Relative Strength of recent Art Ross wins (using PPG)

[daver]

For each season, I measured the PPG of the winner vs. the #10 PPG, #25 PPG and the #50 PPG (min. 40 games) and averaged those three.

For example, Kucherov's 23/24 PPG was 1.78, a 146 points/82 game pace. That was 1.402 times better than #10, 1.648 times better than #25, and 1.935 time better than #50.

Here is how the most recent winners' pts/82 game pace look against Kucherov's 146 point as the base comparable. E.g. each winner's PPG vs. the #10, #25 and #50 for each season in comparison to Kucherov's.


McDavid - 22/23: 153

McDavid - 21/22: 132

McDavid - 20/21: 169

Draisaitl - 19/20: 136

Kucherov - 18/19: 133

McDavid - 17/18: 119

McDavid - 16/17: 128

Kane - 15/16: 134

Crosby - 13/14: 131


Notable others

McDavid 23/24: 143

MacKinnon - 23/24: 140

Draisaitl 22/23 - 128

MacKinnon - 22/23: 125

Crosby - 12/13: 152

Malkin - 11/12: 144

Crosby - 10/11: 169

Sedin 09/10: 127

Ovechkin - 09/10: 140

Malkin - 08/09: 128

Ovechkin - 07/08: 124

Crosby - 06/07: 126

Thornton - 05/06: 128


Comment

I don't think there are too many surprises here in terms of level of performance vs. direct peers. McDavid has the best full 82 season Art Ross win during his career while Ovechkin and Malkin have the best almost full seasons while Crosby has the best partial seasons during the "Big 3" era.

Once you line up recent Art Ross winners and other notables with the early post lockout winners and the two dominant DPE 2.0 winners there does seem to be a bias towards the recent winners. As a group, Kucherov, Draisaitl and MacKinnon seemingly would have been winning more Rosses than Malkin, Crosby and Ovechkin.

McDavid - 22/23: 153

Kucherov - 23/24: 146

Malkin - 11/12: 144

McDavid 23/24: 143

MacKinnon - 23/24: 140

Ovechkin - 09/10: 140

Draisaitl - 19/20: 136

Kane - 15/16: 134

Crosby - 13/14: 131

Kucherov - 18/19: 133

McDavid - 21/22: 132

Draisaitl 22/23 - 128

Malkin - 08/09: 128

Thornton - 05/06: 128

Sedin - 09/10: 127

Crosby - 06/07: 126

MacKinnon - 22/23: 125

Ovechkin - 07/08: 124


The bias is reduced if you remove the #25 and #50 scorers from the calculation. There were more PP points being scored in the early post lockout which elevated the lower Top 50 PPGs.


McDavid - 22/23: 153

Malkin - 11/12: 153

Kucherov - 23/24: 146

McDavid 23/24: 143

MacKinnon - 23/24: 140

Kane - 15/16: 140

Ovechkin - 09/10: 138

Malkin - 08/09: 135

Draisaitl - 19/20: 134

Crosby - 13/14: 134

Kucherov - 18/19: 133

Crosby - 06/07: 133

Draisaitl 22/23 - 131

Thornton - 05/06: 131

MacKinnon - 22/23: 128

Sedin - 09/10: 125

Ovechkin - 07/08: 124

McDavid - 21/22: 122

Click to read more...

 

[Midnight Judges]

A thread about the Art Ross gets distorted into yet another "pretending for Crosby" exercise.

 

[The Macho King]

I frankly don't like adjusting for PPG - especially at a league-wide level. Even over a half season, I feel that PPG ends up with a lot more noise than final totals do.

 

[WarriorofTime]

Why look at PPG? Art Ross is a counting award.

 

[daver]

I frankly don't like adjusting for PPG - especially at a league-wide level. Even over a half season, I feel that PPG ends up with a lot more noise than final totals do.

Adjusting in what sense? Do you mean using PPG instead of raw points totals? This is basically VsX using more data; not using league GPG.

 

[WarriorofTime]

Adjusting in what sense? Do you mean using PPG instead of raw points totals? This is basically VsX using more data; not using league GPG.

