1940. How does expansions or other disturbance affected the level of play, what year is stronger than the last, etc...

Hope I'm not too confusing.

2. ### Ogopogo*Guest

Wez, I would suggest that you try not to overthink this.

My perspective is that a scoring title is a scoring title no matter what year it is. Why do I say that? Well, the NHL offers the best players available every season they operate. If you are the best scorer in the NHL one year, you are equal to the best scorer in the NHL any year because the NHL provides the best hockey talent available. Expansion is irrelevant because the best one year is the same as the best from another year. Point totals are irrelevant.

Now, if someone subscribes to the theory that say, the 2000 season was stronger than the 1954 season, my question is how could they possibly determine that? There is no way to say that one season was stronger or weaker than another. Each season is an entity unto itself and you cannot compare the level of talent from one year to the other.

The way to compare across eras is to assume that all NHL seasons offer the best talent available. If you are the best scorer in the NHL one year that is the same as being the best scorer any year - with adjustments for seasons of particular domination. Then you compare players based on how long and to what degree they dominated.

To say that 1988 was a better season because Robitaille is a better player than Palffy is flawed logic, IMO and will lead you down a path with no end.

3. ### Hockey OutsiderRegistered User

Joined:
Jan 16, 2005
Messages:
4,576
495
Trophy Points:
124
Weztex, I've also done a lot of work with "adjusted stats" and I've run into the same problem that you have. So it's good to see that there are others thinking about the same topics.

One way to approach the problem is to consider the level of parity in the league. If there's a high level of parity, it probably means that the talent pool is high, since there would be few weak players and weak teams against whom you could rack up easy points and easy wins. If there's a low level of parity, there are probably not enough good players to fill every team's roster; good players and teams would easily dominate sub-NHL calibre players.

You can measure parity (between teams) easily if you have the data. I measure it as: standard deviation in teams' win percentage, divided by the league average win percentage. (The league average win percentage is always 50%, except for recent years with the idiotic "overtime loss point"). Here are the results:

1918-1931: Parity is very low. There's a huge gap between good and bad teams. This makes sense, because the NHL also had to compete with two other professional leagues (the PCHA and WCHL). There is a slight increase in parity after 1926, when the NHL became the sole professional league. (Mathematical note: some of the perceived lack of parity might be due to how we're measuring things... we base parity on the standard deviation in teams' win percentages. The schedule was extremely short in this era, so it's more likely to have a "fluke" season. This increases the standard deviation and therefore increases the perceived lack of parity)

1930-1943: Parity is almost as high as it is in the modern era. This makes sense, because the NHL has established itself as the best league in the world. There's a long period of stability.

1944-1945: Parity is extremely low. This makes sense, because most of the top players left to fight in WWII. Their replacements were not NHL calibre.

1946-1967: Parity is almost as high as it is in the modern era. The NHL is in a long period of stability.

1968-1986: Parity is very low (peaking in 1975) but declines gradually. The NHL doubled in size and competed with the WHA, so it makes sense that there was a strain on the talent pool.

1987-present: Parity has remained very high, and is virtually unchanged since the mid-80s. It did spike briefly from 1993-1996 due to the abundance of expansion teams. Again, the NHL has been fairly stable.

I'm not sure how much (if at all) this helps. But it does cause me to discount stats from 1918-20 and 1944-45 by quite a bit.

Last edited: Mar 26, 2007
4. ### MasaoRegistered User

Joined:
Nov 24, 2002
Messages:
11,131
327
Trophy Points:
199
But do they really?

What about WW2? What about all that time when there were almost no europeans in the NHL? What about how difficult it was for Canada to beat the Soviets in 1972? What about CSKA Moscow who tied the Montreal Canadiens 3-3 in 1975? What about all the players who defected to the WHA?

Nowadays, if there's a star player who's born in some far-off town in Siberia, and he has the skill to play in the NHL and be one of the best, there's a pretty good chance he will. I don't think this was always the case. The game is just so much bigger today, and more and more young players are competing for a job in the NHL. Plus, with all the money involved the players are treated completely differently.

