Post-consolidation VsX Benchmarks

[Iain Fyffe]

Just applied percentages with team subjective weight given to each attribute.

That sounds...entirely subjective and arbitrary. How do you confirm that the subjective weights are appropriate? How do you confirm that all relevant factors are considered? How do you confirm that no irrelevant factors are considered?

 

[Canadiens1958]

Evaluations

That sounds...entirely subjective and arbitrary. How do you confirm that the subjective weights are appropriate? How do you confirm that all relevant factors are considered? How do you confirm that no irrelevant factors are considered?

All that matters is the fairness of the process.

 

[Iain Fyffe]

All that matters is the fairness of the process.

Uh, yeah. That was my point. If the subjective weights are off, it's unfair. If not all the relevant factors are included, it's unfair. If some irrelevant factors are included, it's unfair. You're long on generalizations and rather short on specifics.

You're always keen to jump on any non-scouting method used because it doesn't consider some subtle details or another (even if it doesn't claim to), but you never demonstrate that scouting actually accounts for these details appropriately itself.

You say the scouts have accounted for these "variables." Can you demonstrate that they have done so in an appropriate manner? You justification is based on subjective weights by your own admission. So why should we accept that these subjective weights are correct?

 

[Ben Grimm]

Comparing points scored during one season is good for comparing points scored during one season. Since there is no other data involved it's obviously useless for anything else. Removing scoring leaders because one person thinks they scored too many points makes it even more useless.

 

[Canadiens1958]

Potential

Comparing points scored during one season is good for comparing points scored during one season. Since there is no other data involved it's obviously useless for anything else. Removing scoring leaders because one person thinks they scored too many points makes it even more useless.

See the potential of the VsX measure. Toying with it in some AAA leagues as usually with the Pee Wee thru Midget levels you have a youngster dominate for reasons other than talent. Age advantage, usage, etc. Making a few adjustments gives it some promise but will take a longer testing period to draw solid conclusions.

At the NHL level, especially trying to be ambitious and going back to the 1926 consolidation and trying to compare across eras without allowing for rule changes regarding assists, overtime and other scoring quirks over close to 90 seasons simply creates more questions then answers.

Upside is that the resulting questions, if researched and answered will provide some interesting information. So the generated process is worthwhile.

 

[Hockey Outsider]

In case anybody is curious, here are the "Vs. X" results for 1927-2014. I used an adjustment factor of 87 for 2014. I'm not sure if I fully understand how Sturm calculates this, so let me know if it should be updated.

