The 5th best player ever is a goalie.

[BraveCanadian]

Then add in the US, Europe, British Columbia ect.... but when solid evidence gets presented to indicate the direction we get posts like this.

Not going to pull out all the evidence again. Nope. Not again.

 

[blogofmike]

Why would it make more sense to compare to the best 6 forwards? When people adjust scoring they also adjust on a average goals per game basis and not on a average goals scored by the best 6 players basis. I know it's a popular way to do it this way on this forum but it honestly makes absolutely no sense.

I think he meant Top 6 as in first liners and second liners. Which makes sense, since you compare scorers to scorers.

When automotive forums debate the best race car of all time, they talk about how the McLaren F1 or Ford GT40 and how they did compared to the best race cars of their day. They don't take the average of every 4 wheeled conveyance into account so that the sample size includes the average speed of John Deere tractors.

Which is to say, what exactly is the value of including Jim Peplinski's scoring in the Gretzky 84 / Lemieux 89 baselines?

Also roster sizes were 15/16 for Howe, as opposed to 18 for Gretzky and Lemieux, which doesn't impact a comparision against top scorers, but would impact the methodology you outlined in post 102 which includes Jim Nill in samples for 1984 and 1989, (even though he failed to crack the Top 300 scorers of either year,) while potentially excluding the 5th place scorer from each season because they were sub-70 game players,

You mean the 95/96 season? Lemieux slowed down by then and hugely benefited by playing power plays with prime Jagr but purely statistically? He dominated even more than ever before. The guys who did the whole adjusted thing ranked this season above every season of either Gretzky or Lemieux. I mean look at the top ten, every player is a legend and in the case of Jagr, Forsberg and Selanne the best forward ever from their respective countries...

View attachment 794340

1996 Lemieux is more comparable to post-peak Howe. Peak Howe was a better even strength scorer. Here's his best year of his prime:

	50 player	GP	EV Pts	EV PPG	96 player	GP	EV Pts	EV PPG
1​	Howe	70​	70​	1.00​	Jagr	82​	95​	1.16​
2​	M. Richard	65​	50​	0.77​	Nedved	80​	76​	0.95​
3​	Lindsay	67​	50​	0.75​	Lindros	73​	75​	1.03​
4​	Abel	69​	50​	0.72​	Lemieux	70​	73​	1.04​
5​	Schmidt	62​	48​	0.77​	Mogilny	79​	71​	0.90​

Lemieux doesn't really stand out from the pack, while that's a pretty good gap between Howe and the field, Still, 6-team leagues are much more prone to having a scoring race get impacted by a single injury and all, so maybe we expand the sample.

	51-53 Player	Year	GP	EV Pts	EV PPG
1​	Howe	1951​	70​	70​	1.00​
2​	Howe	1953​	70​	67​	0.96​
3​	Howe	1952​	70​	61​	0.87​
4​	Lindsay	1953​	70​	54​	0.77​
5​	M. Richard	1951​	65​	50​	0.77​
6​	Lindsay	1951​	67​	50​	0.75​
7​	Abel	1951​	69​	50​	0.72​
8​	Lindsay	1952​	70​	50​	0.71​
9​	Schmidt	1951​	62​	48​	0.77​
10​	Hergesheimer	1953​	70​	47​	0.67​

Howe seems to stand out even more from 1951-53. How about that? In a 3-year sample no one is on pace to score more than 55 ES points, but Howe's floor is 61.

But still, Wally Hergesheimer? Who is that? Let's cover Howe's rookie season, all the way to the first expansion in 1967-68.