Today's players follow a rigorous training program prepared by professionals. They have the best doctors to look into every boo-boo with the most advanced medical technology. I think NHL life is quite different for Jaromir Jagr than for Howie Morenz, in a time when players were paid peanuts.

Joined:
Jun 2, 2004
Messages:
13,224
105
Trophy Points:
141
Location:
Toronto
This is where I disagree with your methodology. IMO this kind of thinking removes the actual talent of the players from the equation.

Steve Yzerman came 3rd in scoring behind Gretzky and Lemieux in 1989 with 155 points. In the year 2000, Mark Recchi came third behind Jagr and Bure with 91 points. Even taking era into account, there's absolutely no way Mark Recchi ever scores 155 points. To say both are equal isn't fair to Yzerman, who clearly (in my eyes anyways) had the better offensive season.

Let's say Grant Fuhr is the best goalie of the 1980's (argument could be made for Smith as well, but other than that, a weak era in goaltending). Let's say Patrick Roy is the 2nd best goalie of the 90's after Hasek. Does that make Fuhr better than Roy? Clearly it doesn't.

Rob Blake has more Norris Trophies than Brad Park. Martin St. Louis has won more scoring titles than Maurice Richard. Peter Stastny never once made a first OR second all-star team, John Leclair has made 5.

All seasons are not created equal. The level of competition each player faces has to be taken into consideration, or it just doesn't make any sense to me. A strictly empirical approach takes common sense out of the equation.

6. ### Ogopogo*Guest

As I said, the NHL features the best players available. Players at war were not available. Players behind the Iron Curtain or in the WHA were not available to the NHL.

IMO, if you are the best scorer in the NHL one year that is equal to being the best scorer in the NHL any other year. You are the NHL's best.

7. ### Ogopogo*Guest

Why is the 155 put up by Yzerman better than the 91 put up by Recchi? Is it because it is just a big number? The 80s featured inflationary scoring, Yzerman's 155 is a mirage. If that version of Steve Yzerman would have played in 2000, he would have put up about 91 points. Remember, Bernie Nicholls put up 150 that year too - 150 in 1989 is certainly not the same thing as 150 in 1991 or any other year.

There's a lot more to it than you think. I have Roy ahead of Fuhr, Park ahead of Blake, Richard ahead of St. Louis and Stastny ahead of LeClair. Shocking, isn't it?

Joined:
Jun 2, 2004
Messages:
13,224
105
Trophy Points:
141
Location:
Toronto
So, do you believe that at their peaks, Mark Recchi was just as good as Steve Yzerman offensively? Because that's what this sounds like to me. This was the closest (I believe) that either have come to a scoring title. They both peaked at 3rd place.

For me, having seen both players play, I can say without a doubt that Yzerman was the better player.

9. ### Ogopogo*Guest

There is more to hockey than just scoring and there is more to my ratings than just finishing 3rd one time. A career is a cumulation of your seasons. Yzerman had more standout seasons than did Recchi, Yzerman is ahead in my ratings.

Joined:
Jun 2, 2004
Messages:
13,224
105
Trophy Points:
141
Location:
Toronto
yes, there's no disputing that Yzerman has accomplished more.

But my point is that you have Recchi's offensive peak on par with Yzerman's, which I don't believe to be true. Steve Yzerman had an amazing offensive season in 1989, and won the Pearson because of it (although I feel Lemieux deserved it). It's not his fault that he was playing in the shadow of the two greatest centres of all time.

Sometimes I think you get trapped by your methods a little bit. Context IS important.

11. ### Ogopogo*Guest

I disagree with the idea that he was buried behind Gretzky and Lemeiux. EVERY era had their own greats. Players of the 20s were buried behind Shore and Morenz. In the 50s it was Howe and Richard. The 70s was Orr and Esposito. Today it is Crosby and whoever else is going to emerge from the pack.

Remember, Recchi was buried behind Jagr in his prime.

The point is, if you are "in the shadow" of two other players, that means you are less of a player than they are and you are less of a player than players who dominated previous eras. Yzerman is behind Howe, Hull, Mikita, Dionne, Lafleur and many others in my ratings. I have him at #43. Great player but, if you really examine his career, he didn't even establish himself as a clear #3 behind Gretzky and Lemieux.

