Adjusted stats - how valuable?
View Single Post
11-13-2012, 01:22 AM
Just a Fool
Join Date: Feb 2003
Location: Guelph, Ont
Originally Posted by
Czech Your Math
I address as many points of yours that I deem worth response and try to bring context to the discussion. You want to isolate perceived flaws and then it's off to the next often incorrect point. I think your strategy is basically guerilla warfare and hoping that I'll give up and go home. Since I'm tired of repeatedly explaining the same concepts to you, I will likely soon cease discussion for now.
I know how you feel, I'm getting pretty sick of asking questions or bringing up other context only to be fed more concepts disguised as answers.
I've already explained why using raw points even over the same range of seasons can be misleading. Every season is different, and in seasons with higher scoring it's easier to create a substantial gap in raw points than a lower scoring season (with the same % advantage). Gretzky was better in the early 90s and Jagr better in the mid-late 90s, which creates an illusion in raw points for the 90s due to the change in league scoring level. Daniel Sedin's scoring title in 2011, when Crosby & Malkin were injured and Ovechkin began his decline, had a different "quality" than Lemieux beating Gretzky & Yzerman in 1989 and they each had a different quality than someone winning one in depleted O6 during WWII.
Dude, they played in the league together at the same time from 90/91-98/99. 9 of the 10 years of the 90's and you want to convert them to adjusted stats?
The only illusion being created is by you by using adjusted stats!
You're trying to take what actually happened for both players under identical conditions, against the exact same competition and complicate it by converting it to a system that is far from perfect to begin with.
There is absolutely no logical rational for doing this period!
If we were comparing a 26 year old Gretzky to a 26 year old Jagr, that's a whole different ball game but we are not!
Seriously dude, I know I'm stubborn and have some trouble admitting when I'm wrong but you take the cake as well as the whole bakery on this one.
The value of Murray's goals was greater than those of Bossy in many seasons. One could attribute some of that value to Thornton instead, but that doesn't change the fact that the actual goals had more value due to the lower scoring environment.
Murray is an oft-cited anomaly, but it's not really a flaw of adjusted stats. He finished 2nd and 5th in goals in consecutive years. In '02 he beat out top 20 finishers like Naslund, Bondra, Tkachuk, Lindros, Shanahan, Alfredsson, Demitra, Modano, Bure, Gagne, Kariya, Palffy, Kovalev, Briere, Yashin, etc. In '03 he beat out top 20 finishers like Heatley, Kovalchuk, Palffy, Sundin, Hull, Kovalev, Jagr, Fedorov, Thornton, Jokinen, Demitra, Iginla, St. Louis, Mogilny and Kovalev. He had a couple of amazing years relative to his peers, but that's not the fault of adjusted stats.
So you begin by calling it an anomally and yet still try and defend it...killin me man.
Soooo Bossy scores 20% of the Islanders goals and is in on 39% of their total in 82/83 compared to Murray's 18% of Boston's goals and 38% of their total in 02/03.
Despite that, you are perfectly fine believing/defending that Murray's 02/03 season is valued to be 8% higher than Bossy's 82/83?
See I don't know about you but it sure looks to me like Bossy contributed more and had more value to his team winning than Murray did his
These favorable conditions only reinforce my position that subsequent players may have been able to achieve a substantially higher level of raw points during the 80s. You seem anchored to the position that those players were all as good or better than later group of players which I previously listed.
Actually yeah...I am anchored to the idea that Lafleur, Bossy, Stastny and Dionne are as good and/or better than anyone you named.
I think you'll find I am far, far from alone on that.
AS has a different value to different people, I think that much is obvious. Some people basically completely discard them and use raw points or don't consider such data at all. I think most of us who consider them superior to raw data know that there is additional context needed when comparing different player-seasons. Most may prefer to do some additional mental adjustments, but in the longer run I believe it would be better and more useful to have explicit, quantified data that allows for fine-tuning of the numbers when comparing the quality/difficulty of various player-seasons. Until that time, it's still a common ground for many to plan routes to the "final summit", whether those are implicit, mental routes or explicit, quantified routes.
IMO the process should be as follows:
- hypothesize reasons that AS distorts the quality/difficulty of adjusted production in various seasons for various types of players (particularly top tier players)
- analyze and quantify the effects of these possible factors, preferably simultaneously via regression, since many of them may overlap substantially in terms of when they changed (PP opportunities, league GPG, Euro influx, most recent expansion all occurred roughly around the same time).
- use reason to assess the results of the analysis and, if necessary, apply the results to players in a fair way via an adjustment to the formula when used for the purposes of comparing the quality of various player-seasons
I think someone well-versed (better than I) in regression could help tremendously in this effort, but it seems those posters are few and far between, which is one reason such a process could take a while. Such posters may not be inclined to themselves pursue such a project, which would mean that they must be willing to assist before/during such a project when it is undertaken by (an)other individual(s).
Or of course the simple answer, that AS's value is not a constant in every case
Last edited by Rhiessan71: 11-13-2012 at
View Public Profile
Find More Posts by Rhiessan71