Seven year weighted Vs. X score - 1927 to 2014

Rank|Player|Score
1 | Wayne Gretzky | 155.1
2 | Gordie Howe | 126
3 | Phil Esposito | 123.4
4 | Mario Lemieux | 120.4
5 | Jaromir Jagr | 114.6
6 | Bobby Orr | 109.3
7 | Stan Mikita | 108.1
8 | Bobby Hull | 107.1
9 | Jean Beliveau | 105.9
10 | Maurice Richard | 105.7
11 | Guy Lafleur | 104.9
12 | Ted Lindsay | 104.8
13 | Bill Cowley | 103.5
14 | Marcel Dionne | 103.2
15 | Howie Morenz | 102.8
16 | Andy Bathgate | 101.2
17 | Joe Sakic | 97.9
18 | Sidney Crosby | 97.9
19 | Alex Ovechkin | 97.5
20 | Charlie Conacher | 97.1
21 | Bill Cook | 96.6
22 | Doug Bentley | 96.2
23 | Frank Boucher | 95.4
24 | Elmer Lach | 95.4
25 | Max Bentley | 94.9
26 | Mike Bossy | 94.4
27 | Joe Thornton | 94.4
28 | Steve Yzerman | 93.5
29 | Bryan Trottier | 93.5
30 | Syl Apps Sr | 93
31 | Teemu Selanne | 92.9
32 | Martin St. Louis | 92.9
33 | Toe Blake | 92.6
34 | Sweeney Schriner | 91.9
35 | Peter Forsberg | 90.9
36 | Nels Stewart | 90.5
37 | Adam Oates | 90.2
38 | Bernie Geoffrion | 90.2
39 | Busher Jackson | 90
40 | Marty Barry | 89.9
41 | Mark Messier | 89.5
42 | Roy Conacher | 88.8
43 | Mark Recchi | 88.6
44 | Norm Ullman | 88.6
45 | Jean Ratelle | 88.5
46 | Peter Stastny | 88.3
47 | Brett Hull | 88.2
48 | Jari Kurri | 88.1
49 | Gordie Drillon | 88.1
50 | Syd Howe | 87.9
51 | Sid Abel | 87.8
52 | Paul Coffey | 87.7
53 | Bobby Clarke | 87.6
54 | Ron Francis | 87.6
55 | Milt Schmidt | 87.5
56 | Jarome Iginla | 87
57 | Evgeni Malkin | 86.7
58 | Dickie Moore | 86
59 | Pavel Bure | 86
60 | Dale Hawerchuk | 85.9
61 | Henri Richard | 85.6
62 | Frank Mahovlich | 85.5
63 | Paul Kariya | 85.4
64 | Denis Savard | 85.4
65 | Eric Lindros | 85.4
66 | John Bucyk | 85.3
67 | Alex Delvecchio | 84.9
68 | Gilbert Perreault | 84.6
69 | Bryan Hextall | 84.5
70 | Luc Robitaille | 84.4
71 | Ilya Kovalchuk | 84.3
72 | Darryl Sittler | 84.1
73 | Markus Naslund | 83.6
74 | Paul Thompson | 83.2
75 | Aurel Joliat | 83.1
76 | Henrik Sedin | 82.8
77 | Marian Hossa | 82.6
78 | Clint Smith | 82.6
79 | Daniel Alfredsson | 82.6
80 | Mats Sundin | 82.3
81 | Doug Gilmour | 82.3
82 | Theoren Fleury | 82.3
83 | Pierre Turgeon | 82.3
84 | Bill Mosienko | 82.2
85 | John LeClair | 82.1
86 | Pavel Datsyuk | 82
87 | Mike Modano | 81.7
88 | Jeremy Roenick | 81.5
89 | Ted Kennedy | 81.5
90 | Dany Heatley | 81.5
91 | Lynn Patrick | 81.2
92 | Sergei Fedorov | 81
93 | Ziggy Palffy | 80.6
94 | Bernie Nicholls | 80.3
95 | Rod Gilbert | 80.2
96 | Lorne Carr | 80.2
97 | Ken Hodge | 79.9
98 | Cooney Weiland | 79.4
99 | Brendan Shanahan | 79.3
100 | Michel Goulet | 79.3
101 | Patrik Elias | 79.3
102 | Keith Tkachuk | 79.3
103 | Eric Staal | 79.2
104 | Pat LaFontaine | 78.8
105 | Hooley Smith | 78.8
106 | Doug Weight | 78.6
107 | Cecil Dillon | 78.4
108 | Ryan Getzlaf | 78.4
109 | Brad Richards | 78.4
110 | Alexander Mogilny | 78.1
111 | Phil Watson | 78.1
112 | Jason Spezza | 77.6
113 | Alexei Yashin | 77.6
114 | Daniel Sedin | 77.3
115 | Bernie Federko | 77.3
116 | Vincent Lecavalier | 77.2
117 | Henrik Zetterberg | 76.7
118 | Bun Cook | 76.6
119 | Bert Olmstead | 76.3
120 | Bobby Bauer | 76.2
121 | Joe Primeau | 76
122 | Patrick Kane | 76
123 | Bobby Rousseau | 76
124 | Alex Kovalev | 75.8
125 | Nicklas Backstrom | 75.7
126 | Red Kelly | 75.7
127 | Herbie Lewis | 75.6
128 | Jacques Lemaire | 75.5
129 | Raymond Bourque | 75.3
130 | Pavol Demitra | 75.2
131 | Johnny Gottselig | 75
132 | Phil Goyette | 74.9
133 | Lanny McDonald | 74.6
134 | Denis Potvin | 74.3
135 | Vincent Damphousse | 74.2
136 | Bill Thoms | 74.2
137 | Brian Leetch | 74.1
138 | Dit Clapper | 74.1
139 | Rick Middleton | 74
140 | Marc Savard | 73.9
141 | Milan Hejduk | 73.8
142 | Don McKenney | 73.8
143 | Yvan Cournoyer | 73.6
144 | Anze Kopitar | 73.6
145 | Woody Dumart | 73.5
146 | Patrick Marleau | 73.3
147 | Tony Amonte | 73.3
148 | Neil Colville | 73.2
149 | Dave Keon | 73.2
150 | Gaye Stewart | 73.1
151 | Ray Whitney | 73
152 | Rod Brind'Amour | 72.8
153 | Larry Aurie | 72.7
154 | Corey Perry | 72.6
155 | Peter Bondra | 72.5
156 | Alex Tanguay | 72.5
157 | Herb Cain | 72.2
158 | Eddie Wiseman | 72.1
159 | Steve Larmer | 72.1
160 | Mike Ribeiro | 71.9
161 | Glenn Anderson | 71.9
162 | Kent Nilsson | 71.7
163 | Dennis Maruk | 71.7
164 | Rick MacLeish | 71.6
165 | Tod Sloan | 71.5
166 | Joe Mullen | 71.4
167 | Buddy O'Connor | 71.3
168 | Johnny Gagnon | 71
169 | Steve Shutt | 70.8
170 | Al MacInnis | 70.6
171 | Kenny Wharram | 70.6
172 | Joe Nieuwendyk | 70.5
173 | Dave Taylor | 70.4
174 | Pete Mahovlich | 70.3
175 | Rick Martin | 70.2
176 | Dino Ciccarelli | 69.9
177 | Marian Gaborik | 69.9
178 | Ebbie Goodfellow | 69.8
179 | Nicklas Lidstrom | 69.7
180 | Billy Taylor | 69.5
181 | Rene Robert | 69.4
182 | Dave Andreychuk | 69.4
183 | Baldy Northcott | 69.3
184 | Todd Bertuzzi | 69.2
185 | Paul Ronty | 69.1
186 | Bobby Smith | 69.1
187 | Jimmy Ward | 69
188 | Thomas Vanek | 68.8
189 | Steven Stamkos | 68.8
190 | Zach Parise | 68.7
191 | Gus Bodnar | 68.7
192 | Sid Smith | 68.6
193 | Bill Barber | 68.6
194 | Miroslav Satan | 68.6
195 | Phil Kessel | 68.5
196 | Olli Jokinen | 68.5
197 | Mike Gartner | 68.5
198 | Ed Litzenberger | 68.5
199 | Robert Lang | 68.3
200 | Jonathan Toews | 68.2
201 | Scott Gomez | 68.1
202 | Ace Bailey | 68.1
203 | Steve Sullivan | 68
204 | Martin Straka | 68
205 | Pit Martin | 67.9
206 | Craig Janney | 67.9
207 | Kevin Stevens | 67.8
208 | Camille Henry | 67.4
209 | Jason Allison | 67.2
210 | Dean Prentice | 67.1
211 | Brian Propp | 67
212 | Garry Unger | 67
213 | Syl Apps | 66.9
214 | Owen Nolan | 66.9
215 | Claude Provost | 66.9
216 | Butch Goring | 66.7
217 | Art Jackson | 66.7
218 | Rick Nash | 66.6
219 | Peter McNab | 66.5
220 | Petr Nedved | 66.4
221 | Shane Doan | 66.3
222 | Fred Stanfield | 66.2
223 | Glen Murray | 66.2
224 | Pierre Larouche | 66.1
225 | Art Chapman | 66
226 | Alex Zhamnov | 65.9
227 | Charlie Simmer | 65.5
228 | Wilf Paiement | 65.4
229 | Jason Pominville | 65.4
230 | Dennis Hull | 65.2
231 | Joe Carveth | 65.2
232 | Murray Oliver | 65.1
233 | Petr Sykora | 65.1
234 | Barry Pederson | 65
235 | Kirk Muller | 65
236 | Wayne Cashman | 65
237 | Walt Tkaczuk | 64.9
238 | Red Berenson | 64.9
239 | Brian Bellows | 64.9
240 | Vic Stasiuk | 64.8
241 | Phil Housley | 64.8
242 | Johnny Peirson | 64.7
243 | Cory Stillman | 64.7
244 | Grant Warwick | 64.7
245 | Alexander Semin | 64.6
246 | Neal Broten | 64.5
247 | Bill Guerin | 64.5
248 | Vaclav Prospal | 64.5
249 | Greg Adams | 64.5
250 | Bill Hay | 64.3

