Top Offensive Players Since 2005, Using an Objective Scoring System

[Midnight Judges]

The purpose of this thread is to look at the top 30 offensive players (all forwards) since 2005 using an objective scoring system, compare how those results differ from the traditional point system, and discuss which metric grades better against the smell test.

The traditional system used in hockey (1 goal = 1 assist = 1 secondary assist) has no statistical basis. It was created over a period of decades in which the NHL waffled on how many assists to count and when to count them. There is wild inconsistency in the NHL for how to count them – even varying significantly from arena to arena. Today the vast majority of hockey fans are utterly unaware of how we arrived with the traditional system and how inconsistent it is, yet they take these arbitrary numbers as gospel without the slightest statistical justification.

The valuations resulting from the analysis in the linked article (1/.54/.10 for a goal/primary assist/secondary assist respectively) are utilized for the purposes of this thread. The methodology can all be accessed in this link:

What is the objective value of an assist?

This methodology appears to be sound and also consistent with other statistical analyses. Critiques of this methodology are welcome. I’ve yet to encounter any convincing rationales opposed to it.

In this thread Offensive statistic charts are provided in 5 formats:

1. Regular Season 2005-Present Offensive Value
2. Regular Season 2005-Present Per Game Offensive Value
3. Regular Season 2005-Present Per Season Offensive Value
4. Playoffs 2005-Present Offensive Value
5. Playoffs 2005-Present Per Game Offensive Value

NHL.com does not provide secondary assist data prior to the 2009-2010 season so I painstakingly got that data from hockey reference scoring logs. Please note this means there is a slight chance that there is human error on these counts. If anyone notices a data problem or a miscalculation let me know and I will correct it.

Player set: I used the top players for points since 2005 mixed with the top players for goals since 2005, then threw in McDavid for the hell of it. I did not adjust for goals scored per game per season (GPGPS) because most of these players played in all the same seasons and the resulting variance would be small. Even comparing McDavid’s and Ovechkin’s average GPGPS – it’s extremely close and actually a tad higher for McDavid. This may slightly disadvantage the players who weren’t around for the highest scoring seasons – ‘05 through ‘07 or ’17 through ‘19. I could factor in that layer without much difficulty. There are some players who would have entered some of these per game lists but were not among the leaders in raw goals or points (Marc Savard, Pavel Datsyuk, Artemi Panarin).

Charts:

Rank	Player	Regular Season 2005-Present Offensive Value	Traditional Point System (Raw Points) Ranking	Difference
1	Ovechkin	864.22	2	1
2	Crosby	734.64	1	-1
3	E. Staal	631.06	5	2
4	Thornton	626.32	3	-1
5	Malkin	622.38	4	-1
6	Kane	575.68	6	0
7	Iginla	565.98	16	9
8	Marleau	547.7	13	5
9	D. Sedin	540.92	9	0
10	Vanek	533.84	17	7
11	Stamkos	532.82	21	10
12	Perry	531.2	18	6
13	Kessel	529.88	15	2
14	Kopitar	513.48	10	-4
15	Zetterberg	512.62	11	-4
16	Nash	511	29	13
17	Pavelski	509.56	24	7
18	Parise	501.14	26	8
19	Getzlaf	499.24	8	-11
20	Spezza	496.08	14	-6
21	J. Carter	486.56	31	10
22	Hossa	484.62	27	5
23	Toews	481.56	25	2
24	Bergeron	478.18	19	-5
25	Tavares	466.24	30	5
26	St. Louis	456.16	20	-6
27	Kovalchuk	449.84	39	12
28	Backstrom	447.88	12	-16
29	H. Sedin	444.36	7	-22
30	Giroux	415.76	23	-7
31	McDavid	221.04	Meh	Meh

[TBODY] [/TBODY]

A pair of Swedes take a hit under this system – Nick Backstrom and Henrik Sedin no longer receive equal credit for the times when Alex Ovechkin and Daniel Sedin score goals.