Unfortunately, it's pretty obvious that this is just the latest round of "treat Crosby's partial seasons as full seasons" in a not so subtle fashion.

 

[MadLuke]

I was not able to follow why PPG when it is about the strenght of an Art Ross wins ? AS mentioned it is a pure compiling award that is quite about playing the games.

Seem more a relative strength of PPG season among at least 40 games played ?

 

[Vilica]

Rk	Pts Rank	Name	Year	Team	Games	Goals	Assists	Points	Team GF	LA GF	% LA	G%	P%	VsX	VsX Season		Avg VsX
1	1	Connor McDavid	20-21	EDM	56	33	72	105	183	165	1.109	0.180	0.574	69	152.17		145.17
2	1	Connor McDavid	22-23	EDM	82	64	89	153	325	258	1.260	0.197	0.471	113	135.40		135.28
3	1	Nikita Kucherov	23-24	TBL	81	44	100	144	288	253	1.138	0.153	0.500	120	120.00		129.84
5	1	Leon Draisaitl	19-20	EDM	71	43	67	110	223	208	1.072	0.193	0.493	97	113.40		120.64
6	1	Nikita Kucherov	18-19	TBL	82	41	87	128	319	244	1.307	0.129	0.401	116	110.34		119.67
9	1	Sidney Crosby	06-07	PIT	79	36	84	120	267	236	1.131	0.135	0.449	114	105.26		115.99
10	1	Joe Thornton	05-06		81	29	96	125	275	248	1.109	0.105	0.455	106	117.92		114.98
11	1	Alex Ovechkin	07-08	WAS	82	65	47	112	238	223	1.067	0.273	0.471	106	105.66		114.57
12	1	Evgeni Malkin	11-12	PIT	75	50	59	109	273	218	1.252	0.183	0.399	97	112.37		114.06
15	1	Henrik Sedin	09-10	VAN	82	29	83	112	268	227	1.181	0.108	0.418	109	102.75		112.55
16	1	Patrick Kane	15-16	CHI	82	46	60	106	234	219	1.068	0.197	0.453	89	119.10		110.41
18	1	Evgeni Malkin	08-09	PIT	82	35	78	113	258	234	1.103	0.136	0.438	110	102.73		110.16
19	1	Connor McDavid	21-22	EDM	80	44	79	123	285	255	1.118	0.154	0.432	115	106.96		110.03
20	2	Alex Ovechkin	09-10	WAS	72	50	59	109	313	227	1.379	0.160	0.348	109	100.00		109.54
24	1	Sidney Crosby	13-14	PIT	80	36	68	104	242	219	1.105	0.149	0.430	87	119.54		108.33
26	1	Martin St. Louis	12-13	TBL	48	17	43	60	147	127	1.157	0.116	0.408	57	105.26		107.77
29	1	Daniel Sedin	10-11	VAN	82	41	63	104	258	224	1.152	0.159	0.403	99	105.05		105.91
36	1	Connor McDavid	17-18	EDM	82	41	67	108	229	240	0.954	0.179	0.472	102	105.88		102.65
38	1	Connor McDavid	16-17	EDM	82	30	70	100	243	223	1.090	0.123	0.412	89	112.36		102.29
47	3	Jarome Iginla	07-08	CGY	82	50	48	98	226	223	1.013	0.221	0.434	106	92.45		100.25
48	3	Nikita Kucherov	22-23	TBL	82	30	83	113	280	258	1.085	0.107	0.404	113	100.00		99.91
117	1	Jamie Benn	14-15	DAL	82	35	52	87	257	218	1.179	0.136	0.339	86	101.16		91.04

Here's all 19 Art Ross winners, plus a couple extra ones (Ovechkin's 09-10 and the 2 seasons nearest to 100), sorted by Average VsX. Recent seasons have peaked slightly higher than the original generational years post-lockout, but the Ovechkin 09-10 season and Malkin 11-12 season show the power of playing all 82. Both of them playing to their season averages, playing all 82 games, would've ended at an Average VsX of around 125, moving them into the top 5 of the list. Even Crosby's 06-07, where he missed 3 games (and lost out on maybe 5 points), would erase the gap between Kucherov's 18-19 and his season. But because of the missed games, they peaked in the 115 area instead.

