Adjusted Scoring

Triffy

Registered User
Jun 23, 2006
337
3
Helsinki
For the purposes of "relative to era" comparisons I would consider his skating to be like that of a Tim Kerr or Dave Andreychuk. They still accomplished great things despite their skating. I'm worried more about results, not exactly how they achieved those results.

I have no idea how good of a skater Tim Kerr was. Andreychuk I know wasn't good. But I suppose you mean't those two were, relative to their era, bad skaters. But they were bad skaters when average skating ability was good. Denneny was not. Denneny's skating ability was bad, even for his era. The offside rule was introduced the season after Denneny quit. I don't know how he played the game. But I could imagine that he'd spend much time just waiting for the puck in the offensive zone where he was lethal.

I also prefer players who base their game on universal hockey abilities, such as skating, stickhandling and hockey sense. If you're good at those, you can go anywhere and do well. But that's just my opinion.
 

Kyle McMahon

Registered User
May 10, 2006
13,301
4,352
It's too bad that there's a lot of anti-pre-26 sentiment out there. If someone was the top goal-scorer in the NHL, NHA, or PCHA, there is no evidence to suggest that they wouldn't be one of the top 3 or 4 goal-scorers in an imaginary consolidated league. It's not their fault what the state of top-level hockey was in their time.

Well exactly. In a consolidated league somebody would have been the top goal scorer, yet some people seem to punish everyone. My theory is, punish the guys who you think would have only been a top-10 scorer instead of top-3 if the leagues were consolidated, but whoever you think would have led a consolidated league, as well as thier own, should not be penalized.
 

MXD

Original #4
Oct 27, 2005
50,668
16,394
Well exactly. In a consolidated league somebody would have been the top goal scorer, yet some people seem to punish everyone. My theory is, punish the guys who you think would have only been a top-10 scorer instead of top-3 if the leagues were consolidated, but whoever you think would have led a consolidated league, as well as thier own, should not be penalized.

Well, it's also that some guys arguably played their best hockey post consolidation. Bill Cook and Herb Gardiner comes to mind.

The thing with Gardiner is that, in a consolidated league, there are way much more guys that can bring what Gardiner does.
 

Hockey Outsider

Registered User
Jan 16, 2005
9,056
13,985
A few people have requested either the raw data I used, or the final results. You can download an Excel file with all the data here: http://rapidshare.com/files/179526850/AdjustedScoring_v7.xlsx.html

Those of you with an older version of Excel should try this one: http://rapidshare.com/files/179577432/AdjustedScoring_v7.xls.html

...Would this method (with the milestones seasons...) sligthly overrate wartime players? Herb Cain and Billy Taylor in Mark Messier's territory just seems wrong to me.

Yes, that's one of the weaknesses of this method. I know that Cain and Taylor aren't close to Messier; this is where a qualitative look at the quality of the competition is important.

It's too bad that there's a lot of anti-pre-26 sentiment out there. If someone was the top goal-scorer in the NHL, NHA, or PCHA, there is no evidence to suggest that they wouldn't be one of the top 3 or 4 goal-scorers in an imaginary consolidated league. It's not their fault what the state of top-level hockey was in their time.

Just to be clear, I agree with you on this point. A player finishing 4th in scoring in the NHL in 1920 is probably equivalent to a player finishing around 8th in a hypothetical consolidated North American league.

I get your point, and agree in general. But I can understand why people drop players like Denneny down: today's game is based mostly on skating, and his skating ability was weak even for his era. It's just hard to imagine a player like that excell in any given conditions.

I think one could argue that Denneny could have been similar to Andreychuk in the modern NHL -- slow, heavy, and willing to play a bit dirty at times. Robitaille is another example of a ridiculously slow goal-scorer, but he really lacks Denneny's size and physical play.