Rank	Player	Regular Season 2005-Present Per Game Offensive Value	Traditional Point System PPG (Min 270 GP)	Difference
1	Ovechkin	0.797250923	4	3
2	Crosby	0.779045599	2	0
3	McDavid	0.770174216	1	-2
4	Malkin	0.730492958	3	-1
5	Stamkos	0.714235925	9	4
6	Kane	0.63751938	6	0
7	Tavares	0.620825566	17	10
8	Iginla	0.609892241	38	30
9	Thornton	0.592544939	11	2
10	St. Louis	0.592415584	10	0
11	E. Staal	0.57736505	37	26
12	Toews	0.55161512	36	24
13	D. Sedin	0.545832492	25	12
14	Perry	0.537651822	63	49
15	Kessel	0.532008032	52	37
16	Parise	0.5297463	60	44
17	Pavelski	0.52913811	59	42
18	Spezza	0.52	23	5
19	Vanek	0.518794947	73	54
20	Kopitar	0.511944167	20	0
21	Kovalchuk	0.511181818	13	-8
22	Getzlaf	0.507357724	18	-4
23	Giroux	0.50702439	19	-4
24	Backstrom	0.500424581	15	-9
25	Marleau	0.498362147	76	51
26	J. Carter	0.496489796	95	69
27	Nash	0.482075472	65	38
28	Zetterberg	0.473770795	20	-8
29	Bergeron	0.465155642	58	29
30	H. Sedin	0.439090909	21	-9
31	Hossa	0.370221543	30	-1

[TBODY] [/TBODY]

Per game stats show more variances vs points in terms of difference from the traditional points system. The variances highlighted here ring true to me. Toews jumps tremendously from 36th to 12th. The hockey media knows he is a more impactful player than the traditional points system indicates but instead of reaching to the obvious (the flaws with the traditional point system) they’ve attempted to justify his success with defensive play. Same goes for Jerome Iginla, who has been cursed with being on bad teams – something hockey fans and media are incredibly bad at analyzing. The traditional system arbitrarily punishes players like Iginla/Kovalchuk who get stuck on bad teams - because the players around them score less and they get fewer secondary assists.

Other players jump tremendously – Marleau, Carter, Perry – players you know are impactful and yet their regular season PPG numbers have them at 76, 95, and 73 respectively. This is a red flag for the traditional system. 72 players had more offensive impact per game than Hart winner Corey Perry through all his best years? I doubt it.

The names at the top of the list are not surprising. Alex Ovechkin is the most valuable offensive player on a per game basis. Crosby comes in second. Surprisingly McDavid is less valuable on a per game basis than Ovechkin and Crosby’s career average despite them both being in their 30s. It highlights how extraordinary these two players are to a clearer degree than the traditional point system. McDavid has tons of work to do if he wants to be mentioned with them.

Malkin, Stamkos, Kane, Tavares rounding out these leaders feels correct to me. Stamkos’s regular season value is tempered by his playoff failures.

Rank	Player	Regular Season 2005-Present Per Reg. Season Offensive Value
1	Ovechkin	61.73
2	McDavid	55.26
3	Crosby	52.47428571
4	Kovalchuk	49.98222222
5	Stamkos	48.43818182
6	Kane	47.97333333
7	Malkin	47.87538462
8	Iginla	47.165
9	Tavares	46.624
10	St. Louis	45.616
11	E. Staal	45.07571429
12	Thornton	44.73714286
13	Kessel	40.76
14	Hossa	40.385
15	Toews	40.13
16	Zetterberg	39.43230769
17	Nash	39.30769231
18	Pavelski	39.19692308
19	Marleau	39.12142857
20	D. Sedin	38.63714286
21	Vanek	38.13142857
22	Perry	37.94285714
23	Backstrom	37.32333333
24	Kopitar	36.67714286
25	Parise	35.79571429
26	Getzlaf	35.66
27	Spezza	35.43428571
28	J. Carter	34.75428571
29	Giroux	34.64666667
30	Bergeron	34.15571429
31	H. Sedin	31.74

[TBODY] [/TBODY]

This is the highly underrated per season basis – the actual annual offensive value a player provides to his team. This is the chart that separates the Gretzky’s from the Lemieux’s. Interestingly McDavid is superior to Crosby despite a lesser value on a per game basis.