The biggest gainers are Crosby and Ovechkin (both in the top twenty). Thornton moves into the top thirty, and Malkin into the top sixty.

 

[Hockey Outsider]

Ten year weighted Vs. X score - 1927 to 2014

Rank|Player|Score
1 | Wayne Gretzky | 146.2
2 | Gordie Howe | 119.8
3 | Phil Esposito | 114.7
4 | Mario Lemieux | 113.9
5 | Jaromir Jagr | 107.7
6 | Stan Mikita | 103
7 | Maurice Richard | 102.2
8 | Bobby Hull | 101.6
9 | Jean Beliveau | 101.1
10 | Marcel Dionne | 97.9
11 | Ted Lindsay | 97.7
12 | Andy Bathgate | 96.8
13 | Joe Sakic | 94.8
14 | Bill Cowley | 93.7
15 | Guy Lafleur | 92.7
16 | Bobby Orr | 92.1
17 | Howie Morenz | 91.8
18 | Joe Thornton | 90.3
19 | Frank Boucher | 90.2
20 | Mike Bossy | 90.1
21 | Steve Yzerman | 89.7
22 | Teemu Selanne | 89.6
23 | Bryan Trottier | 88.2
24 | Toe Blake | 88.2
25 | Nels Stewart | 87.9
26 | Martin St. Louis | 87.8
27 | Elmer Lach | 87.4
28 | Syl Apps Sr | 87.1
29 | Adam Oates | 87
30 | Bill Cook | 87
31 | Alex Ovechkin | 86
32 | Mark Messier | 85.7
33 | Doug Bentley | 85.6
34 | Charlie Conacher | 85.5
35 | Mark Recchi | 85.2
36 | Max Bentley | 85.2
37 | Ron Francis | 84.9
38 | Bernie Geoffrion | 84.8
39 | Brett Hull | 84.5
40 | Norm Ullman | 84.5
41 | Jean Ratelle | 84.4
42 | Peter Forsberg | 84
43 | Marty Barry | 83.5
44 | Sweeney Schriner | 83.5
45 | Alex Delvecchio | 83.3
46 | Paul Coffey | 82.9
47 | Syd Howe | 82.8
48 | Sidney Crosby | 82.6
49 | Jarome Iginla | 82.5
50 | Jari Kurri | 82.5
51 | Dale Hawerchuk | 82.4
52 | Frank Mahovlich | 82.3
53 | Busher Jackson | 82.2
54 | Peter Stastny | 82.1
55 | Bobby Clarke | 82
56 | John Bucyk | 81.9
57 | Luc Robitaille | 81.7
58 | Gilbert Perreault | 80.8
59 | Milt Schmidt | 80.7
60 | Mats Sundin | 80.5
61 | Henri Richard | 80.3
62 | Paul Kariya | 80.1
63 | Darryl Sittler | 79.9
64 | Pierre Turgeon | 79.6
65 | Denis Savard | 79.2
66 | Roy Conacher | 79.1
67 | Theoren Fleury | 78.9
68 | Sid Abel | 78.9
69 | Mike Modano | 78.5
70 | Ilya Kovalchuk | 78.5
71 | Daniel Alfredsson | 78.4
72 | Marian Hossa | 78.4
73 | Doug Gilmour | 78.3
74 | Rod Gilbert | 78.2
75 | Sergei Fedorov | 77.9
76 | Eric Lindros | 77.8
77 | Jeremy Roenick | 77.2
78 | Henrik Sedin | 77
79 | Ted Kennedy | 77
80 | Aurel Joliat | 76.8
81 | Dickie Moore | 76.7
82 | Pavel Datsyuk | 76
83 | Bill Mosienko | 76
84 | Pavel Bure | 75.6
85 | Brendan Shanahan | 75.5
86 | Bryan Hextall | 75.4
87 | Markus Naslund | 75.3
88 | Keith Tkachuk | 75
89 | Patrik Elias | 74.9
90 | Bernie Nicholls | 74.8
91 | Lorne Carr | 74.5
92 | Bernie Federko | 74.4
93 | Clint Smith | 74.4
94 | Dany Heatley | 74.3
95 | Brad Richards | 74.3
96 | Alexander Mogilny | 74.1
97 | John LeClair | 73.5
98 | Doug Weight | 73.5
99 | Daniel Sedin | 73.3
100 | Pat LaFontaine | 73.3
101 | Michel Goulet | 73.2
102 | Eric Staal | 73
103 | Ken Hodge | 72.8
104 | Phil Watson | 72.6
105 | Alexei Yashin | 72.4
106 | Vincent Lecavalier | 72.3
107 | Raymond Bourque | 72.1
108 | Paul Thompson | 71.7
109 | Vincent Damphousse | 71.7
110 | Lynn Patrick | 71.7
111 | Ziggy Palffy | 71.7
112 | Hooley Smith | 71.6
113 | Red Kelly | 71.5
114 | Cooney Weiland | 71.4
115 | Henrik Zetterberg | 71.3
116 | Bert Olmstead | 70.7
117 | Dave Keon | 70.5
118 | Bun Cook | 70.5
119 | Jacques Lemaire | 70.4
120 | Alex Kovalev | 70.2
121 | Dit Clapper | 70.1
122 | Brian Leetch | 70