12. ### WeztexRegistered User

Joined:
Feb 6, 2006
Messages:
1,437
86
Trophy Points:
107
Ogopogo, the list I am working on works in a similar way that the one you defend. For now, I must say it gave me great results on an overall view. However, like Hockey Outsider, I've run into some results that every fan could contest only by common sense. Does your list put Naslund over Hawerchuk and Lafontaine? If so, I think that this need some working.

The only solution I've found so far is to arbitrarily weight every season based on what we know and that why I started this thread here. While I don't like the fact of throwing opinions into mathematics, I think the use of known facts can only be benificial into a project of this kind.

Thanks Hockey Outsider, that's a tremendous effort you put there. I arreciate it and will take the parity into account. Still, fully rely on this would tend to overate today's players because of the higher-than-ever parity.

I guess, I'll weight every season year by year and see what it looks like. I'll let you know if it looks good.

13. ### pnepRegistered User

Joined:
Mar 10, 2004
Messages:
2,552
55
Trophy Points:
101
Location:
Novosibirsk,Russia
I use Players "Best Seasons" to measure quality of opposition.

Example:

Maurice Richard "HHOF Monitor" PTS by seasons:

1942-43 -- 6,00
1943-44 -- 152,00
1944-45 -- 427,50
1945-46 -- 196,50
1946-47 -- 511,50
1947-48 -- 134,00
1948-49 -- 113,00
1949-50 -- 233,50
1950-51 -- 515,00
1951-52 -- 128,00
1952-53 -- 182,50
1953-54 -- 332,00
1954-55 -- 327,00
1955-56 -- 284,50
1956-57 -- 258,00
1957-58 -- 217,00
1958-59 -- 91,00
1959-60 -- 96,00

Maurice Richard "Best Season" is 1950-51.

I calculate "Best Seasons" for all NHL players and make this chart:

NHL Season -- # of "Best Seasons" -- Players Played -- % "Best Seasons"
==============================
1917-18 -- 23 -- 45 -- 51,00%
1918-19 -- 9 -- 36 -- 25,00%
1919-20 -- 21 -- 49 -- 43,00%
1920-21 -- 9 -- 48 -- 19,00%
1921-22 -- 8 -- 47 -- 17,00%
1922-23 -- 7 -- 44 -- 16,00%
1923-24 -- 13 -- 54 -- 24,00%
1924-25 -- 30 -- 83 -- 36,00%
1925-26 -- 36 -- 107 -- 34,00%
1926-27 -- 38 -- 150 -- 25,00%
1927-28 -- 32 -- 148 -- 22,00%
1928-29 -- 27 -- 147 -- 18,00%
1929-30 -- 33 -- 159 -- 21,00%
1930-31 -- 42 -- 183 -- 23,00%
1931-32 -- 28 -- 151 -- 19,00%
1932-33 -- 27 -- 171 -- 16,00%
1933-34 -- 36 -- 175 -- 21,00%
1934-35 -- 28 -- 178 -- 16,00%
1935-36 -- 24 -- 170 -- 14,00%
1936-37 -- 30 -- 176 -- 17,00%
1937-38 -- 34 -- 166 -- 20,00%
1938-39 -- 17 -- 155 -- 11,00%
1939-40 -- 26 -- 149 -- 17,00%
1940-41 -- 38 -- 151 -- 25,00%
1941-42 -- 38 -- 160 -- 24,00%
1942-43 -- 44 -- 143 -- 31,00%
1943-44 -- 71 -- 151 -- 47,00%
1944-45 -- 49 -- 132 -- 37,00%
1945-46 -- 23 -- 140 -- 16,00%
1946-47 -- 33 -- 150 -- 22,00%
1947-48 -- 31 -- 153 -- 20,00%
1948-49 -- 37 -- 147 -- 25,00%
1949-50 -- 53 -- 170 -- 31,00%
1950-51 -- 34 -- 170 -- 20,00%
1951-52 -- 22 -- 156 -- 14,00%
1952-53 -- 41 -- 161 -- 25,00%
1953-54 -- 21 -- 154 -- 14,00%
1954-55 -- 30 -- 157 -- 19,00%
1955-56 -- 26 -- 147 -- 18,00%
1956-57 -- 25 -- 150 -- 17,00%
1957-58 -- 21 -- 159 -- 13,00%
1958-59 -- 25 -- 145 -- 17,00%
1959-60 -- 31 -- 155 -- 20,00%
1960-61 -- 38 -- 160 -- 24,00%
1961-62 -- 17 -- 151 -- 11,00%
1962-63 -- 21 -- 155 -- 14,00%
1963-64 -- 20 -- 168 -- 12,00%
1964-65 -- 22 -- 171 -- 13,00%
1965-66 -- 20 -- 184 -- 11,00%
1966-67 -- 15 -- 179 -- 8,00%
1967-68 -- 71 -- 326 -- 22,00%
1968-69 -- 49 -- 329 -- 15,00%
1969-70 -- 53 -- 326 -- 16,00%
1970-71 -- 75 -- 388 -- 19,00%
1971-72 -- 57 -- 382 -- 15,00%
1972-73 -- 81 -- 405 -- 20,00%
1973-74 -- 95 -- 439 -- 22,00%
1974-75 -- 114 -- 503 -- 23,00%
1975-76 -- 87 -- 493 -- 18,00%
1976-77 -- 99 -- 506 -- 20,00%
1977-78 -- 90 -- 511 -- 18,00%
1978-79 -- 69 -- 502 -- 14,00%
1979-80 -- 151 -- 656 -- 23,00%
1980-81 -- 132 -- 641 -- 21,00%
1981-82 -- 116 -- 686 -- 17,00%
1982-83 -- 109 -- 678 -- 16,00%
1983-84 -- 107 -- 693 -- 15,00%
1984-85 -- 114 -- 675 -- 17,00%
1985-86 -- 101 -- 693 -- 15,00%
1986-87 -- 81 -- 688 -- 12,00%
1987-88 -- 135 -- 746 -- 18,00%
1988-89 -- 80 -- 734 -- 11,00%
1989-90 -- 113 -- 730 -- 15,00%
1990-91 -- 115 -- 743 -- 15,00%
1991-92 -- 102 -- 788 -- 13,00%
1992-93 -- 126 -- 789 -- 16,00%
1993-94 -- 148 -- 872 -- 17,00%
1994-95 -- 102 -- 808 -- 13,00%
1995-96 -- 116 -- 857 -- 14,00%
1996-97 -- 106 -- 849 -- 12,00%
1997-98 -- 105 -- 836 -- 13,00%
1998-99 -- 131 -- 902 -- 15,00%
1999-00 -- 136 -- 923 -- 15,00%
2000-01 -- 149 -- 975 -- 15,00%
2001-02 -- 157 -- 966 -- 16,00%

High % "Best Seasons" --> Low quality of opposition...

Graph:

PS
you may use Goals or PTS to calculate players 'Best Season', result almost identical.

Last edited: Oct 28, 2008
14. ### pnepRegistered User

Joined:
Mar 10, 2004
Messages:
2,552
55
Trophy Points:
101
Location:
Novosibirsk,Russia
all numbers without unknown stats from leagues in early years (NHA, PCHA, WCHL, WHL)

Last edited: Mar 27, 2007
15. ### Frightened Inmate #2Registered User

Joined:
Jun 26, 2003
Messages:
4,385
0
Trophy Points:
0
Location:
Calgary
The main problem that I have with Ogopogo's theory is that in the games early days it was an elite sport for the most part and accordingly the number of people privileged to play it was dramatically reduced. If there are only a relative handful of players available to play a game does that mean that the best automatically equates to the best later when the number of players is dramatically increased, I subscribe to the theory that with more players playing the game there are bound to be better players.