When looking at P% and %LA, most of these seasons are in the mid 40% amongst the years, and how many goals their team scored in relation to league average dictates their point totals. Take the 2 McDavid years of 17-18 and 22-23, by P% he had just over 47% of his team's goals both years. The difference is that the Oilers scored nearly 100 goals more in 22-23, and that led to an additional 45 points for McDavid. You can get into the nitty-gritty of his EV points versus PP points, and where those points came from, but that's more for forecasting future seasons.

In terms of Art Ross winning seasonal strength, there's a huge gap between McDavid's 16-17 and Benn's 14-15. Overall, that McDavid year is 38th of all post-lockout seasons, while Benn's year is 117th. The Iginla/Kucherov (and Pastrnak 22-23, as that was tied in points with Kucherov) benchmark of 100 settle in at 47-49, or basically the 50 best seasons (in terms of points) in the 19 years post-lockout qualify as above 100. It also shows the clustering - you have ~50 seasons above 100, and ~75 seasons between 90-100.

Finally, by setting the VsX by averaging seasons instead of leaving it to the vagaries of seasonal scoring, the gaps among seasons actually mean something. Instead of Ovechkin's 09-10, Kucherov's 22-23, and Tavares' 14-15 all being exactly 100 as the VsX benchmarks of their respective years, their Average VsX of 109.54, 99.91 and 89.99 represent the 10 and 20 point gap that's a more accurate reflection of their relative dominance.

 

[daver]

Rk	Pts Rank	Name	Year	Team	Games	Goals	Assists	Points	Team GF	LA GF	% LA	G%	P%	VsX	VsX Season		Avg VsX
1	1	Connor McDavid	20-21	EDM	56	33	72	105	183	165	1.109	0.180	0.574	69	152.17		145.17
2	1	Connor McDavid	22-23	EDM	82	64	89	153	325	258	1.260	0.197	0.471	113	135.40		135.28
3	1	Nikita Kucherov	23-24	TBL	81	44	100	144	288	253	1.138	0.153	0.500	120	120.00		129.84
5	1	Leon Draisaitl	19-20	EDM	71	43	67	110	223	208	1.072	0.193	0.493	97	113.40		120.64
6	1	Nikita Kucherov	18-19	TBL	82	41	87	128	319	244	1.307	0.129	0.401	116	110.34		119.67
9	1	Sidney Crosby	06-07	PIT	79	36	84	120	267	236	1.131	0.135	0.449	114	105.26		115.99
10	1	Joe Thornton	05-06		81	29	96	125	275	248	1.109	0.105	0.455	106	117.92		114.98
11	1	Alex Ovechkin	07-08	WAS	82	65	47	112	238	223	1.067	0.273	0.471	106	105.66		114.57
12	1	Evgeni Malkin	11-12	PIT	75	50	59	109	273	218	1.252	0.183	0.399	97	112.37		114.06
15	1	Henrik Sedin	09-10	VAN	82	29	83	112	268	227	1.181	0.108	0.418	109	102.75		112.55
16	1	Patrick Kane	15-16	CHI	82	46	60	106	234	219	1.068	0.197	0.453	89	119.10		110.41
18	1	Evgeni Malkin	08-09	PIT	82	35	78	113	258	234	1.103	0.136	0.438	110	102.73		110.16
19	1	Connor McDavid	21-22	EDM	80	44	79	123	285	255	1.118	0.154	0.432	115	106.96		110.03
20	2	Alex Ovechkin	09-10	WAS	72	50	59	109	313	227	1.379	0.160	0.348	109	100.00		109.54
24	1	Sidney Crosby	13-14	PIT	80	36	68	104	242	219	1.105	0.149	0.430	87	119.54		108.33
26	1	Martin St. Louis	12-13	TBL	48	17	43	60	147	127	1.157	0.116	0.408	57	105.26		107.77
29	1	Daniel Sedin	10-11	VAN	82	41	63	104	258	224	1.152	0.159	0.403	99	105.05		105.91
36	1	Connor McDavid	17-18	EDM	82	41	67	108	229	240	0.954	0.179	0.472	102	105.88		102.65
38	1	Connor McDavid	16-17	EDM	82	30	70	100	243	223	1.090	0.123	0.412	89	112.36		102.29
47	3	Jarome Iginla	07-08	CGY	82	50	48	98	226	223	1.013	0.221	0.434	106	92.45		100.25
48	3	Nikita Kucherov	22-23	TBL	82	30	83	113	280	258	1.085	0.107	0.404	113	100.00		99.91
117	1	Jamie Benn	14-15	DAL	82	35	52	87	257	218	1.179	0.136	0.339	86	101.16		91.04