	47-68 Player	Year	GP	EV Pts	EV PPG
1​	Howe	1951​	70​	70​	1.00​
2​	Bo. Hull	1960​	70​	69​	0.99​
3​	Howe	1953​	70​	67​	0.96​
4​	Moore	1959​	70​	67​	0.96​
5​	Mikita	1967​	70​	65​	0.93​
6​	F. Mahovlich	1961​	70​	64​	0.91​
7​	Bo. Hull	1962​	70​	63​	0.90​
8​	H. Richard	1963​	67​	61​	0.91​
9​	Howe	1952​	70​	61​	0.87​
10​	Ratelle	1968​	74​	61​	0.82​
11​	7 Players incl. Howe 68			60​

Of course, that 70 ES points wouldn't hold up for long. Three players (including Gordie Howe) would beat it in 1969. (Technically Herb Cain's 74 ES points in 1944 was the most until 1969, but Boston may have been scoring PP goals as ES goals. Otherwise 70 ES points is the most anyone would score in the first 50 years of the NHL.)

Howe's peak stands up pretty well when compared to the best of the best, even when giving up an era advantage.

 

[daver]

If the league size in general reflects the expansion of talent then quite clearly the 3rd best player out of 150 (98th percentile) should be equal to a 98th percentile player out of a 1000 which would be 20th.

Your two points - saying that the 3rd best player is equal to the 5th best player today and saying the league talent depth reflects the expansions - completely contradict each other.

I merely presented stats to show that it is not as simple as multiplying everything by five to account for 30 plus teams vs. 6 teams.

You need to present some evidence to back up your claim that players from earlier eras clearly benefitted from lower talent/lower # of players.

Most HOH posters acknowledge that a 5th place scoring finish in the O6 is not the same as 5th place in the current era based on the "percentile" angle.

But the stats do not back up the premise that you simply multiply by 6.

IMO, GOAT talent is GOAT talent in any era. There is no reason to think that Howe doesn't dominate in any era, relatively speaking. He loses Art Rosses to Wayne and Mario, and maybe to peak Jagr, Crosby, and McDavid. A 50 to 70 year difference in playing time is just too long to speculate.

I do not accept the premise that his domination needs an asterix because there are more players in the current league. Do you put an asterix by Wayne and Mario then too?

 

[daver]

We spoke about Howe's domination against peers vs Gretzky/Lemieux's respective domination against their peers. I used the scoring against the average forward who has played the majority of games in a season (roughly 90% or more). That is in order to get rid of the players who played a low amount of games in a season which interestingly was a huge amount of players in the 1980s. The other guy used a completely arbitrary vs 6th 10th 20th best player not at all accounting for the size of the league.

I think domination against one's peers can be used to evaluate the talent of a player. Of course the size of the talent pool matters way more but since it is way more difficult to assess and since the NHL's talent pool did evolve at least partially together with the size of the league the domination against one's peers in the NHL can be used.

Of course I was accounting for league size, just not in the way that you want that establishes bright lines of differentiation.

It doesn't pass the smell test to say claim that Beliveau's, Howe's or Hull's, 3rd place finish in any given season equated to a Top 20/25 finish in the current era.

 

[bobholly39]

With 5 points in last night's game, McDavid is now in 3rd place in scoring race. This after an abysmal start to the year, where it seemed he was headed to an off-year. If he sees this season through - and sweeps Ross/Hart while helping Oilers climb back in playoffs - let alone any potential playoff success on top of that this year - it will be yet another truly outstanding accomplishment on his resume, to add to 2021/23 seasons + 22 playoffs.

The point is - I still think there's a chance McDavid manages to break up this big 4, and ends up with an argument for #5 for sure, but possible even a "big 5" where he's not necessarily just #5.

In the absolute best case scenario with no injuries and upper end of projections for missed time - I still say Crosby ~ Howe, or even Crosby > Howe could have been a real possibility. And outside of playoffs, McDavid's prime is somehow looking to be even above those upward projections for Crosby had he had no injuries. So for McDavid, I would say:

- Ends up with strongest case at #5 - extremely likely. Over 50% likelihood (playoffs will be the make or break it point most likely)
- Unanimous #5 (meaning his case is so strong, no one else from Crosby/Beliveau/Hull etc can be argued above) - I'd say this is ~25% likely. He definitely needs some strong playoff successes, and continued prime, but so far so good.
- "Big 5" - meaning not only is he unanimously above the field for #5, but he has a legitimate claim to be argued above some of the big 4, depending on what you like (just like some say Lemieux > Howe vs Howe > Lemieux). Maybe ~5-10% chance? The fact that this is even possible is incredible.