123 | Johnny Gottselig | 69.8
124 | Yvan Cournoyer | 69.6
125 | Bobby Rousseau | 69.6
126 | Patrick Marleau | 69.6
127 | Woody Dumart | 69.5
128 | Cecil Dillon | 69.5
129 | Rod Brind'Amour | 69.4
130 | Pavol Demitra | 69.4
131 | Ray Whitney | 69.4
132 | Jason Spezza | 69.3
133 | Lanny McDonald | 69.2
134 | Steve Larmer | 68.8
135 | Phil Goyette | 68.8
136 | Tony Amonte | 68.7
137 | Denis Potvin | 68.6
138 | Evgeni Malkin | 68.5
139 | Rick Middleton | 68.4
140 | Don McKenney | 68.3
141 | Milan Hejduk | 68.2
142 | Herbie Lewis | 68
143 | Peter Bondra | 67.9
144 | Alex Tanguay | 67.6
145 | Joe Mullen | 67.4
146 | Dino Ciccarelli | 67.3
147 | Bill Thoms | 67.3
148 | Herb Cain | 67.2
149 | Joe Nieuwendyk | 67.2
150 | Al MacInnis | 67
151 | Glenn Anderson | 66.7
152 | Bobby Smith | 66.7
153 | Dave Andreychuk | 66.7
154 | Ryan Getzlaf | 66.5
155 | Nicklas Lidstrom | 66.5
156 | Mike Ribeiro | 66.4
157 | Mike Gartner | 66.3
158 | Gordie Drillon | 66.2
159 | Marc Savard | 66.1
160 | Pete Mahovlich | 66
161 | Bill Barber | 65.6
162 | Dave Taylor | 65.2
163 | Dennis Maruk | 65.1
164 | Ebbie Goodfellow | 64.9
165 | Brian Propp | 64.9
166 | Buddy O'Connor | 64.9
167 | Dean Prentice | 64.6
168 | Scott Gomez | 64.3
169 | Eddie Wiseman | 64.3
170 | Marian Gaborik | 64.3
171 | Larry Aurie | 64.2
172 | Tod Sloan | 64.1
173 | Rick MacLeish | 63.9
174 | Olli Jokinen | 63.7
175 | Pit Martin | 63.5
176 | Shane Doan | 63.3
177 | Miroslav Satan | 63.3
178 | Kenny Wharram | 63.2
179 | Owen Nolan | 63.1
180 | Rick Martin | 63
181 | Steve Shutt | 62.9
182 | Garry Unger | 62.8
183 | Phil Housley | 62.7
184 | Alex Zhamnov | 62.3
185 | Steve Sullivan | 62.3
186 | Rick Nash | 62.2
187 | Craig Janney | 62
188 | Todd Bertuzzi | 61.9
189 | Brian Bellows | 61.8
190 | Gus Bodnar | 61.8
191 | Jimmy Ward | 61.7
192 | Robert Lang | 61.5
193 | Tom Lysiak | 61.5
194 | Butch Goring | 61.3
195 | Johnny Gagnon | 61.3
196 | Pat Verbeek | 61.2
197 | Martin Straka | 61
198 | Claude Provost | 60.9
199 | Petr Sykora | 60.8
200 | Murray Oliver | 60.8
201 | Rene Robert | 60.8
202 | Pierre Larouche | 60.7
203 | Camille Henry | 60.6
204 | Petr Nedved | 60.3
205 | Kirk Muller | 60.1
206 | Anze Kopitar | 60
207 | Ryan Smyth | 59.9
208 | Corey Perry | 59.9
209 | Thomas Vanek | 59.8
210 | Cory Stillman | 59.8
211 | Peter McNab | 59.7
212 | Gaye Stewart | 59.6
213 | Dennis Hull | 59.6
214 | Vyacheslav Kozlov | 59.6
215 | Wayne Cashman | 59.6
216 | Neil Colville | 59.5
217 | Kevin Stevens | 59.5
218 | Wilf Paiement | 59.4
219 | Jason Arnott | 59.3
220 | Bill Guerin | 59.3
221 | Neal Broten | 59.3
222 | Andrew Brunette | 59.3
223 | Cliff Ronning | 59.1
224 | Sergei Gonchar | 59
225 | Eddie Shore | 58.9
226 | Bob Nevin | 58.9
227 | Vic Hadfield | 58.8
228 | Jozef Stumpel | 58.8
229 | Larry Murphy | 58.7
230 | Jean Pronovost | 58.5
231 | Fred Stanfield | 58.5
232 | George Armstrong | 58.5
233 | Steve Thomas | 58.5
234 | Red Berenson | 58.4
235 | Gary Roberts | 58.3
236 | Ivan Boldirev | 58.3
237 | Chris Drury | 58.2
238 | Tomas Sandstrom | 58
239 | Michael Nylander | 57.9
240 | Glen Murray | 57.8
241 | Mush March | 57.8
242 | Greg Adams | 57.8
243 | Baldy Northcott | 57.8
244 | Rick Tocchet | 57.8
245 | Bobby Bauer | 57.7
246 | Saku Koivu | 57.5
247 | Joe Primeau | 57.4
248 | Mike Ridley | 57.4
249 | Kent Nilsson | 57.4
250 | Robert Reichel | 57.3