I will use an analogy, if I invented a game tomorrow and only a few people played it and I was the best... but in 100 years it was a popular game would that mean that I was the equal of the best player later. Logically the answer has to be no./

16. ### Stone87Registered User

Joined:
Mar 20, 2005
Messages:
1,750
0
Trophy Points:
0
Location:
Rochester
You can divide goals by the average number of goals scored for that season and then multiply that number times a common number. You would have to do the same for games played. I find that this actually gives a pretty good representation of who is the best. Assists are a little trickier and I've never tried to figure out if they have changed over time. Under my system the all-time goal scoring list would be as follows.

1.Gordie Howe - 1052
2.Wayne Gretzky - 799
3. Brett Hull - 768
4. Phil Esposito - 749
5. Maurice Richard - 749
6. Bobby Hull - 728
7. Jaromir Jagr - 695
8. Luc Robitaille - 679
9.Stan Mikita - 678
10.Teemu Selanne - 678
11.Mark Messier - 676
12. Steve Yzerman - 673
13.Brendan Shanahan - 673
14.Marcel Dionne - 659
15.Jean Beliveau - 649
16.Joe Sakic - 646
17.Mario Lemieux - 641
18.Mike Gartner - 640
19.Dave Andreychuk - 633
20.Frank Mahovolich - 623
21.Joe Niuewendyk - 585
22.Norm Ullman - 584
23.Mats Sundin - 582
24.Alex Delvecchio - 572
25.Peter Bondra - 571

This system isn't perfect but it helps in a variety of ways. The first problem is most of the top scorers are from the 1980's because of the freakishly high scoring games. Gretzky, Robitaille, Messier, Yzerman, and Dionne were great players and still have great stats even with the adjustment. Andreychuk and Gartner just had extremely long careers. Players like Ciccarelli, Kurri, LaFleur, and Francis however are not nearly as good as their stats would suggest. It also puts more players from the 50's and 60's into the top 25. The weakest era is definatly the 1970's though scoring was higher than it is now and yet players did not put up big numbers. This is because the difference between good and bad players decreased as a result of overexpansion.

Then I guess to apply this more towards your project just try and figure out how many 50 goal scorers per year and that sort of thing. To see which is the strongest.

Last edited: Mar 27, 2007
17. ### Nalyd PsychoRegistered User

Joined:
Feb 27, 2002
Messages:
24,415
1
Trophy Points:
0
Location:
No Bandwagon
Ogopogo one name puts a serious dent in your system...

Herb Cain.

18. ### WisentRegistered User

Joined:
Nov 15, 2003
Messages:
3,667
0
Trophy Points:
0
Occupation:
Student
Location:
Mannheim
I agree with your idea. At least if you are lookiing at the NHL and state that they offer the best talent available (which is doubtful still IMO, but we don`t want to go into that discussion again) you should at least take a look in the parity of the season and decades.

19. ### Czech Your MathRegistered User

Joined:
Jan 25, 2006
Messages:
4,852
3
Trophy Points:
86
Location:
bohemia
H.O. and Pnep each presented some good data, and I tend to agree with their general conclusions. A blending of the results from different methods may yield better results than those from any single method.

You might want to read a couple of other studies:

"The Relationship Between Goal-Scoring and League Talent Level" at Puckerings.com is an interesting hypothesis, although not the most scientific study.

"League Equivalencies" at HockeyAnalytics.com (Research category) doesn't address differing NHL talent levels directly. However, the relatively simple (conceptually) technique used to equate performance in other leagues to the NHL is intriguing, and the same technique could be used to compare consecutive NHL seasons to one another.

20. ### pnepRegistered User

Joined:
Mar 10, 2004
Messages:
2,552
55
Trophy Points:
101
Location:
Novosibirsk,Russia
Last edited: Mar 27, 2007

Joined:
Jun 2, 2004
Messages:
13,224
105
Trophy Points:
141
Location:
Toronto
My point isn't that Steve Yzerman would have dominated any other era. My point is that just because a player is #1 in his era, doesn't mean he's equal to the #1 from a different era. You make it sound like the greatest players of the game have been evenly divided through the ages, which is obviously untrue.