Regarding goalies - as of today I usually slot Crosby at #5, and Roy at #6. Hasek I have around ~11th or so (better peak than Roy, but lesser career, playoffs being a big differentiator). So I don't have a goalie at #5 myself, but not exactly far.

 

[rmartin65]

Yeah, it seems inevitable that McDavid ends up at least number 5, so long as he stays healthy over the next couple of years.

I don't think he breaks up the Big 4; he either makes it a Big 5, or there may be this idea that there is a Big 3 (Orr/Gretzky/Howe) and a Next 2 (Lemieux/McDavid).

 

[Michael Farkas]

One of my favorite things in these exercises to justify not putting the real evaluation work in, people name a player that I know is really skilled, but a little bit lost to the sands of time, and go "can you believe [this guy] finished Xth in 195X?!?!"

We watched Jonathan Cheechoo lead the league in goals, a Jamie Benn Art Ross, Chris Kreider just finished 3rd in goals, etc. and that's without counting all of the random goalies that end up towards the top of the stats pack due to team effects haha

You're never ever ever going to be able just to wrap this up in a nice little box. At some point, the work is going to need to be done if you want better/real answers.

 

[jigglysquishy]

One of my favorite things in these exercises to justify not putting the real evaluation work in, people name a player that I know is really skilled, but a little bit lost to the sands of time, and go "can you believe [this guy] finished Xth in 195X?!?!"

We watched Jonathan Cheechoo lead the league in goals, a Jamie Benn Art Ross, Chris Kreider just finished 3rd in goals, etc. and that's without counting all of the random goalies that end up towards the top of the stats pack due to team effects haha

You're never ever ever going to be able just to wrap this up in a nice little box. At some point, the work is going to need to be done if you want better/real answers.

This comes up a lot in casual conversation. It's usually connected with a good/great player lost to time.

People on the main board have criticized the 50s because Bernie Geoffrion led in goals and they didn't recognize the name.

 

[MadLuke]

Wuold be a bit like one day someone looking at Zhamnov 3rd place in the lock-out year and be a bit baffled by the talk of how loaded the early 90s was in top end talent.

Even the many of the scoring leaders were out of the league by 29 once the talent crisis resolved itself (post war or something) to downplay an era can be dangerous, injuries, Wade Redden, MIke Richards, etc.... players fall off hard in all eras sometime, we could imagine even more so in a world player were not protected by long superbe contract, with less spots to move too.

 

[Overrated]

Of course I was accounting for league size, just not in the way that you want that establishes bright lines of differentiation.

It doesn't pass the smell test to say claim that Beliveau's, Howe's or Hull's, 3rd place finish in any given season equated to a Top 20/25 finish in the current era.

It actually does. Howe, Beliveau, Hull were all the best player in the NHL at some point but all three of them would show up in the top3 or top5 finishes all throughout their 30s. In 68/69 for example 40 year old Gordie and 37 year old Beliveau finished #3 and #6 in PPG. Sure they had better than average longevity but so do Crosby, Ovechkin and Malkin (the three best players from the last era) yet last season they finished respectively at #16 #40 & #26. You yourself said Crosby is better than Hull and Beliveau and closer to Howe yet he didn't even make the top15 in points and was #19 in PPG last season.

 

[wetcoast]

Not going to pull out all the evidence again. Nope. Not again.
View attachment 794903

That's probably because there is no hard evidence for either side as it's not as simple as a formula or that talent can be measured across long periods of times nevermind a ton of other variables.

 

[Michael Farkas]

Ovechkin also finished 37th in points in his prime, tied with Teddy Purcell and behind Val Filppula. What gets inferred from that data point about....anything?

 

[Bear of Bad News]

Filppula had Hall of Fame hair coloring/bleaching game, so it's no shame to be on a list with him.

 

[Johnny Engine]

Ovechkin also finished 37th in points in his prime, tied with Teddy Purcell and behind Val Filppula. What gets inferred from that data point about....anything?