Thornton moves into the top twenty, and St. Louis is in the top thirty. It's remarkable that Ovechkin is 31st considering he's only played nine seasons (one of which was forgettable). Iginla and Crosby both move into the top fifty.

 

[TheDevilMadeMe]

HO, 87 is the correct benchmark for 2013-14. Getzlaf isn't even close to being an outlier as the #2 scorer.

Question - would it be relatively easy for you to re-run these numbers using Sturm's war-year adjustments? They are not as mathematically formulaic, but IMO they give a better estimate of the offensive value of the war years players.

Edit: See Post 131 for the new benchmarks from 1942-43 to 1945-46.

Edit: See also Post 133 for an adjustment for the over-awarding of assists in Chicago in 1948-49.

 

[Hockey Outsider]

I updated and spot-checked against the players Sturm calculated in posts 132 to 138. I'm still getting two differences:

- Cowley: agrees
- D Bentley: agrees
- M Bentley: I calculate 91.2, he has 89.4. Difference is due to his best season (1947), which I calculate as 114 (72 points compared to benchmark of 63) - Sturm had 102.
- Lach: agrees (out by 0.2 due to rounding)
- Howe: agrees (out by 0.1 due to rounding)
-Blake: I calculate 86.4, he has 85.3. Difference due to 7th best season (1947) which I calculate as 79 (50 points compared to benchmark of 63) - Sturm had 73.

Maybe it's a coincidence, or maybe I'm missing an adjustment to 1947? I left the benchmark unchanged at 63. Will post the update (won't be for several days unfortunately), but wanted to confirm the numbers first.

 

[TheDevilMadeMe]

I updated and spot-checked against the players Sturm calculated in posts 132 to 138. I'm still getting two differences:

- Cowley: agrees
- D Bentley: agrees
- M Bentley: I calculate 91.2, he has 89.4. Difference is due to his best season (1947), which I calculate as 114 (72 points compared to benchmark of 63) - Sturm had 102.
- Lach: agrees (out by 0.2 due to rounding)
- Howe: agrees (out by 0.1 due to rounding)
-Blake: I calculate 86.4, he has 85.3. Difference due to 7th best season (1947) which I calculate as 79 (50 points compared to benchmark of 63) - Sturm had 73.

Maybe it's a coincidence, or maybe I'm missing an adjustment to 1947? I left the benchmark unchanged at 63. Will post the update (won't be for several days unfortunately), but wanted to confirm the numbers first.

Good catch.

It looks like Sturminator used Maurice Richard's 2nd place finish at 71 as the benchmark for 1946-47 when he calculated the scores for the war year players, rather than the 63 that he should have used and is listed in the OP. So your numbers would be correct... Except this creates another major problem - he used the assumed 71 benchmark from 1946-47 as a basis for his war time adjustment for 1945-46... not sure what to do about that:

1945-46: Some argue that this is not really a war year, but I disagree. The league was clearly still quite weak, Apps only played 40 games, Schmidt was not in hockey shape, and the scoring leaderboard is just a hot mess. What to do?