When we're talking about the two of the 3 most offensively dominant players of all time (Lemieux and Gretzky), I think it needs to be taken into account. Do you believe that Jagr was as good as Lemieux or Gretzky? Those two players burried the field on a regular basis. Yes, Jagr won scoring titles. Yes, Jagr was by far the best offensive player in the game in the late 90's. However, I've seen them both with my own eyes, and he had nothing on Lemieux. There is a massive difference in my mind finishing 3rd to Jagr and Sakic as opposed to Wayne and Mario.

If I can get you to agree that Jagr was not as good offensively as Gretzky and Lemieux (right?), then it should stand to reason that a player could be BETTER than Jagr, and still be WORSE than Gretzky and Lemieux. If this hypothetical player were to have played in 1989, he would be third place. If he were to play in 1999, he would be first. This is the issue.

Last edited: Mar 27, 2007
22. ### J0e Th0rnton*Guest

Im debating entering this thread before Bilros gets here with his Nonsense about Orr sucking because all 70's players are amateurs compared to today. and how "Lidstrom is better than Orr ever was"

The truth is, the Era's are different. In style mostly. As far as skill goes, they are not all that different. Some less informed individuals will try and brainwash people that hockey before the mid 80's were all amateur non athlete players who would not last against the talent of today, but its really a delusional attempt to fool themselves into thinking hockey players today are better.

Scoring was low in certain era's. Scoring picked up in the 70's because of the rise of Offensive defensemen and they changed the way the game was played from defense to offense first.
Scoring blew up in the 80's because of certain stars making their appearance, and system changes. Some defensemen under the new systems, played as a fourth forward instead of a defenseman, and thus, it became a different game.

Mid 90's-2004. Goaltending equipment much bigger+ the initiation of the instigator rule. The instigator removed the idea of players policing themselves, and as a result, clutching and grabbing, something players didn't let others get away with back in the day, became commonplace. Combining this with the defensive first systems and goaltending equipment twice as large, but half as heavy as the old days, and you had low scoring games.

Players adapted to their era's, but some went against the grain.
Trottier was a spectacular defensive forward in an offensive minded time. Imagine his numbers if he threw defense out the window like others and just scored. But then he wouldn;t have been trots.

I could go into further detail, but why bother.

23. ### BM67Registered User

Joined:
Mar 5, 2002
Messages:
4,641
1
Trophy Points:
141
Location:
In "The System"
Which is why they now use point percentage, rather than the W%. The difference is that instead of measuring your point total against the most points you could have gotten, 82/164=.500, they measure against all the points given out, 92/190=.484 (82 wins plus 26 loser points).

2741/(2460+281)=.500 for last year

So a team that won all 82 games, but went to OT 12 times, would have a W% of 1.000 but a P% of .932, and a team that went to OT 22 times would need 93 points to have a .500 P%, which would be a .567 W%.

24. ### WisentRegistered User

Joined:
Nov 15, 2003
Messages:
3,667
0
Trophy Points:
0
Occupation:
Student
Location:
Mannheim
Style of course as well but take a look at the table presented in this thread. Parity levels vary greatly from time to time. That is what may inflate some people`s stats.

25. ### Sens RuleRegistered User

Joined:
Sep 22, 2005
Messages:
23,032
22
Trophy Points:
116
By the same logic though the best player of any era could be the best player ever. Just becuse only 1/5th as many people played hockey in 1958 doesn't mean that Bobby Hull and Gordie Howe can't be 2 of the best 6 or 7 players ever to play the game. Or when 1/25th or 1/50th as many played Eddie Shore was the best. He could be one of the very best ever. Or he could be the 500th or 100th best ever.

You can see it in the Olympics with the 100m dash. The first 2 or 3 or 4 Olympics it was an eilte sport with very slow times. It is evident there would have been thousands of faster people than those competing in the Olympics. By say 1936 Jesse Owens could run it in 10 seconds flat. He could well have been as good a sprinter as anyone in the Olympics today, with a bit better training and equipment he cuts .1 or .2 seconds off his time. But those first few Olympics with times of 12 or 11 seconds those guys are no where near the best ever.

But hockey is much more complex than running in a straight line for 100 metres. It is so much about the mental aspect of the game.