Purcell was the highest scoring Newfoundlander that year, out of 550K people (all of whom can be assumed to be in the NHL talent pool, including...ESPECIALLY...Phonse from Nicky's Nose Cove), while Ovechkin came 4th among Russians, out of 143 million people, but since they got 22 medals in Sochi (but not in hockey so make it 23 sports they're playing), we can assume each sport has a talent pool of 6.2 million, and if you divide that by 4 you see that Ovechkin is probably 3 times more impressive than Purcell, which is pretty good for ol' Teddy because who knew he had it in 'em.

 

[rmartin65]

Ovechkin also finished 37th in points in his prime, tied with Teddy Purcell and behind Val Filppula. What gets inferred from that data point about....anything?

Ovechkin is the youngest of the three, so he’s clearly superior based on talent growth over time.

 

[blogofmike]

One of my favorite things in these exercises to justify not putting the real evaluation work in, people name a player that I know is really skilled, but a little bit lost to the sands of time, and go "can you believe [this guy] finished Xth in 195X?!?!"

In my head, I was writing that in a sarcastic voice...

 

[Michael Farkas]

In my head, I was writing that in a sarcastic voice...

I didn't really intend to single you out here, so I apologize for it perhaps coming across like that...it's just been a very common theme in the recent talent pool shenanigans. The "say old timey sounding name to confirm I don't do any homework on this" just isn't the dunk that some folks appear to think it is...

 

[BraveCanadian]

That's probably because there is no hard evidence for either side as it's not as simple as a formula or that talent can be measured across long periods of times nevermind a ton of other variables.

It isn't simple, but all the evidence that we do have does not support your theory.

 

[wetcoast]

Ovechkin also finished 37th in points in his prime, tied with Teddy Purcell and behind Val Filppula. What gets inferred from that data point about....anything?

Ovi had a poor year?

 

[blogofmike]

A lot of this seems like territory that's been covered.

That's the upside of VsX. X is a high water mark that's often the 2nd place scorer. If you score 80% of X, it doesn't matter if you are 5th place in 1954 or 25th place in 2024. And someone clever built in caveats for outlier seasons for when 2nd place isn't a good benchmark.

 

[jigglysquishy]

A lot of this seems like territory that's been covered.

That's the upside of VsX. X is a high water mark that's often the 2nd place scorer. If you score 80% of X, it doesn't matter if you are 5th place in 1954 or 25th place in 2024. And someone clever built in caveats for outlier seasons for when 2nd place isn't a good benchmark.

It's a really neat feature of X

Season	X	>0.9X	>0.8X	>0.7X
1962-63	81	5	7	15
1963-64	78	6	9	14
1964-65	83	3	5	8
1965-66	78	7	7	14
1966-67	70	6	9	13
Average (last 5 O6 years)	78	5.4	7.4	12.8
2018-19	116	4	11	25
2019-20	97	5	11	17
2020-21	69	9	18	32
2021-22	115	8	15	32
2022-23	113	10	18	31
Average (last 5 years)	102	7.2	14.6	27.4

For 0.8X and 0.7X, it is roughly twice as many players hit that target today as in the O6 era.

As an side, the X for 2018-19 has always sat wrong with me. If you look at the scoring tables, McDavid should clearly not be the benchmark. He is 16% ahead of 5th place, but because of the gaps between 2->3->4->5, he never hits the 5% rule. If we treat Kane's 110 as a benchmark, the numbers fall more in alignment with the 2021-2023 trends.

It's by no means a perfect system, but it does capture a growing league fairly elegantly.

 

[MadLuke]

For 0.8X and 0.7X, it is roughly twice as many players hit that target today as in the O6 era.

This look nice to me (and maybe it is just goes in the direction of all our prior, so our brain like it)

The 0.9x do change but not that much, the 0.8x change a lot but no world beating and than it goes off the rail.

