View Single Post
10-31-2012, 04:58 PM
#133
Registered User

Join Date: Jan 2006
Location: bohemia
Country:
Posts: 4,846
vCash: 500
Quote:
 Originally Posted by Rhiessan71 Most of it is you skirting around admitting there is a big flaw (and there is a big flaw that you will see if you do the exercise I mentioned earlier)
I honestly couldn't quite grasp what you wanted me to try to calculate or look at. If you could explain it further/differently, and it's not too time-intensive, then perhaps I could try to do so.

Quote:
 Originally Posted by Rhiessan71 As far as giving up the fight, I haven't and I do understand the math.
I was just joking, I know you aren't going to give up.

However, you evidently do not understand the math, or you wouldn't have even needed to attempt calculations about whether Gretzky's points were reduced more than some random player's. You would have known, just by knowing the actual process, that Gretzky would be favored more than any other player by HR.com's process of adjusting stats. No actual calculations would even be necessary, since it's a structural issue in the formula that someone with familiarity and understanding of the process/formual would know just by thinking about it for a moment.

Quote:
 Originally Posted by Rhiessan71 It does concern me though that you don't understand that a flaw of any sort gets magnified the further you go from the median in any formula giving results based on averages.
First, I don't necessarily agree that there's a substantial flaw. As far as actual value (proportion of goals/game scored, since goals equate to wins), the formula/process is perfect in principle and in its relative simplicity for achieving the desired results. You, and others, allege that there are flaws when applying it to certain individuals and/or subsets, for the purpose of comparing their adjusted stats in terms of difficulty and/or the likelihood that they would translate equivalently to different seasons. This is a separate issue, although I recognize its important applications. You assume not only that there is a flaw (and it would be difficult for the data not to have any flaw in such an application), but that it is of a substantial magnitude. Also, most of those who propose to correct any such flaw, seem to do so by using a very similar methodology to that which may have created any such flaw in the first place: by using some measurement and applying it equally across the board, regardless of the underlying reason for the change in measurement. As I've stated and shown examples of, this may not improve the results, and could actually distort them.

Quote:
 Originally Posted by Rhiessan71 But don't you worry, I will soon provide an extensive analysis and breaking down of all of your premises provided throughout this thread.
Whew, I was really worried that there wouldn't be a long-winded, biased post that would be as much confusing as illuminating, and I might not get sent off on some wild goose chase, or not allowed to explain some relatively simple concept for the third or fourth time. Thanks for allaying any such concerns!

Quote:
 Originally Posted by Rhiessan71 As far as the issue of exclusion...I even bolded the part of your post where you said exactly that. It's not about the boogey man and it's not about understanding adjusted stats. It's about understanding that it's not adjusted stats vs every other way. That is it about adjusted stats + everything else.
I agree that adjusted stats are far from the only way to assess or evaluate players' production and value, although I do believe they are often the best way, and certainly the fairest, most objective way. When we were talking of exclusion, it was my impression that you meant that when I (and many others) use adjusted stats, we basically exclude raw stats in the process. If so, then yes, I basically exclude the raw ingredient once the superior, finished product is available.

Last edited by Czech Your Math: 10-31-2012 at 05:03 PM.