HFBoards

Go Back   HFBoards > NHL Western Conference > Central Division > Minnesota Wild
Notices

The Annual "Rip HF For the Wild Prospect Rankings" Thread

Reply
 
Thread Tools
Old
08-30-2010, 04:05 PM
  #26
this providence
Chips in Bed Theorem
 
this providence's Avatar
 
Join Date: Oct 2008
Location: St. Paul
Posts: 9,459
vCash: 500
While looking around the "Prospects" forum I found this well written idea for the ranking system here. Much better than their current structure in my opinion. Too bad they'd never re-do it...

Quote:
Originally Posted by Qvist View Post
I don't think it makes any sense at all, but that it is on the contrary deeply flawed already at the level of conception. My apologies to the many people who put so much work into the grades, but your efforts deserve to be rewarded with a more useful result and the problem is with the basic design of the system.

It is inherently unlikely that any prospect will reach any ceiling that has been correctly determined, and whether he does so or not depends largely on factors that are not presently known and cannot be meaningfully evaluated. Even if the system managed to meaningfully assess what it tries to assess (which it clearly doesn't), it would still be deeply flawed: The real issue is not the likelihood of a prospect reaching a largely speculatively determined level, it's what sort of player he will become if he doesn't. So, player X has a potential of 7.5, but only D likelihood of achieving it. Great, what does that tell me, other than that someone has determined a theoretical upside that he is very unlikely to reach? What would be considerably more useful to know is what he's expected to reach if he doesn't reach 7.5.

What we have is a system which attempts to quantify and measure the two things about a prospect that are both the most difficult to assess correctly and the most difficult to base on any predetermined criteria. This alone makes it neccessarily largely speculative, as well as inavoidably inconsistent. In addition to this, "potential" is evidently subject to a variable definition and is expressed, insanely, along a 20-point scale. A 20-point scale. How can you expect the slightest consistency when a large number of people spread around the world get to employ a 20-point scale? In any case, it is wholly illusory to think that you can have any sort of meaningful opinion about whether the potential of an 18-year old player is 7.5 rather than 7. And. even if the system worked, it still wouldn't tell us much.

In summation, already at the level of logic and theory, the rating system is deeply and fundamentally flawed. In practice, its output is worse than useless.

The time is long overdue for a fundamental revision of it. Here are a few ideas in that regard.

1. A rating of long-term potential. This should express what level of player the prospect is expected to become if his development process is generally successful (ie, "reasonable aim" rather than "likely outcome"), as opposed to maximally successful. This has to be rough, and it has to use a small number of grades, otherwise an adequate level of consistency is impossible and the accuracy so low as to render the exercise meaningless. Four grades would be more than sufficient - NHL star, NHL high-impact player, NHL regular, Marginal or part-time NHL player. Meaningful criteria with player examples could be developed as a guideline for the grades. As much as possible, grades should be based on the level of expectation of the NHL team, insofar as that is known or can be found out.

2. Short-term development goals. This would in effect tell you how far along the curve towards that potential the player is, and define some meaningful criteria for expected progress in the season ahead. Again, reasonable aim rather than likely outcome - more or less, what the player should achieve in the upcoming season in order to stay on course towards the potential defined in #1. This ought also to be based on a finite number of predetermined options, such as "increase point total significantly", "Earn more ice time", "establish himself as a regular starter", "improve his defensive zone coverage", "put on muscle" or "make the NHL squad". The extent to which goals are reached or exceeded would then provide some indication of whether the potential ranking needed to be adjusted up or down at subsequent revisions.

3. Risk factors. Like 2, these should also be generic and selected from a finite and predetermined list of options, and should list those factors that are considered to represent significant and lingering question marks, as opposed to stages in the development process the player has not yet reached - such as "injury-prone", "lack of consistency", "small size", "questionable attitude" or "lack of foot speed".

Additional detail should go into the commentary.

In my opinion, this would be a kind of system that could reasonably aim for a useful level of accuracy, and which it could be possible to implement with a reasonable degree of consistency. It would also be far more useful than the present one.

The aim of a prospect rating system cannot be and should not be to correctly identify how good a player the prospect will become. That is impossible - nobody does or can know it, it depends on future unknown factors. What it must aim for is to correctly identify what reasonable aims are as of today. This is not the same thing. If a prospect undergoes a torrential and unexpected development during age 19-21, then a well-functioning prospect rating should have his potential set lower at age 18 than he actually turned out to have in the end. In the present system however, that would represent a failure to correctly assess potential. It shouldn't be an outcome guessing game, it should be about correctly identifying where the arrow is pointing as of now. Otherwise we might as well call an astrologer.