I.e. making a top 5 is maybe just 33% a bigger deal now than back then, but making a top 20 is much much harder now than back then if I read those numbers correctly (if you achieved to put yourself in a roster spot that made it possible which would be quite the can of worm to compare vs now but that would be the case of the type of player being talked about), which would be the intuition of many.

 

[The Panther]

With 5 points in last night's game, McDavid is now in 3rd place in scoring race. This after an abysmal start to the year, where it seemed he was headed to an off-year. If he sees this season through - and sweeps Ross/Hart while helping Oilers climb back in playoffs - let alone any potential playoff success on top of that this year - it will be yet another truly outstanding accomplishment on his resume, to add to 2021/23 seasons + 22 playoffs.

The point is - I still think there's a chance McDavid manages to break up this big 4, and ends up with an argument for #5 for sure, but possible even a "big 5" where he's not necessarily just #5.

In the absolute best case scenario with no injuries and upper end of projections for missed time - I still say Crosby ~ Howe, or even Crosby > Howe could have been a real possibility. And outside of playoffs, McDavid's prime is somehow looking to be even above those upward projections for Crosby had he had no injuries. So for McDavid, I would say:

- Ends up with strongest case at #5 - extremely likely. Over 50% likelihood (playoffs will be the make or break it point most likely)
- Unanimous #5 (meaning his case is so strong, no one else from Crosby/Beliveau/Hull etc can be argued above) - I'd say this is ~25% likely. He definitely needs some strong playoff successes, and continued prime, but so far so good.
- "Big 5" - meaning not only is he unanimously above the field for #5, but he has a legitimate claim to be argued above some of the big 4, depending on what you like (just like some say Lemieux > Howe vs Howe > Lemieux). Maybe ~5-10% chance? The fact that this is even possible is incredible.

Regarding goalies - as of today I usually slot Crosby at #5, and Roy at #6. Hasek I have around ~11th or so (better peak than Roy, but lesser career, playoffs being a big differentiator). So I don't have a goalie at #5 myself, but not exactly far.

His position in the scoring race is even stronger than it appears, given that he missed 2 games (and then played injured) and that the Oilers have played fewer games than most teams. This is the current overall PPG scoring rate:
1.66 Kucherov
1.61 McDavid
1.61 MacKinnon

This is the PPG scoring rate since Nov. 23rd:
2.18 McDavid (with a best-in-the-NHL +17)
1.90 MacKinnon
1.79 Kucherov

Assuming no injury or whatnot, I have to think McDavid's current chances of winning yet another Art Ross to be substantial -- maybe 50%?

Really be something amazing if he could do it. An Art Ross for 2023-24 would give him SIX, tied with Howe and Lemieux. However, Kucherov is having an amazing year, so it's not going to be easy. Probably depends a bit on how things go for Tampa the rest of the way. If they continue to underwhelm, Kuch might see that object in the rear-view mirror overtaking him soon.

Speaking of, it now appears that Nate MacKinnon is the offensive equivalent in the current era to defenseman Brad Park of the past. That is, his career may end up being defined by the near-misses of several major trophies and many 2nd-place finishes.

 

[Vilica]

The 19-20 season is actually the most instructive when looking at VsX and comparing it to pre-expansion seasons, because the pandemic cut off the season at 70ish games. League average scoring that year was 208, and when looking at the 60s, that same scoring level was present for most of the seasons [58-59 203, 59-60 206, 60-61 210, 61-62 211, 62-63 208, 63-64 194, 64-65 201, 65-66 213, 66-67 209] in the same number of games. My formula thinks 97 as the benchmark is a bit high, the VsX average is just over 91 (also, the 18-19 benchmark is at 107, not 116), and going back to that O6 era, it thinks that 90 point threshold (plus or minus a little) should be the benchmark for most seasons except 63-64 and 64-65. Look at these two seasons:

Pts Rank	Name	Year	Team	Games	Goals	Assists	Points	Team GF	LA GF	% LA	G%	P%	VsX	VsX Season		Avg VsX	VsX
3	Stan Mikita	62-63	CHI	65	31	45	76	194	208	0.933	0.160	0.392	81	93.83		83.35	91.18
10	Jack Eichel	19-20	BUF	68	36	42	78	193	208	0.928	0.187	0.404	97	80.41		85.54	91.18

Those seasons are virtually identical, the VsX scores should be much closer to Avg VsX than their actual VsX.

