seventieslord
Student Of The Game
I came up with this idea recently as a way to build off of this rough career even strength data I posted in another thread:
Of course, these numbers include the entire careers of all these players (minus two insigificant seasons for Hull), but cut off the entire prime of Gordie Howe. It is amazing that he keeps up with these players at all, let alone leads the pack.
it was commented by @Hockey Outsider :
And I would have to 100% agree. It would be hard to imagine that if Howe's years as an 18-30 year old were included, that his career ES numbers would look any worse; they'd almost certainly look significantly better.
Just to summarize, here are the numbers of all the players in the above lists:
and also, because these are full career numbers, which can be affected by longevity, here are the 8-year peak numbers I posted in that thread:
For Howe's best 8 year span, i just took the first 8 years available with the data that is there: 59-60 through 66-67, when he was 31-38.
But those are almost certainly not his best 8 years. So I went about recreating ES statistics for him. Here's how I did it:
1. Use hockey-reference to calculate Detroit's total ESGF and ESGA each season (ESGF is just the sum of all ESG in the player stats panel, ESGA is their total GA minus PPA and SHA)
2. Determine Detroit's ESGA in the games Howe missed in the first 20 years of his career (easy to do using gamelogs and the fact that he missed just 42 games through 1970)
3. Determine what percentage of Detroit's ESGA Howe tended to be on the ice for in games that he played from 1959-1970: 34, 38, 33, 34, 40, 19, 32, 32, 34, 33, 30. Aside from the 40 and 19, this hovered remarkably close to 33% his whole career.
4. Working backwards, assume Howe was on the ice for 33% of ESGA for the previous 10 seasons of his prime, then build in a natural dropoff (Howe's build-up in duties from age 18-20) of 31%, 29% and 22%
5. With those assumptions, one can now estimate the number of ESGA Howe was on the ice for
6. To estimate Howe's ESGF, start by looking at how many ESGF he had each season: 22, 38, 31, 51, 70, 61, 67, 48, 42, 42, 57, 49, 49.
7. Take a look at seasons with known ESGF numbers. What percentage of ESGF did Howe score a point on? It turns out that this number is also highly reliable: from 59-60 onwards: 67, 84, 73, 72, 70, 70, 72, 71, 73, 79, 77, 73. This was an average of 73% through 1971. however, it is fair to assume he was a bigger catalyst to the offense at his physical peak, so in assigning these numbers, I built in an age 22 peak that Howe rapidly builds up to, then slowly falls off from: 65, 70, 73.5, 75, 76.5, 76, 75.5, 75, 74.5, 74, 73.5, 73, 72.5. It's not a stretch to say that if we know how many ES points Gordie Howe scored (and we do), then we also have a pretty good idea how many goals he was on the ice for. These numbers are reliable and repeatable annually.
8. With Howe's estimated GF and GA in hand, it's now possible to create an R-on for him. It's also now possible to subtract those numbers from the team totals to get the "off" numbers, and an R-off. Here's the raw data:
Right off the bat, I would say that these numbers pass the smell test. Howe was posting on-off ratios of 1.85-2.10 in three of the four seasons following this sample at age 31-34, and he was assumedly better at 28-30. Ratios of 2.0+ are not unheard of in single seasons for all-time greats.
Check out Detroit's sharp decline from 1955 onwards. They were a tire fire with Howe off the ice for the late 50s. The 0.44 mark in 58-59 was an aberration, but in the 7 seasons surrounding that, they hovered from 0.61-0.89 while Howe stayed at 1.16-188 (aside from 1961). This is another reason the numbers pass the smell test: they pick up right where the real numbers left off.
Interestingly, Howe's 1959 R-on is his worst since his rookie season, but he still gets his best on-off ratio of his career, because Detroit was just so bad without him.
So, where does this put Howe on the charts I previously posted:
For his full career:
And for 8-year peak:
I hope you find this useful.
Where do you place Ovechkin on your personal list of the greatest players of all time?
Right... does that apply to Bobby Hull as well? Their careers are pretty damn similar. This is a table I put together last year out of curiosity comparing Crosby, Believeau, Hull, and Ovechkin hypothetically in a Canadian-only league if anyone is interested. Canada only league Ovechkin...
forums.hfboards.com
Here are the the skaters listed by @jigglysquishy as potential top-20 players of all-time, whose careers took place mostly or all past 1959, ranked by their team's R-off (GF-GA ratio at even strength with the player out of the lineup and/or on the bench):
Hull 1.19
Lidstrom 1.17
Ovechkin 1.07
Gretzky 1.04
Orr 1.03
Crosby 0.99
Bourque 0.96
Jagr 0.93
Howe (928 games post-1959) 0.85
Lemieux 0.84
I don't think that supporting cast is something that OV can really complain about.