Here's all 19 Art Ross winners, plus a couple extra ones (Ovechkin's 09-10 and the 2 seasons nearest to 100), sorted by Average VsX. Recent seasons have peaked slightly higher than the original generational years post-lockout, but the Ovechkin 09-10 season and Malkin 11-12 season show the power of playing all 82. Both of them playing to their season averages, playing all 82 games, would've ended at an Average VsX of around 125, moving them into the top 5 of the list. Even Crosby's 06-07, where he missed 3 games (and lost out on maybe 5 points), would erase the gap between Kucherov's 18-19 and his season. But because of the missed games, they peaked in the 115 area instead.

When looking at P% and %LA, most of these seasons are in the mid 40% amongst the years, and how many goals their team scored in relation to league average dictates their point totals. Take the 2 McDavid years of 17-18 and 22-23, by P% he had just over 47% of his team's goals both years. The difference is that the Oilers scored nearly 100 goals more in 22-23, and that led to an additional 45 points for McDavid. You can get into the nitty-gritty of his EV points versus PP points, and where those points came from, but that's more for forecasting future seasons.

In terms of Art Ross winning seasonal strength, there's a huge gap between McDavid's 16-17 and Benn's 14-15. Overall, that McDavid year is 38th of all post-lockout seasons, while Benn's year is 117th. The Iginla/Kucherov (and Pastrnak 22-23, as that was tied in points with Kucherov) benchmark of 100 settle in at 47-49, or basically the 50 best seasons (in terms of points) in the 19 years post-lockout qualify as above 100. It also shows the clustering - you have ~50 seasons above 100, and ~75 seasons between 90-100.

Finally, by setting the VsX by averaging seasons instead of leaving it to the vagaries of seasonal scoring, the gaps among seasons actually mean something. Instead of Ovechkin's 09-10, Kucherov's 22-23, and Tavares' 14-15 all being exactly 100 as the VsX benchmarks of their respective years, their Average VsX of 109.54, 99.91 and 89.99 represent the 10 and 20 point gap that's a more accurate reflection of their relative dominance.

Thanks for the data.

So the Average VsX (VsX in brackets) ranks the seasons this way:

McDavid 22/23 - 135 (135)
Kucherov 23/24 - 130 (120)

vs. as noted in the OP (PPG vs. #10/#25/#50)

McDavid 153 (or 149 using raw point totals, not PPG)
Kucherov 146 (or 144 using raw point totals, not PPG)

The "VsX" gap seems too big, my method seems to pass the raw numbers statistical smell test, both for PPG and raw numbers, given a cursory look at league GPG, the " Average VsX" seemingly factors in League GF and point % of team's goals into the calculation as it gets them as close.

Can I trouble you to show me how the "Average VsX" is calculated for those two seasons. e.g. how Kucherov's 144 translates to 130. In getting to the 130, are the p% for the other players used in the calculation factored in?

 

[daver]

Recent seasons have peaked slightly higher than the original generational years post-lockout, but the Ovechkin 09-10 season and Malkin 11-12 season show the power of playing all 82. Both of them playing to their season averages, playing all 82 games, would've ended at an Average VsX of around 125, moving them into the top 5 of the list.

There does seem to be some bias towards recent seasons as Malkin's 11/12 season especially was as dominant than Kuckerov's using vs. #10/#25/#50.

I am wondering if this bias has something to do with p%. You say that his Average VsX would be around 125 if he played a full season. What is the P% in that calculation? It would presumably increase from 0.399.