This chart is also necessary because it illustrates durability and per game offensive values for players who have been in the NHL for fewer than 14 seasons. Durability is inexplicably discarded as meaningless by much of hockey fandom whereas in other sports it is logically respected and treasured (Cal Ripken, Brett Favre, etc).

Rank	Player	Playoffs 2005-Present Offensive Value
1	Crosby	110.12
2	Malkin	99.46
3	Ovechkin	86.94
4	Kane	77.54
5	Zetterberg	75.94
6	Hossa	69.04
7	Marleau	68.74
8	Getzlaf	68.62
9	Pavelski	68.16
10	Toews	66.36
11	Backstrom	62.36
12	Thornton	58.64
13	Bergeron	58.2
14	Perry	55.82
15	J. Carter	52.18
16	Kessel	48.84
17	Parise	47.7
18	Giroux	39.54
19	Spezza	38.08
20	Kopitar	36.94
21	St. Louis	35.84
22	D. Sedin	34.4
23	H. Sedin	32.02
24	Stamkos	31.72
25	Iginla	31.7
26	E. Staal	30.08
27	Vanek	27.34
28	Nash	26.96
29	Tavares	17.92
30	Kovalchuk	16.56
31	McDavid	5.84

[TBODY] [/TBODY]

Raw offensive value for playoffs 2005-present. Games played is necessary context here as the quantities vary substantially. Crosby-Malkin-Ovechkin-Kane-Zetterberg feels right to me.

Rank	Player	Playoffs 2005-Present Per Game Offensive Value
1	Ovechkin	0.67921875
2	Crosby	0.671463415
3	Malkin	0.613950617
4	Kane	0.610551181
5	Iginla	0.598113208
6	Tavares	0.578064516
7	Giroux	0.573043478
8	Kessel	0.56137931
9	Zetterberg	0.554306569
10	Getzlaf	0.54896
11	E. Staal	0.51862069
12	Toews	0.5184375
13	Kovalchuk	0.5175
14	Spezza	0.514594595
15	Pavelski	0.508656716
16	Backstrom	0.50699187
17	Parise	0.491752577
18	Marleau	0.491
19	St. Louis	0.490958904
20	D. Sedin	0.484507042
21	Perry	0.473050847
22	Kopitar	0.467594937
23	Stamkos	0.453142857
24	Bergeron	0.451162791
25	McDavid	0.449230769
26	J. Carter	0.434833333
27	H. Sedin	0.432702703
28	Thornton	0.407222222
29	Vanek	0.396231884
30	Hossa	0.336780488
31	Nash	0.302921348

[TBODY] [/TBODY]

Ovechkin is the most valuable player - offensively speaking - in the playoffs on a per game basis since 2005, with Crosby second. It's not surprising seeing as how Ovechkin is within 1 goal of Crosby in 36 fewer playoff games. The distance between them and Malkin/Kane is substantial. It shows who is generational and who is in the pack.

Again, Iginla and Kovalchuk fair well. Rick Nash and Steven Stamkos do not. A look farther back at the younger years for the Sedins, Thornton, Hossa, Iginla, etc. would be more equitable, but I had to draw the line somewhere.

Anyway, let me know what you think.

Click to read more...

 

[JoVel]

Objective you say?

 

[TooManyHumans]

Somehow I knew the conclusion that would be reached when I saw the thread title and author. Lol.

 

[Trap Jesus]

Objective you say?

Objective: make Ovechkin look better

 

[Filthy Dangles]

This seems to undervalue playmakers who get a greater ratio of points via assists

Also you have the weight of a 2A as .10 while
Your linked article has it as .19

Also knowing the OP, it’s hard to believe this objective as much as it is agenda driven

 

[K Fleur]

“Objective”- MJ

 

[Tad Mikowsky]

Cmon OP, you’re not at all objective.

 

[traparatus]

Thanks for putting this together, it's a lot of work.

Point system in the NHL always struck me as arbitrary. Why two assists? Why not one or three? Obviously, lumping all of those things together into a single stat and calling it POINTS is random as hell.

 

[Midnight Judges]

Also you have the weight of a 2A as .10 while
Your linked article has it as .19

You're not correct:

If I was to make a formula, I think until further proof manifests, it would be prudent to value them at 10% or less at least when evaluating a forwards offensive production.