Luckily, we can repeat the process we used for the 1944-45 season here. Max Bentley won the scoring championship in consecutive seasons starting in 1945-46. The problem in this year is that his margin is much bigger than in the following season in a full-strength league. So if we assume that Max is the legitimate scoring champ but recalculate the margin between #1 and #2 to establish our benchmark for the 1945-46 season, this is what we get:

72/71 [1946-47 margin of victory] = 1.01
61/1.01 [Bentley's 1945-46 scoring output/above margin] = 60

I propose 60 as a benchmark for the 1945-46 season.

Unfortunately, it looks like he is in the middle of one of his breaks from hfboards, so I wouldn't expect him to clarify any time soon.

 

[Sturminator]

I updated and spot-checked against the players Sturm calculated in posts 132 to 138. I'm still getting two differences:

- Cowley: agrees
- D Bentley: agrees
- M Bentley: I calculate 91.2, he has 89.4. Difference is due to his best season (1947), which I calculate as 114 (72 points compared to benchmark of 63) - Sturm had 102.
- Lach: agrees (out by 0.2 due to rounding)
- Howe: agrees (out by 0.1 due to rounding)
-Blake: I calculate 86.4, he has 85.3. Difference due to 7th best season (1947) which I calculate as 79 (50 points compared to benchmark of 63) - Sturm had 73.

Maybe it's a coincidence, or maybe I'm missing an adjustment to 1947? I left the benchmark unchanged at 63. Will post the update (won't be for several days unfortunately), but wanted to confirm the numbers first.

Your numbers are correct, HO. Thank you for double-checking my work, and sorry I was not here at the time to clarify my intent.

 

[Sturminator]

So, I think now is a good time to move the ball a bit further here. A couple of things I'd like to work on:

1) We should integrate HO's corrections to the wartime fudges into the final product. I will update the OP to show both the original and the new numbers, but we should also probably update the spreadsheets.

2) We should incorporate the adjustment discussed here into D. Bentley and R. Conacher's 1948-49 assist numbers, and their final numbers. I will update the OP to make a note of it (though it doesn't change the benchmark), but we should also probably update the spreadsheets.

3) I think we should probably do away with the weighting system, per TDMM's suggestion. The truth of the matter is that the weights really don't make much of a difference in the numbers, but they add an unnecessary level of complexity and opacity which makes the system, as a whole, less accessible and more difficult to work with. As clarity and ease of use are among the first-order goals of this project, I think it is time for the weighting system to go.

4) I'm going to try to come up with a system to handle the Bruins-mania era of the early 70's, similar to my efforts with the wartime fudges. Not quite sure how this will look at present, but I'm open to suggestions.

-------------------------------------------

I'd be grateful for any help with the above. Maybe we can get an improved system working in time for the next round of end-of-season VsX numbers? That's my present goal, at any rate.

 

[Sturminator]

So, here are my initial thoughts/notes on a potential Bruins-era fudge:

1) The goal should be to "normalize, not remove" the top scorers from the Bruins. What I mean is this: we should not assume that the Bruins players (most specifically Orr and Esposito) did not exist, but rather, we should set our assumptions with their scoring performance normalized to that of a "standard" top scorer among defensemen and forwards, respectively.

1a) The problem here is that normalizing Orr essentially just takes him out of the equation because there is no other defenseman who has ever led the NHL in scoring in the post-consolidation era. Paul Coffey was 2nd in scoring once, so he could be a normalization target, but he was playing with Gretzky at the time, and has some normalization issues of his own. Ultimately, normalizing Orr to the performance of an average top-scoring defenseman means removing him from the scoring lists for the purposes of this project.​

2) Normalizing the forwards is easier. I propose that we simply take all Bruins forwards at the top of the scoring list in any given season (sometimes only Espo, sometimes more), and normalize them to a standard top scorer among forwards: just take the scoring of the #2 guy and add 5% to that. I haven't calculated the actual average gap between #1 and #2, but I think 5% is a reasonable number.

3) After normalization, just apply the standard VsX method.

---------------------------------------------------

Results if the above is applied, with reference to existing benchmarks:

1967-68: ok

1968-69: ok

1969-70: Mikita = ok. If we normalize Espo to "standard top-scoring forward", Mikita remains the benchmark

1970-71: BROKEN. If we normalize Espo-Bucyk-Hodge to "standard top-scoring forward", the benchmark becomes Ullman at 85 (was averaged - 100)

1971-72: BROKEN. If we normalize Espo to "standard top-scoring forward", the benchmark becomes Ratelle at 109 (was Orr - 117)

1972-73: ok

1973-74: BROKEN. If we normalize Espo-Hodge-Cashman to "standard top-scoring forward", the benchmark becomes Clarke at 87 (was averaged - 106)

1974-75: BROKEN. If we normalize Espo to "standard top-scoring forward", the benchmark becomes Dionne at 121 (was Esposito - 127)

So, using the above methodology for a Bruins-era fudge, we're looking at minor-to-moderate downward shifts of the benchmark for 1971-72 and 1974-75 (the Orr/Espo seasons), and large downward shifts for 1970-71 and 1973-74 (the seasons when multiple Bruins forwards outscored every other forward in the league). On first blush, I think it passes the smell test; large adjustments should be expected and desirable for the two seasons in which the entire Bruins offense simply broke the league.

Comments, criticisms, and personal attacks are welcome.