R: on:
Orr: 2.01
Lidstrom: 1.41
Bourque: 1.37
Hull: 1.37
Crosby: 1.36
Jagr: 1.35
Howe: 1.26
Gretzky: 1.26
Lemieux: 1.23
Ovechkin: 1.21
Beliveau was 1.33 on, 1.33 off, in the last 748 games of his career. (yeah, I guess he counts if Howe does)
For total team ES performance numbers you'd have to weigh the on and off figures (approximately 40% of ES minutes were played by the forwards on average, and 45-50% by defensemen)
On/off:
Orr: 195%
Howe: 148%
Lemieux: 146%
Jagr: 145%
Bourque: 143%
Crosby: 137%
Lidstrom: 121%
Gretzky: 121%
Hull: 115%
Ovechkin: 113%
Of course, these numbers include the entire careers of all these players (minus two insigificant seasons for Hull), but cut off the entire prime of Gordie Howe. It is amazing that he keeps up with these players at all, let alone leads the pack.
it was commented by @Hockey Outsider :
Second, Howe's numbers look incredible. What's important to emphasize (as you did on the previous page) is this only includes his career from ages 31 onwards. It's missing four of his Hart trophies, and five of his Art Ross trophies. I wonder how much better his numbers would have been during his peak. (On the other hand, this excludes the slow start from his first 2-3 seasons, and that probably would have dragged his average down).
And I would have to 100% agree. It would be hard to imagine that if Howe's years as an 18-30 year old were included, that his career ES numbers would look any worse; they'd almost certainly look significantly better.
Just to summarize, here are the numbers of all the players in the above lists:
R on | R off | Ratio | |
Orr | 2.01 | 1.03 | 195% |
Howe (1959+) | 1.26 | 0.85 | 148% |
Lemieux | 1.23 | 0.84 | 146% |
Jagr | 1.35 | 0.93 | 145% |
Bourque | 1.37 | 0.96 | 143% |
Crosby | 1.36 | 0.99 | 137% |
Lidstrom | 1.41 | 1.17 | 121% |
Gretzky | 1.26 | 1.04 | 121% |
Hull | 1.37 | 1.19 | 115% |
Ovechkin | 1.21 | 107 | 113% |
and also, because these are full career numbers, which can be affected by longevity, here are the 8-year peak numbers I posted in that thread:
R on | R off | Ratio | |
Orr | 2.19 | 1.10 | 1.99 |
Bourque | 1.47 | 0.91 | 1.62 |
Crosby | 1.64 | 1.03 | 1.59 |
Jagr | 1.45 | 0.92 | 1.58 |
Gretzky | 1.67 | 1.15 | 1.5 |
Lemieux | 1.41 | 0.95 | 1.48 |
Howe (1959+) | 1.25 | 0.86 | 1.45 |
Ovechkin | 1.4 | 0.97 | 1.44 |
Lidstrom | 1.51 | 1.22 | 1.24 |
Hull | 1.31 | 1.22 | 1.07 |
For Howe's best 8 year span, i just took the first 8 years available with the data that is there: 59-60 through 66-67, when he was 31-38.
But those are almost certainly not his best 8 years. So I went about recreating ES statistics for him. Here's how I did it:
1. Use hockey-reference to calculate Detroit's total ESGF and ESGA each season (ESGF is just the sum of all ESG in the player stats panel, ESGA is their total GA minus PPA and SHA)
2. Determine Detroit's ESGA in the games Howe missed in the first 20 years of his career (easy to do using gamelogs and the fact that he missed just 42 games through 1970)
3. Determine what percentage of Detroit's ESGA Howe tended to be on the ice for in games that he played from 1959-1970: 34, 38, 33, 34, 40, 19, 32, 32, 34, 33, 30. Aside from the 40 and 19, this hovered remarkably close to 33% his whole career.
4. Working backwards, assume Howe was on the ice for 33% of ESGA for the previous 10 seasons of his prime, then build in a natural dropoff (Howe's build-up in duties from age 18-20) of 31%, 29% and 22%
5. With those assumptions, one can now estimate the number of ESGA Howe was on the ice for
6. To estimate Howe's ESGF, start by looking at how many ESGF he had each season: 22, 38, 31, 51, 70, 61, 67, 48, 42, 42, 57, 49, 49.
7. Take a look at seasons with known ESGF numbers. What percentage of ESGF did Howe score a point on? It turns out that this number is also highly reliable: from 59-60 onwards: 67, 84, 73, 72, 70, 70, 72, 71, 73, 79, 77, 73. This was an average of 73% through 1971. however, it is fair to assume he was a bigger catalyst to the offense at his physical peak, so in assigning these numbers, I built in an age 22 peak that Howe rapidly builds up to, then slowly falls off from: 65, 70, 73.5, 75, 76.5, 76, 75.5, 75, 74.5, 74, 73.5, 73, 72.5. It's not a stretch to say that if we know how many ES points Gordie Howe scored (and we do), then we also have a pretty good idea how many goals he was on the ice for. These numbers are reliable and repeatable annually.