 

[JD1]

There is no objective way of looking at scoring stats and assigning a consistent value to a goal, a primary assist and a secondary assist. Any attempt to do that lacks an understanding of the game.

A goal scorer can score a goal on a primarily self generated play.

A primary assist is awarded to a player that makes an end to end rush, goes by two guys in the neutral, pulls a how do you do on a D man and dishes cross ice to a guy for a tap in.

A secondary assist is awarded to a guy that fights an extended battle in the corner against two opponents, comes out with the puck, makes a short dish to a teammate who immediately goes cross crease to an open team mate for a tap in goal.

3 goals. 3 different players led the effort to score. You need to watch and understand the game to know who contributed. Reading stats from 15 years ago doesn't do much.

I dont know you OP or your posting history but it seems the boys are calling you out as a OV lover. I've seen OV's team mates, a couple of hundred times, control a puck for 30 seconds in the O zone,dish it to OV for the one timer and he contributed a millisecond to zone possession. Good for him. He scored. Gets the fame and the big bucks. But it seems you're trying hard to make a case where none exists.

 

[NOTENOUGHJTCGOALS]

Any adjusted point metric that has Crosby above Giroux is not worth reading.

 

[BoredBrandonPridham]

You’re conflating PP points with 5v5 points here which is a glaring issue. Your list is basically ordered by PP time.

It would be better off splitting them into PP and 5v5, and also using p60 given there were almost double the PP opportunities circa 2005 as there are presently.

 

[OppositeLocK]

I dunno how objective this is. I know not every goal is the same, or even assist.

This is why the eye test and watching the actual games is so important. Sometimes a player contributes so much in three periods, lifting sticks at the right moment, breaking up offence, maintaining pressure, driving the offence, and they won't end up with a single point.

While someone could have an easy tap in goal. Look, good players will end up on the scoresheet, but you can't just look at stats.

How many points did Crosby lose over the years because his teammates couldn't finish a beautiful set up? You can't measure that.

 

[tarheelhockey]

As a matter of readability, there’s no reason to show these numbers to the 9th decimal.

 

[Muikea Bulju]

There should be a secondary statistic on who actually was mostly responsible for the goal

It might be all on the goal-scorer. If McDavid takes the puck in his D-zone and creates a goal all by himself from completely nothing, the goal should be accounted for him by at least 90%

In some other cases, the passer or the guy with the second, or even the third assist might have been the one actually mostly creating the whole thing. If a guy like Forsberg, Gretzky or McDavid (again) coasts around the offensive zone for 40 seconds and finally passes to an open scoring chance to some goon, and it goes in off of his skate or something, the goal is mostly due to the passer (or even the third passer)

I think the league should create a "jury" to vote on every goal in the league, and award different players different shares of the "responsibility" of the goal scored.

Then add the results together, and publish a stat on it on nhl.com. The current stats can remain, just add another one to the side

Then we could have a more objective way of saying that someone is actually better at offense than someone else

 

[OppositeLocK]

There should be a secondary statistic on who actually was mostly responsible for the goal

It might be all on the goal-scorer. If McDavid takes the puck in his D-zone and creates a goal all by himself from completely nothing, the goal should be accounted for him by at least 90%

In some other cases, the passer or the guy with the second, or even the third assist might have been the one actually mostly creating the whole thing. If a guy like Forsberg, Gretzky or McDavid (again) coasts around the offensive zone for 40 seconds and finally passes to an open scoring chance to some goon, and it goes in off of his skate or something, the goal is mostly due to the passer (or even the third passer)

I think the league should create a "jury" to vote on every goal in the league, and award different players different shares of the "responsibility" of the goal scored.

Then add the results together, and publish a stat on it on nhl.com. The current stats can remain, just add another one to the side

Then we could have a more objective way of saying that someone is actually better at offense than someone else

Wonderful idea. Make something more objective by basing it on something completely subjective.

 

[Kettil33]

Here's the flaw in discounting assists.