 

[Sturminator]

At this point, I think it is the 1970-71 and 1973-74 seasons that are most problematic. The above solution may be workable, but it may also go too far. Above, I normalized all Bruins forwards into a single normalized top forward scorer. This may not be the way to go. If we, instead, normalize all three (in both cases) Bruins forwards into three top scorers all tied at #1 in scoring, it changes our benchmarks in the following ways:

1970-71:

Original:
1. Esposito - 152
2. Orr - 139
3. Bucyk - 116
4. Hodge - 105
5. Hull - 96
6. Ullman - 85
7. Cashman - 79
8. McKenzie - 77
9. Keon - 76
9. Beliveau - 76
...average as benchmark: 100

First Fudge:

[Bruin] - 101
Hull - 96
Ullman - 85 = benchmark

Second Fudge:

Esposito - 101
Bucyk - 101
Hodge - 101
Hull - 96
Ullman - 85
Cashman - 79
MacKenzie - 77
Stanfield - 76
...average as benchmark: 90

and...

1973-74:

Original:
1. Esposito - 145
2. Orr - 122
3. Hodge - 105
4. Cashman - 89
5. Clarke - 87
6. Martin - 86
...average as benchmark: 106

First Fudge:

[Bruin] - 91
Clarke - 87 = benchmark

Second Fudge:

Esposito - 91
Hodge - 91 = benchmark
Cashman - 91
Clarke - 87
Martin - 86

I dunno...I tend to think the second set of fudges are better for these seasons.

 

[TheDevilMadeMe]

What's the issue with just using the highest non-Bruin in the years broken by the Bruins?

 

[Sturminator]

What's the issue with just using the highest non-Bruin in the years broken by the Bruins?

You mean as the benchmark, on the assumption that Espo leads the league in a more normal way, but no other Bruins are at the top in a world where Bobby Orr is not ridiculous? That may ultimately be the best method; it is certainly the clearest.

 

[seventieslord]

so pretty much back to the way I had it then?

 

[Sturminator]

Well, simply using the first non-Bruin in 1970-71 is problematic because the gap between Hull and Ullman is more than 10%. So, the VsX system would choose Ullman as the #3 scorer in this case. Maybe that is the best way to go, but maybe not. I mean, in reality, he was the 6th leading scorer in that season. Setting him as the benchmark feels a little galling.

For the 1973-74 season, using Clarke's scoring as the benchmark doesn't look like a problem from a methodological standpoint.

It's just...I dunno, I recognize that the system as it stands is likely too harsh at least in the benchmarking for 70-71 and 73-74 (outside of these two seasons, it seems much less problematic), but I don't want to overrate these players, either. I think the wartime fudge works because it found a middle way between the extremes. I'm not sure just completely ignoring the Bruins is the way to go. That feels more like the lower or "most generous" extreme, and I'd prefer a more moderate approach.

 

[Hockey Outsider]

3) I think we should probably do away with the weighting system, per TDMM's suggestion. The truth of the matter is that the weights really don't make much of a difference in the numbers, but they add an unnecessary level of complexity and opacity which makes the system, as a whole, less accessible and more difficult to work with. As clarity and ease of use are among the first-order goals of this project, I think it is time for the weighting system to go.

I agree with this point. It doesn't provide useful information and makes the system appear unnecessarily complicated.

I looked at the players with the top 250 seven-year VsX using the traditional weighting (18-19-20-19-18-17-16), then compared the results to the unweighted system.

Of the 250 players, only 33 had a difference of greater than 1.0%, and just seven had a difference of more than 2.0%. Those players are:

1. Claude Giroux (3.3% - four strong seasons, two weak ones)
2. Steven Stamkos (3.1% - same as above)
3. Paul Haynes (2.3% - significant drop-off in seasons six and seven)
4. Ace Bailey (2.2% - significant drop-off in seasons five, six and seven)
5. Jason Allison (2.0% - same as above)
6. Billy Taylor (2.0% - significant drop off in seasons six and seven)
7. Paul Ronty (2.0% - drops in seasons five and six, then huge drop in season seven)

Under the weighted method, only 91 players have a VsX of 80.0 or higher. Just three players` scores change by as little as 1.0% when we compared it to the unweighted results (Hextall - 1.2%; Orr - 1.1%; Malkin - 1.0%). Many ``top heavy`` players like Lafleur, Conacher, Bure, etc are affected by less than 1%.

Conclusion: I agree with Sturm`s idea. The weighting system adds unnecessary complexity to the project (both in terms of making the calculation more difficult, and making the system appear confusing and inaccessible to readers). It has a minimal impact on the rankings (and mostly hurts players with very short primes, who wouldn't fare well in this system anyway). Let's do away with the weightings.

 

[TheDevilMadeMe]

The weightings make a lot more of a difference when talking about 10 year primes, right?

 

[Hockey Outsider]

The weightings make a lot more of a difference when talking about 10 year primes, right?

You're right. If we look at the top 250 players (based on the standard 10-year weighted VsX) and compare it to the unweighted version, all but 17 players have their scores change by at least 1.0%.

The only exceptions are Delvecchio, Sundin, Gilbert, Francis, Gartner, N. Stewart, Housley, Turgeon, Propp, Sakic, Damphousse, Robitaille, Ronning, Ciccarelli, Oates, M. Richard, B. Smith. This means (unsurprisingly) that these players were remarkably consistent.