Further to that, shouldn't Malkin's 11/12 P% (and Ovechkin's too) reflect their average per game P%, not over 82 games. This seems to be be biased towards players who played more games.

If VsX is supposed to be about raw point totals and not PPG, then, IMO, P% should exclude the games they missed. I.e. the current calculation takes the "they didn't play all the games" too far.

 

[daver]

Using raw points instead of PPG (in brackets), here are the numbers for a few notables :

McDavid 22/23 - 149 (153)
Kucherov 23/24 - 144 (146)
MacKinnon 23/24 - 140 (140)
Kane 15/16 - 132 (134)
Malkin 11/12 - 130 (144)
Kucherov 18/19 - 128 (133)
Crosby 13/14 - 126 (131)
McDavid 21/22 - 123 (132)

Kucherov and MacKinnon get closer to 22/23 McDavid. Malkin takes the expected significant drop which, IMO, undervalues his level of play.

 

[klefbombs shoulder]

Why exactly does Crosby's 12-13 season get considered when he played 36 games. I thought the cut off was 40 GP? Also Draisaitl's 2020-21 season is a "notable other", he scored 84 points in 56 games, 3rd place had 69 points.

 

[Vilica]

There does seem to be some bias towards recent seasons as Malkin's 11/12 season especially was as dominant than Kuckerov's using vs. #10/#25/#50.

I am wondering if this bias has something to do with p%. You say that his Average VsX would be around 125 if he played a full season. What is the P% in that calculation? It would presumably increase from 0.399.

Further to that, shouldn't Malkin's 11/12 P% (and Ovechkin's too) reflect their average per game P%, not over 82 games. This seems to be be biased towards players who played more games.

If VsX is supposed to be about raw point totals and not PPG, then, IMO, P% should exclude the games they missed. I.e. the current calculation takes the "they didn't play all the games" too far.

I don't have the Malkin numbers off-hand, but I do have some Ovechkin 09-10 numbers:

	Name	Year	Team	Games	Goals	Assists	Points	Team GF	LA GF	% LA	G%	P%	VsX	VsX Season		Avg VsX
2	Alex Ovechkin	09-10	WAS	72	50	59	109	313	227	1.379	0.160	0.348	109	100.00		109.54
2	Alex Ovechkin	09-10	WAS	72	50	59	109	268	199	1.347	0.187	0.407	109	100.00		124.95
1	Alex Ovechkin	09-10	WAS	82	59	68	127	313	227	1.379	0.188	0.406	109	116.51		127.62

The Capitals scored 45 goals in the 10 games he missed, and thus you have to back out 10 games worth of league average, and then you get into questions if that is accurate, or if you should use the actual totals from those 10 games, or if you should hold % LA constant, and then it becomes a series of dominos adjusting every player, which is too much work for this. At the beginning of a season, every player has the opportunity to play all 82 games, and their actual results are what should govern this stat, not their theoretical maximum.

To give a couple of examples, in the 09-10 season, Ovechkin averaged just over 1.5 points per game. If after game 80, with 2 games left, he's sitting at 58 goals and 66 assists, in theory, he'll score 1 goal and have 2 assists in the final 2 games for 127 points. In practice, he's scoring 60 goals. As another example, Crosby missed 1 game in 09-10, against Chicago, a 2-1 loss. That reduces his Team GF from 249 to 248, and in theory the LA GF from 227 to 224, which raises his Average VsX from 109.54 to 111.00. But when you think about it, why is he getting a benefit from not playing a game? Pittsburgh could probably really have used him in that game against Chicago, and his absence most likely contributed to the loss in OT.

I have done some work with backing out games missed for seasons, but only for that more accurate G%/P% number. A lot of the time, the variance in player numbers is simply missed games. A player's general G%/P% tends to be similar season-over-season. The other aspect is that a 5 point gap in Average VsX really isn't that big, especially over 100. Outlier seasons generally are marked by positive variance in the randomness categories of a particular player-season, and there are so many that trying to adjust for them is an exercise in futility. It's more important to show an ability to put up a season in the 120s rather than knowing a particular season was 124.95 or 127.62 by VsX.

In terms of Average VsX calculations, it is very simple.