I think the only way to truly adjust scoring stats for a war year, is to do the following:

a) assume that the missing star players were still in the league, assume they played whatever percentage of the season they usually did, and determine where you would slot them in among the league leaders in goals, assists, and points. Credit them temporarily with the corresponding amounts of goals, assists, and points.

b) adjust the inflated scoring totals by multiplying by the inverse of what factor scoring increased by, compared to the average of the last three pre-war years. i.e. 75 points might only mean 55 points. In other words, assume that if the status quo was maintained, scoring levels would stay the same (you may also want to assume a slight upward trend throughout those seasons to account for scoring getting to the level it was at in 1945-46)

I like Method B more -- I'd feel more comfortable adjusting the statistics that a player actually accumulated, rather than making them up for players who weren't in the league. Clearly, a player like Syd Howe would have scored a fair number of points even if the NHL was at full strength from 1943-45, we just know that it would be less than ~55 points per season.
 
Last edited:

Hockey Outsider

Registered User
Jan 16, 2005
9,056
13,985
Norm Ullman — Amazing. I knew he was underrated. But on the list, he's equal to one of my favorites, Syl Apps. And Sakic, who I'm surprised to see that high, too. And like Sakic, Ullman was also a great playoff performer and a good defensive forward. Ullman barely made it to the HOH Top 100 list last time. Hopefully he rises a bit in this year's update. The gap between Sakic & Yzerman & Apps (who I all see quite similar) and him shouldn't be that big.

I like Ullman too. His peak was a bit higher than I expected, but his amazing 1964-65 season pushes him into the top thirty (in peak value). His playoff record is inconsistent but he has a stunning peak from 1963-66. Throw in a solid two-way game and few years as a Hart nominee and I think he's a borderline top 100 player.

I also like Apps and although he matches, and possibly exceeds, Sakic or Yzerman's peak, he's a fair bit behind both of them in terms of longevity.

Cy Denneny — There are only 2 LWs (Bobby Hull and Ted Lindsay) ahead of Denneny on the list, which is impressive. But as HO already pointed out using Joe Primeau as an example, this method too has its flaws. Denneny seems to have been dominant in playmaking during his time. That's what the 78 assist peak average tells me. The problem is, no one remembers him as a great playmaker. Or am I wrong here? I should definitely re-evaluate his position among LWs. He stands at #11 now.

Denneny is one of the toughest players to evaluate. His scoring record is phenomenal (surprisingly, not too far behind Jagr or Mikita). The two main criticisms are 1) could such a slow, overweight player thrive in the modern era? (I say probably yes, Andreychuk is a decent comparable) and 2) would Denneny have succeeded in a consolidated North American league (again, I say probably yes, he might not have finished in the top three every year, but he still likely would have been in the top 8-10 at the very worst).

At the same time, I don't think he had the 12th greatest peak in NHL history (but even if we took 30% off his totals he'd still rank around the Francis/Bure/Sundin level).

The Bentleys — This is an interesting discussion. Doug and Max one after the other, like it should be. But wait, not on a list that ranks the players by offensive numbers? This surprises me. I always thought Max had the better offensive numbers. But Doug actually got similar results, while being the more responsible defensively. Of course it needs to be pointed out that Max has easily the better playoff resume, thanks to his years as a Maple Leaf. Again, one of those pairs that shouldn't be that far apart from each other on the HOH Top 100 list.

Max was definitely the better scorer if we exclude (or even if we adjust downwards) the WWII years. Max was a tremendous playoff performer relative to his era; Doug was brilliant in his one length playoff run (1944) but otherwise gets an "incomplete" since he was stuck on weak teams his entire career.

Nels Stewart vs. Frank Boucher — One of my numerous misevaluations, it apperas. I made a statistical comparison between Stewart and Boucher some time ago. I can't remember the details, but I think they seemed to be very close offensively. Here Boucher is way ahead of Stewart. Then I look at Boucher's incredible assist totals. Is that just the same "flaw" once again?

Essentially yes. Boucher and Stewart are essentially equal in raw stats (though Ol' Poison did have the longer career). My formula is essentially giving Boucher extra credit because, as a playmaker stuck in an era that gave out very few assists, he must have been a better scorer just to keep even with Stewart, a goal-scorer who benefited (relatively speaking) from how few assists were awarded.