8. With Howe's estimated GF and GA in hand, it's now possible to create an R-on for him. It's also now possible to subtract those numbers from the team totals to get the "off" numbers, and an R-off. Here's the raw data:
year | age | GP | sched | ESP | ~GF% | ~ESGF | TmESGF | TmESGA | TmESGA* | ~GA% | ~ESGA | R-on | GFWO | GAWO | R-off | On/off |
1947 | 18 | 58 | 60 | 22 | 0.65 | 33.85 | 150 | 162 | 157 | 0.22 | 34.54 | 0.98 | 116.15 | 127.46 | 0.91 | 1.08 |
1948 | 19 | 60 | 60 | 38 | 0.7 | 54.29 | 152 | 123 | 123 | 0.29 | 35.67 | 1.52 | 97.71 | 87.33 | 1.12 | 1.36 |
1949 | 20 | 40 | 60 | 31 | 0.735 | 42.18 | 152 | 109 | 75 | 0.31 | 23.25 | 1.81 | 109.82 | 85.75 | 1.28 | 1.42 |
1950 | 21 | 70 | 70 | 51 | 0.75 | 68.00 | 174 | 128 | 128 | 0.33 | 42.24 | 1.61 | 106.00 | 85.76 | 1.24 | 1.30 |
1951 | 22 | 70 | 70 | 70 | 0.765 | 91.50 | 197 | 116 | 116 | 0.33 | 38.28 | 2.39 | 105.50 | 77.72 | 1.36 | 1.76 |
1952 | 23 | 70 | 70 | 61 | 0.76 | 80.26 | 164 | 99 | 99 | 0.33 | 32.67 | 2.46 | 83.74 | 66.33 | 1.26 | 1.95 |
1953 | 24 | 70 | 70 | 67 | 0.755 | 88.74 | 167 | 101 | 101 | 0.33 | 33.33 | 2.66 | 78.26 | 67.67 | 1.16 | 2.30 |
1954 | 25 | 70 | 70 | 48 | 0.75 | 64.00 | 132 | 93 | 93 | 0.33 | 30.69 | 2.09 | 68.00 | 62.31 | 1.09 | 1.91 |
1955 | 26 | 64 | 70 | 42 | 0.745 | 56.38 | 150 | 92 | 87 | 0.33 | 28.71 | 1.96 | 93.62 | 63.29 | 1.48 | 1.33 |
1956 | 27 | 70 | 70 | 42 | 0.74 | 56.76 | 126 | 116 | 116 | 0.33 | 38.28 | 1.48 | 69.24 | 77.72 | 0.89 | 1.66 |
1957 | 28 | 70 | 70 | 57 | 0.735 | 77.55 | 146 | 125 | 125 | 0.33 | 41.25 | 1.88 | 68.45 | 83.75 | 0.82 | 2.30 |
1958 | 29 | 64 | 70 | 49 | 0.73 | 67.12 | 134 | 150 | 135 | 0.33 | 44.55 | 1.51 | 66.88 | 105.45 | 0.63 | 2.38 |
1959 | 30 | 70 | 70 | 49 | 0.725 | 67.59 | 120 | 177 | 177 | 0.33 | 58.41 | 1.16 | 52.41 | 118.59 | 0.44 | 2.62 |
Right off the bat, I would say that these numbers pass the smell test. Howe was posting on-off ratios of 1.85-2.10 in three of the four seasons following this sample at age 31-34, and he was assumedly better at 28-30. Ratios of 2.0+ are not unheard of in single seasons for all-time greats.
Check out Detroit's sharp decline from 1955 onwards. They were a tire fire with Howe off the ice for the late 50s. The 0.44 mark in 58-59 was an aberration, but in the 7 seasons surrounding that, they hovered from 0.61-0.89 while Howe stayed at 1.16-188 (aside from 1961). This is another reason the numbers pass the smell test: they pick up right where the real numbers left off.
Interestingly, Howe's 1959 R-on is his worst since his rookie season, but he still gets his best on-off ratio of his career, because Detroit was just so bad without him.
So, where does this put Howe on the charts I previously posted:
For his full career:
R on | R off | Ratio | |
Orr | 2.01 | 1.03 | 195% |
Howe | 1.53 | 0.95 | 162% |
Lemieux | 1.23 | 0.84 | 146% |
Jagr | 1.35 | 0.93 | 145% |
Bourque | 1.37 | 0.96 | 143% |
Crosby | 1.36 | 0.99 | 137% |
Lidstrom | 1.41 | 1.17 | 121% |
Gretzky | 1.26 | 1.04 | 121% |
Hull | 1.37 | 1.19 | 115% |
Ovechkin | 1.21 | 107 | 113% |
And for 8-year peak:
R on | R off | Ratio | |
Orr | 2.19 | 1.10 | 1.99 |
Howe | 1.9 | 0.97 | 1.96 |
Bourque | 1.47 | 0.91 | 1.62 |
Crosby | 1.64 | 1.03 | 1.59 |
Jagr | 1.45 | 0.92 | 1.58 |
Gretzky | 1.67 | 1.15 | 1.5 |
Lemieux | 1.41 | 0.95 | 1.48 |
Ovechkin | 1.4 | 0.97 | 1.44 |
Lidstrom | 1.51 | 1.22 | 1.24 |
Hull | 1.31 | 1.22 | 1.07 |
I hope you find this useful.