Year	LA GF	VsX	New GF	Goals	Assists	Points	VsX Season	Year	LA GF	VsX	New GF	Goals	Assists	Points	VsX Season
05-06	248	106	282.31	43.13	98.02	141.15	133.16	05-06	248	106	312.40	61.52	85.55	147.07	138.75
06-07	236	114	268.65	41.04	93.28	134.32	117.83	06-07	236	114	297.29	58.54	81.41	139.95	122.77
07-08	223	106	253.85	38.78	88.14	126.92	119.74	07-08	223	106	280.91	55.32	76.93	132.24	124.76
08-09	234	110	266.37	40.70	92.49	133.19	121.08	08-09	234	110	294.77	58.05	80.72	138.77	126.15
09-10	227	109	258.40	39.48	89.72	129.20	118.53	09-10	227	109	285.95	56.31	78.31	134.62	123.50
10-11	224	99	254.99	38.96	88.54	127.49	128.78	10-11	224	99	282.17	55.57	77.27	132.84	134.18
11-12	218	97	248.16	37.91	86.17	124.08	127.92	11-12	218	97	274.61	54.08	75.20	129.28	133.28
12-13	127	57	144.57	22.09	50.20	72.28	126.82	12-13	127	57	159.98	31.50	43.81	75.31	132.13
13-14	219	87	249.30	38.09	86.56	124.65	143.27	13-14	219	87	275.87	54.33	75.55	129.87	149.28
14-15	218	86	248.16	37.91	86.17	124.08	144.28	14-15	218	86	274.61	54.08	75.20	129.28	150.32
15-16	219	89	249.30	38.09	86.56	124.65	140.05	15-16	219	89	275.87	54.33	75.55	129.87	145.92
16-17	223	89	253.85	38.78	88.14	126.92	142.61	16-17	223	89	280.91	55.32	76.93	132.24	148.59
17-18	240	102	273.20	41.74	94.86	136.60	133.92	17-18	240	102	302.33	59.53	82.79	142.33	139.53
18-19	244	116	277.75	42.43	96.44	138.88	119.72	18-19	244	116	307.36	60.53	84.17	144.70	124.74

The left player is Kucherov's 23-24, the right player is McDavid's 22-23. Here's their season stat lines, as a reminder.

	Name	Year	Team	Games	Goals	Assists	Points	Team GF	LA GF	% LA	G%	P%	VsX	VsX Season		Avg VsX
1	Connor McDavid	22-23	EDM	82	64	89	153	325	258	1.260	0.197	0.471	113	135.40		135.28
1	Nikita Kucherov	23-24	TBL	81	44	100	144	288	253	1.138	0.153	0.500	120	120.00		129.84

It's much easier to demonstrate Kucherov's results because his P% is exactly 50. Essentially, you take all those % items, LA, G%, and P%, and transpose them to the 05-06 season, where league average was 248. Thus, 13.8% above league average becomes 282.31 goals, and 50% of that is 141.15. Then, you compare that point total to the actual VsX of the season, 106, and Kucherov's VsX for the 05-06 year is 133.16. Repeat that for 06-07, and so on, until you have the VsX numbers for all 14 years, and then just take an average of those to create the final number (the actual calculation is 1817.71/14 for Kucherov compared to 1893.9/14 for McDavid).

[I created this in 2021, and thought the disruption of the pandemic was enough of a shock to not want to use the 19-20 season at the time, though I'm guessing Average VsX would be marginally more accurate if I had continued on with subsequent seasons. Rounded league average is a bit of a blunt instrument though, as that 253 number is essentially 8086 goals by 32 teams, or 252.68. Just 7 less goals in the 1312 games of the season would be 252.47, which would get rounded down to 252, and Kucherov's Average VsX would increase from 129.84 to 130.35. To get down to 243, which would raise Kucherov's Average VsX to the 135 level, requires just 295 less goals, or somewhere between a fifth and a quarter of a goal a game. (As an example, there were 446 empty-net goals in 23-24, so that would cover the entire decrease, so then you'd have to back out all empty net points, thus MacKinnon would have 132, Kucherov 130, McDavid 127, but then that's not the stats players are targeting. Kucherov's play and deployment changes if he's 2 points behind MacKinnon instead of 4 points ahead.) That's why points are points, league average is league average, and the only real adjustments are some Boston teams in the 70s. They are so far above league average that I adjusted their league average to be the average of just the O6 teams, not all the teams. I've also spent way too much time on this post, so out it goes into the world.]