Here's another set:

Pts Rank	Name	Year	Team	Games	Goals	Assists	Points	Team GF	LA GF	% LA	G%	P%	VsX	VsX Season		Avg VsX	VsX
3	Norm Ullman	66-67	DET	68	26	44	70	212	209	1.014	0.123	0.330	70	100.00		76.40	91.61
15	Kyle Connor	19-20	WPG	71	38	35	73	213	208	1.024	0.178	0.343	97	75.26		80.06	91.18
16	Mark Scheifele	19-20	WPG	71	29	44	73	213	208	1.024	0.136	0.343	97	75.26		80.06	91.18

Because Winnipeg played 71 games, their league average should probably be 211 or 212, which pushes their Avg VsX down to 78ish, but again, look how closely the goals for, the percentage of league average, and the players' point percentages match up. [In any case, a gap less than 5 between two seasons isn't really significant in any sense of the matter. Luck and randomness are just as likely for that gap as talent.] They're basically the exact same level of season, and yet there's a 25 point gap in their VsX because 2nd place in an 18 forward league has much more variance than 2nd place in a 90-96 forward league.

A crude threshold for an impact season is when a forward accumulates points on more than 40% of their team's goals. In the period between 49-50 and 66-67 (18 seasons), it happened 27 times. Concentrating on Boston and New York, it was accomplished by Andy Bathgate 5 times for the Rangers, and Murray Oliver once for the Bruins. If you look at the Boston teams in that period, their top scoring forward normally topped out between 30-35%, and sometimes not even that high. There was no structural reason why a forward on Boston couldn't hit that 40% threshold (as evidenced by Bathgate doing it 5 times for a New York team that scored about the same number of goals), they just never did. Bronco Horvath had a couple of high 30s, Johnny Bucyk was steadily around 33, Milt Schmidt had a 35, but Murray Oliver is the only Boston player that hit 40%.

That 18 seasons is the same between 05-06 and 22-23, and if I counted correctly, it occurred 71 times in that period. Obviously with 30+ teams and more players, there's more opportunity, but you look at the bad teams pre-expansion (Chicago, Boston and New York), and see their top forwards topping out at 30%, and you realize that the gap is lack of talent, not lack of opportunity. Take as an example some teams that were tanking, like the 14-15 Coyotes and Sabres - Sam Gagner was Arizona's top scoring forward, Tyler Ennis Buffalo's, and their stat lines (Gagner 41 points, 165 goals scored, 24.8%, Ennis 46 points, 153 goals scored, 30.0%). Compare them to the bad teams in the O6 era, and see the statistical similarities. If we know what Gagner and Ennis are, presumably their O6 comparables are probably about the same (in terms of talent). If a player is getting top-line deployments/PP time, and not breaking the 30% threshold, they're basically replacement-level as a top-line forward, especially if their team is below league average in goals for. [If a team is significantly above league average in goals for, their team point scoring is much more likely to be wonky, and the 30% threshold is not as applicable.]

 

[daver]

The general idea, IMO, is to not believe that one can apply a hard and fast formula but can make it reasonable to accept general statistical presumptions.

It is a fact that Crosby, McDavid and peak/prime Jagr are statistically closer to Howe than they are to Hull and Beliveau when measuring how each player finished relative to the #1/#2 scorer in points and/or PPG.

I can buy the idea that with the expansion of talent since Howe/the O6, three Howe level players have emerged in the last 40 years.

I can also buy the idea that standing out in your respective era has value also that cannot be automatically lowered due to the era in which one played. I.e. Hull and Beliveau stood out in their era as much as Crosby, McDavid (so far) and Jagr did in theirs.

As mentioned, in an earlier post, where a deeper dive into the numbers is useful is when you get past the Top 3 scorers. It is clear that is was easier to finish Top 5/Top 10 in the O6.