About one-fifth of players (57 to be exact) change by more than 3.0% - I'd consider this to be a fairly small number given the fairly heavy weighting that went into the initial formula.

Thirteen players' results vary by at least 5%: Getzlaf, Crosby, Orr, Northcott, Perry, Kopitar, G. Stewart, Colville, Nilsson, Malkin, Drillon, Bauer, Primeau. Not surprisingly, these are all players who had short primes (either due to injuries or because they're currently playing). When we stop applying the weighting to their best years, their overall score suffers.

Looking at players with a weighted score of 80.0 or higher (sixty players total), nine players' results change by 3.0% or more: M. Bentley, Schriner, Morenz, Lafleur, Conacher, Ovechkin, Crosby and Orr.

There`s definitely more of a change here, but even still it`s fairly small. I still think the added work effort & complexity isn`t worth it.

 

[TheDevilMadeMe]

Thanks for going through all that.

There`s definitely more of a change here, but even still it`s fairly small. I still think the added work effort & complexity isn`t worth it.

Right. I think the weighting was done to make 10-year VsX something of an all-in-one number for evaluating a post-expansion players offensive value, and in that case, it made sense to weigh the best 5 or so years higher than the next best 5 or so (to simplify things).

But in practice, most people will just use 7 year VsX alongside 10 year VsX, anyway.

 

[Sturminator]

Right. I think the weighting was done to make 10-year VsX something of an all-in-one number for evaluating a post-expansion players offensive value, and in that case, it made sense to weigh the best 5 or so years higher than the next best 5 or so (to simplify things).

But in practice, most people will just use 7 year VsX alongside 10 year VsX, anyway.

Yeah, there is some utility in the 10 year weighting, but the benefit is, I think, clearly outweighed by the cost. We'll just have to go back to acknowledging on the margins which players are particularly hurt or helped by an unweighted system.

Thanks for your work here, HO.

What do you two think about the problem of the early 1970's? Are we better off just using Ullman and Clarke as the benchmarks for the two really broken years in question, or is a "middle way" like the second fudge method above preferable?

Should we implement a general rule regarding seasons in which a forward/defenseman from the same team finish in the top-2 in scoring? This would affect the 1983-84 benchmark, as well, though mildly so (would fall from Coffey's 126 to Goulet's 121 in this season).

 

[BM67]

1970-71 Points by teams
Bos 1096
Chi 759
Mtl 745
NYR 685
AVG 641
Tor 637
Van 620
LA 610
Buf 575
StL 574
Pit 572
Det 550
Phi 532
CGS 514
Min 510

1970-71 Top 10
1. Esposito - 152
2. Orr - 139
3. Bucyk - 116
4. Hodge - 105
5. Hull - 96
6. Ullman - 85
7. Cashman - 79
8. McKenzie - 77
9. Keon - 76
9. Beliveau - 76

Take the scorers and divide by (team + Avg) then multiply by (Avg + Avg)

Adjusted 1970-71 Top 10
1. Esposito - 112
2. Orr - 103
3. Hull - 89
4. Bucyk - 86
5. Ullman - 85
6. Hodge - 77
7. Keon - 76
7. Perreault - 76
7. Drouin - 76
10. Tkaczuk - 73

Mahovlich playing with 2 teams would require a further step, and he would probably be at 73 as well.

 

[Sturminator]

Thank you for that work, BM. It provides another interesting take on the problem. I guess the benchmark would be set at Hull's 89 points under your system (leaving Orr out, as I think we should). I think this is obviously too complex to use in what is supposed to be an accessible system, but the perspective it provides is quite useful.

After considering the problem a bit more, I think we should move forward with the following rule - call it the Orr Rule:

In a season in which a defenseman places in the top-2 in scoring:

1) His score is removed.

and

2) Forwards from his team at the top of the scoring table (before scorers from other teams) have their scores normalized to #2 + 5% for purposes of VsX benchmarking.

This is basically fudge #2, from the above. It would specifically affect the benchmarks in the following ways:

1970-71:

Old:
1. Esposito - 152
2. Orr - 139
3. Bucyk - 116
4. Hodge - 105
5. Hull - 96
6. Ullman - 85
7. Cashman - 79
8. McKenzie - 77
9. Keon - 76
9. Beliveau - 76
...average as benchmark: 100

New:
1. Esposito - 101
1. Bucyk - 101
1. Hodge - 101
4. Hull - 96
5. Ullman - 85
6. Cashman - 79
7. MacKenzie - 77
8. Stanfield - 76
...average as benchmark: 90
*Orr removed

1971-72:

Old:
2. Orr - 117

New:
2. Ratelle - 109
*Orr removed

1973-74:

Old:
1. Esposito - 145
2. Orr - 122
3. Hodge - 105
4. Cashman - 89
5. Clarke - 87
6. Martin - 86
...average as benchmark: 106

New:
1. [various players] - 91
*Orr removed

1974-75:

Old:
2. Esposito - 127

New:
2. Dionne - 121
*Orr removed

1983-84:

Old:
2. Coffey - 126

New:
2. Goulet - 121
*Coffey removed

----------------------------------------------------------

HO, could you run the VsX numbers with these adjustments (keeping the weights for the moment so we can compare like-to-like) so we can see how they affect the career numbers of the various non-Bruins who had peak years in the seasons where we have modified benchmarks? Thanks in advance.