 

[daver]

Why exactly does Crosby's 12-13 season get considered when he played 36 games. I thought the cut off was 40 GP?

40 games is 50% in an 82 game season, 12/13 was a 48 game season.

I don't have the Malkin numbers off-hand, but I do have some Ovechkin 09-10 numbers:

	Name	Year	Team	Games	Goals	Assists	Points	Team GF	LA GF	% LA	G%	P%	VsX	VsX Season		Avg VsX
2	Alex Ovechkin	09-10	WAS	72	50	59	109	313	227	1.379	0.160	0.348	109	100.00		109.54
2	Alex Ovechkin	09-10	WAS	72	50	59	109	268	199	1.347	0.187	0.407	109	100.00		124.95
1	Alex Ovechkin	09-10	WAS	82	59	68	127	313	227	1.379	0.188	0.406	109	116.51		127.62

The Capitals scored 45 goals in the 10 games he missed, and thus you have to back out 10 games worth of league average, and then you get into questions if that is accurate, or if you should use the actual totals from those 10 games, or if you should hold % LA constant, and then it becomes a series of dominos adjusting every player, which is too much work for this. At the beginning of a season, every player has the opportunity to play all 82 games, and their actual results are what should govern this stat, not their theoretical maximum.

To give a couple of examples, in the 09-10 season, Ovechkin averaged just over 1.5 points per game. If after game 80, with 2 games left, he's sitting at 58 goals and 66 assists, in theory, he'll score 1 goal and have 2 assists in the final 2 games for 127 points. In practice, he's scoring 60 goals. As another example, Crosby missed 1 game in 09-10, against Chicago, a 2-1 loss. That reduces his Team GF from 249 to 248, and in theory the LA GF from 227 to 224, which raises his Average VsX from 109.54 to 111.00. But when you think about it, why is he getting a benefit from not playing a game? Pittsburgh could probably really have used him in that game against Chicago, and his absence most likely contributed to the loss in OT.

I have done some work with backing out games missed for seasons, but only for that more accurate G%/P% number. A lot of the time, the variance in player numbers is simply missed games. A player's general G%/P% tends to be similar season-over-season. The other aspect is that a 5 point gap in Average VsX really isn't that big, especially over 100. Outlier seasons generally are marked by positive variance in the randomness categories of a particular player-season, and there are so many that trying to adjust for them is an exercise in futility. It's more important to show an ability to put up a season in the 120s rather than knowing a particular season was 124.95 or 127.62 by VsX.

In terms of Average VsX calculations, it is very simple.

Year	LA GF	VsX	New GF	Goals	Assists	Points	VsX Season	Year	LA GF	VsX	New GF	Goals	Assists	Points	VsX Season
05-06	248	106	282.31	43.13	98.02	141.15	133.16	05-06	248	106	312.40	61.52	85.55	147.07	138.75
06-07	236	114	268.65	41.04	93.28	134.32	117.83	06-07	236	114	297.29	58.54	81.41	139.95	122.77
07-08	223	106	253.85	38.78	88.14	126.92	119.74	07-08	223	106	280.91	55.32	76.93	132.24	124.76
08-09	234	110	266.37	40.70	92.49	133.19	121.08	08-09	234	110	294.77	58.05	80.72	138.77	126.15
09-10	227	109	258.40	39.48	89.72	129.20	118.53	09-10	227	109	285.95	56.31	78.31	134.62	123.50
10-11	224	99	254.99	38.96	88.54	127.49	128.78	10-11	224	99	282.17	55.57	77.27	132.84	134.18
11-12	218	97	248.16	37.91	86.17	124.08	127.92	11-12	218	97	274.61	54.08	75.20	129.28	133.28
12-13	127	57	144.57	22.09	50.20	72.28	126.82	12-13	127	57	159.98	31.50	43.81	75.31	132.13
13-14	219	87	249.30	38.09	86.56	124.65	143.27	13-14	219	87	275.87	54.33	75.55	129.87	149.28
14-15	218	86	248.16	37.91	86.17	124.08	144.28	14-15	218	86	274.61	54.08	75.20	129.28	150.32
15-16	219	89	249.30	38.09	86.56	124.65	140.05	15-16	219	89	275.87	54.33	75.55	129.87	145.92
16-17	223	89	253.85	38.78	88.14	126.92	142.61	16-17	223	89	280.91	55.32	76.93	132.24	148.59
17-18	240	102	273.20	41.74	94.86	136.60	133.92	17-18	240	102	302.33	59.53	82.79	142.33	139.53
18-19	244	116	277.75	42.43	96.44	138.88	119.72	18-19	244	116	307.36	60.53	84.17	144.70	124.74