Wow! Very impressed by Syl Apps' per-game numbers. It's not completely surprising, as I knew he missed a few games every year and did very well in per-game metrics, but it's still very impressive to see him in sixth place all-time. This result, along with his high ranking in seventieslord's Hart study, is making me think we may have underrated him in the top 100 list. I'll certainly have him higher on my list.

Apps was a remarkably consistent scorer. He was a top-ten scorer during six of his seven full seasons. In his worst healthy year, he was 12th in scoring. During the three seasons he was injured ('40, '42 and '43), he ranked 1st, 6th and 4th in points per game. At first I was a little surprised to see him rank so high, but I think it makes sense given his consistency (and the fact that he retired young).
 
Last edited:

seventieslord

Student Of The Game
Mar 16, 2006
36,080
7,131
Regina, SK
I like Method B more -- I'd feel more comfortable adjusting the statistics that a player actually accumulated, rather than making them up for players who weren't in the league. Clearly, a player like Syd Howe would have scored a fair number of points even if the NHL was at full strength from 1943-45, we just know that it would be less than ~55 points per season.

It was meant as a two-pronged approach. I mean, we can multiply all the totals from those seasons by 0.8 if we want, but in the end if we adjust using a formula like yours it ends up the same as if we hadn't multiplied by 0.8. Right?

My thinking was that if we assume Schmidt and Apps, for example, are in the league, and assume them to be, for example, 2nd and 4th in league scoring with a realistic number of points, that doesn't change how many points the other players have, but it does change their "score" under your system to something more realistic - it somewhat removes the "reward" of being in a league devoid of talent.

What Method A also helps to do is put into better context the top-10 finishes in those seasons. For example, one can say "yeah, so and so was 3rd in goals, but it was a war year.", completely discounting it. In reality it should not be completely discounted, but it should be noted how many more superior players would have finshes ahead of him if they weren't off at war.
 

MXD

Original #4
Oct 27, 2005
50,668
16,394
Can somebody send me the raw data on non-Vista format? (.xls...)
 

Hockey Outsider

Registered User
Jan 16, 2005
9,056
13,985
Can somebody send me the raw data on non-Vista format? (.xls...)

http://rapidshare.com/files/179577432/AdjustedScoring_v7.xls.html

It was meant as a two-pronged approach. I mean, we can multiply all the totals from those seasons by 0.8 if we want, but in the end if we adjust using a formula like yours it ends up the same as if we hadn't multiplied by 0.8. Right?

Correct.

My thinking was that if we assume Schmidt and Apps, for example, are in the league, and assume them to be, for example, 2nd and 4th in league scoring with a realistic number of points, that doesn't change how many points the other players have, but it does change their "score" under your system to something more realistic - it somewhat removes the "reward" of being in a league devoid of talent.

What Method A also helps to do is put into better context the top-10 finishes in those seasons. For example, one can say "yeah, so and so was 3rd in goals, but it was a war year.", completely discounting it. In reality it should not be completely discounted, but it should be noted how many more superior players would have finshes ahead of him if they weren't off at war.

Although this would be fairly subjective, I think it makes sense. It might be difficult but it should probably be done on a team-by-team basis -- for example, if the Kraut line and Roy Conacher were in the NHL in 1944, it seems highly unlikely that Herb Cain would have received enough ice time to come anywhere close to the top ten in scoring.
 
Last edited:

Hockey Outsider

Registered User
Jan 16, 2005
9,056
13,985
Going through links that don't work anymore in the Past Studies thread, this needs updated please.

This approach has essentially been replaced by the VsX methodology. Both approaches are conceptually similar, but I think VsX is stronger, both in terms of the theory behind it, and how reasonable the results look. For that reason, I'm not planning to update this thread.

The link to the revised VsX thread is - http://hfboards.mandatory.com/threads/reference-vsx-comprehensive-summary-1927-to-2017.2215905/
 

Ad

Upcoming events

Ad

Ad

-->