And consistency is an absolute key. If the system is perfect in all respects other than that each team's prospects are graded according to different understandings of the scale, that alone is enough to make the whole effort largely wasted. It is worth underlining again that the more options you have on the grading scale, the worse this problem becomes. Even if you do nothing else, for god's sake scrap the 1-10 scale with .5 decimals and replace it with something saner.

this providence is offline   Reply With Quote
Old
08-30-2010, 04:12 PM
  #27
Jarick
Moderator
Doing Nothing
 
Jarick's Avatar
 
Join Date: Jul 2007
Location: St Paul, MN
Country: United States
Posts: 23,417
vCash: 500
I think the current system is fine.

7.5D tells you that the kid has a chance to be a high impact player, but is a longshot. 6B tells you the kid has a really safe bet of being a role player. In the end, rankings don't mean squat, it's all about actual performance.

Jarick is offline   Reply With Quote
Old
08-30-2010, 04:21 PM
  #28
this providence
Chips in Bed Theorem
 
this providence's Avatar
 
Join Date: Oct 2008
Location: St. Paul
Posts: 9,459
vCash: 500
Quote:
Originally Posted by Jarick View Post
I think the current system is fine.

7.5D tells you that the kid has a chance to be a high impact player, but is a longshot. 6B tells you the kid has a really safe bet of being a role player. In the end, rankings don't mean squat, it's all about actual performance.
That's right it really means nothing and the rankings are truly too subjective by different parties to be anywhere near consistent. I mean, why do rankings at all when there's going to be so much 'play' in between how each person views a prospect. Take for example Pittsburgh's write up which was right after the Wild's:

1. (1) Eric Tangradi, LW, 8B
2. (2) Simon Despres, D, 8C
3. (NR) Beau Bennett, RW, 7.5C
4. (5) Ben Hanowski, RW, 7C
5. (NR) Tom Kuehnhackl, RW, 7C
6. (8) Robert Bortuzzo, D, 7C
7. (6) Carl Sneep, D, 7C
8. (4) Brian Strait, D, 7C
9. (3) Dustin Jeffrey, C/W, 6.5B
10. (11) Nick Petersen, RW, 7C
11. (15) Philip Samuelsson, D, 7C
12. (NR) Ken Agostino, RW, 7D
13. (10) Alex Velischek, D, 7C
14. (NR) Bryan Rust, RW, 7C
15. (7) Nick Johnson, RW, 6.5B
16. (17) Brad Thiessen, G, 7C
17. (18) Keven Veilleux, C/RW, 7D
18. (13) Alex Grant, D, 6.5C
19. (14) Nicholas D'Agostino, D, 6.5C
20. (NR) Joe Rogalski, D, 7D


1. (NR) Mikael Granlund, C, 8C
2. (2) Marco Scandella, D, 7C
3. (1) Tyler Cuma, D, 7C
4. (3) Colton Gillies, LW, 6.5B
5. (4) Matt Hackett, 7C
6. (6) Cody Almond, 6.5B
7. (NR) Johan Larsson, 6.5B
8. (NR) Casey Wellman, C/RW, 6.5C
9. (NR) Joel Broda, C, 6.5C
10. (7) Erik Haula, LW, 7D
11. (5) Maxim Noreau, D, 6.5C
12. (NR) Nate Prosser, D, 6.5C
13. (8) Justin Falk, D, 6.5C
14. (NR) Brett Bulmer, RW, 6.5C
15. (NR) Jason Zucker, LW, 6.5C
16. (NR) Chad Rau, C, 7D
17. (9) Kris Foucault, LW, 6.5D
18. (12) Carson McMillan, RW, 6C
19. (14) Jere Sallinen, LW, 6.5D
20. (10) Petr Kalus, LW 6.5D


I know the Wild's prospect pool isn't any where near the tops in the league but neither is Pittsburgh's. And look at the difference in the rankings. I don't know how much most of you all have seen some of the Pittsburgh prospects but there's a few that have no business being ranked where they are.

this providence is offline   Reply With Quote
Old
08-30-2010, 04:35 PM
  #29
Jarick
Moderator
Doing Nothing
 
Jarick's Avatar
 
Join Date: Jul 2007
Location: St Paul, MN
Country: United States
Posts: 23,417
vCash: 500
A ranking is only as good as who's writing it. Does the guy know who these players are or is he just guessing based off reputation (Gillies)?

Jarick is offline   Reply With Quote
Reply

Forum Jump


Bookmarks

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off



All times are GMT -5. The time now is 07:27 AM.

monitoring_string = "e4251c93e2ba248d29da988d93bf5144"
Contact Us - HFBoards - Archive - Privacy Statement - Terms of Use - Advertise - Top - AdChoices

vBulletin Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
HFBoards.com is a property of CraveOnline Media, LLC, an Evolve Media, LLC company. 2014 All Rights Reserved.