The left player is Kucherov's 23-24, the right player is McDavid's 22-23. Here's their season stat lines, as a reminder.

	Name	Year	Team	Games	Goals	Assists	Points	Team GF	LA GF	% LA	G%	P%	VsX	VsX Season		Avg VsX
1	Connor McDavid	22-23	EDM	82	64	89	153	325	258	1.260	0.197	0.471	113	135.40		135.28
1	Nikita Kucherov	23-24	TBL	81	44	100	144	288	253	1.138	0.153	0.500	120	120.00		129.84

It's much easier to demonstrate Kucherov's results because his P% is exactly 50. Essentially, you take all those % items, LA, G%, and P%, and transpose them to the 05-06 season, where league average was 248. Thus, 13.8% above league average becomes 282.31 goals, and 50% of that is 141.15. Then, you compare that point total to the actual VsX of the season, 106, and Kucherov's VsX for the 05-06 year is 133.16. Repeat that for 06-07, and so on, until you have the VsX numbers for all 14 years, and then just take an average of those to create the final number (the actual calculation is 1817.71/14 for Kucherov compared to 1893.9/14 for McDavid).

[I created this in 2021, and thought the disruption of the pandemic was enough of a shock to not want to use the 19-20 season at the time, though I'm guessing Average VsX would be marginally more accurate if I had continued on with subsequent seasons. Rounded league average is a bit of a blunt instrument though, as that 253 number is essentially 8086 goals by 32 teams, or 252.68. Just 7 less goals in the 1312 games of the season would be 252.47, which would get rounded down to 252, and Kucherov's Average VsX would increase from 129.84 to 130.35. To get down to 243, which would raise Kucherov's Average VsX to the 135 level, requires just 295 less goals, or somewhere between a fifth and a quarter of a goal a game. (As an example, there were 446 empty-net goals in 23-24, so that would cover the entire decrease, so then you'd have to back out all empty net points, thus MacKinnon would have 132, Kucherov 130, McDavid 127, but then that's not the stats players are targeting. Kucherov's play and deployment changes if he's 2 points behind MacKinnon instead of 4 points ahead.) That's why points are points, league average is league average, and the only real adjustments are some Boston teams in the 70s. They are so far above league average that I adjusted their league average to be the average of just the O6 teams, not all the teams. I've also spent way too much time on this post, so out it goes into the world.]

Thanks for the response.

I may be too conservative but I generally take P% as a tiebreaker for players with similar production. The same with the gap between a player and their teammates. Kucherov was likely downgraded before this season vs. McDavid given how much more McDavid (with Drai) was a one man/two man show for the Oilers. This season Kucherov had the most impressive season of the two based on P% and gap between him and the #2 scorer on his team.

This qualifies a narrative that I promote that "great players produce regardless" which renders things like P% to the tiebreaker area and not something that moves a player up or down a clear statistical tier.

As you mention, there isn't much difference between players who are within 5 points of the Average VsX so the difference between McDavid's 22/23 and Kucherov's 23/24 is not very big or affected a lot by P%/team scoring.

One thing to consider is the affect that significant changes to the scoring environment can have on the elite scorers. I have started a thread on that topic: Scoring levels of the "pack" over the past 30 years in relation to League GPG

It was more difficult for superstar talent to separate themselves from the pack during lower scoring seasons.