You count every goal you score. If you don't score, its because you didn't score. But most "assists" don't count. You make a perfect pass, set a guy up on a breakaway, and he doesn't score - no assist. But you couldn't have made a better play. For every assist a set up man actually gets, he probably makes several similar plays. In other words, although goals are what counts on the scoreboard, its generally the assist getters that drive play, and creating multiple assist worthy opportunities for every goal that actually gets put up on the board. So, ya, if those plays are only worth .4 of a goal, then if you make 5 of them for every goal scored, assist should count 2 points ... if you want to objectively count contributions to offensive play. 1 point seems a fair reflection ... at least for 1st assists, but the logic against heavy discounts still applies to some degree to second assists as well.

 

[Midnight Judges]

You’re conflating PP points with 5v5 points here which is a glaring issue. Your list is basically ordered by PP time.

There is also a strong correlation with ES TOI and TOI in general. This is because playing more games and minutes increases the likelihood that a player will add more value to their team. It's not a bad thing.

If we were comparing top tier players to scrubs who never see PP time, you'd have a stronger point, but in this case almost everyone on the list is a top line forward. They were deployed on the PP to the extent that it helped their team.

It would be better off splitting them into PP and 5v5, and also using p60 given there were almost double the PP opportunities circa 2005 as there are presently.

An adjustment based on power play opportunities makes sense, but obviously it wouldn't impact the relative rankings of players who were in the NHL for all 14 of the seasons in the charts. It may slightly reduce values for the older players relative to the guys who entered the league in 2008 and later.

 

[Sidney the Kidney]

Objective = anyone who reduces the value of assists in order to come to a conclusion they want

 

[OppositeLocK]

Objective = anyone who reduces the value of assists in order to come to a conclusion they want

Yes, and I'm sure you'd find it all very objective if Sidney the Kidney was number 1.

 

[Muikea Bulju]

I think the league should create a "jury" to vote on every goal in the league, and award different players different shares of the "responsibility" of the goal scored.

I would also add a twist to it: if some god-mode player dangles around the Offensive zone for 2 minutes, and forces the tired players to just ice it, and if the next line that comes in to take the oz draw then scores a goal against the tired D, some of the "award", could be handed to the guys who weren't even on the ice, but were on earlier.

And if some whiners don't like the idea of a "jury" to vote on stats, fearing a single crosby-hater might skew the stats, you could just cope every single sport with a panel of judges giving points: just throw out the highest+lowest scores, and take an average out of the rest. And have a large enough "jury", 1-3 persons from every fanbase.

 

[Midnight Judges]

I dunno how objective this is. I know not every goal is the same, or even assist.

This is why the eye test and watching the actual games is so important. Sometimes a player contributes so much in three periods, lifting sticks at the right moment, breaking up offence, maintaining pressure, driving the offence, and they won't end up with a single point.

While someone could have an easy tap in goal. Look, good players will end up on the scoresheet, but you can't just look at stats.

How many points did Crosby lose over the years because his teammates couldn't finish a beautiful set up? You can't measure that.

These things tend to even out over the course of ~ 20,000 minutes.

If a player is truly setting guys up all the time, the assists will reflect that.

 

[OppositeLocK]

I would also add a twist to it: if some god-mode player dangles around the Offensive zone for 2 minutes, and forces the tired players to just ice it, and if the next line that comes in to take the oz draw then scores a goal against the tired D, some of the "award", could be handed to the guys who weren't even on the ice, but were on earlier.

And if some whiners don't like the idea of a "jury" to vote on stats, fearing a single crosby-hater might skew the stats, you could just cope every single sport with a panel of judges giving points: just throw out the highest+lowest scores, and take an average out of the rest. And have a large enough "jury", 1-3 persons from every fanbase.

This is literally the most confusing and arbitrary scoring system I've ever heard.

 

[Midnight Judges]

As a matter of readability, there’s no reason to show these numbers to the 9th decimal.

Ha! Fair enough.

 

[Sidney the Kidney]

Yes, and I'm sure you'd find it all very objective if Sidney the Kidney was number 1.

No, I'd rather a system that values both equally and doesn't punish one type of player (goal scorer or playmaker) in doing so. That's an objective analysis. One that doesn't play favorites to a specific type of player.

Your post might have validity if I made an "objective analysis" where assists are given twice the value of goals or where goals are only worth a fraction of assists. That's pretty much the extent of what the OP is doing for goal scoring.