Author Topic: Hall of Fame probability  (Read 4611 times)

0 Members and 0 Guests are viewing this topic.

Hall of Fame probability
« on: January 27, 2009, 03:29:41 PM »

Offline Jeff

  • CelticsBlog CEO
  • Paul Silas
  • ******
  • Posts: 6673
  • Tommy Points: 301
  • ranter
this is probably a better topic for the offseason, but I re-found this list and thought it was pretty cool

http://www.basketball-reference.com/leaders/hof_prob_career.html

the way the numbers are calculated is explained here, but I admit I am too lazy to read it

http://www.basketball-reference.com/about/hof_prob.html
Faith and Sports - an essay by Jeff Clark

"Know what I pray for? The strength to change what I can, the inability to accept what I can't, and the incapacity to tell the difference." - Calvin (Bill Watterson)

Re: Hall of Fame probability
« Reply #1 on: January 27, 2009, 03:37:31 PM »

Online Donoghus

  • Global Moderator
  • Walter Brown
  • ********************************
  • Posts: 32629
  • Tommy Points: 1730
  • What a Pub Should Be
Three immediate thoughts come to mind regarding the active players:

1)  Pierce is .07 lower than T-Mac?

2)  Ray Allen at .8672 below Carter on the list?

3)  Grant Hill at .8225 despite debilitating injuries over and over again throughout this career.


2010 CB Historical Draft - Best Overall Team

Re: Hall of Fame probability
« Reply #2 on: January 27, 2009, 03:40:21 PM »

Offline Roy Hobbs

  • In The Rafters
  • The Natural
  • *********************************
  • Posts: 33333
  • Tommy Points: 6430
  • Doc could learn a thing or two from Norman Dale
So, we're looking at 14 probably Hall of Famers:

1.   Shaquille O'Neal 1.0000
2.   Tim Duncan   1.0000
3.   Kobe Bryant   1.0000
4.   Kevin Garnett   1.0000
5.   Allen Iverson   1.0000
6.   LeBron James   0.9987
7.   Dirk Nowitzki   0.9973
8.   Steve Nash   0.9970
9.   Jason Kidd   0.9855
10.   Tracy McGrady   0.9787
11.   Paul Pierce   0.9065
12.   Vince Carter   0.9034
13.   Ray Allen   0.8672
14.   Grant Hill   0.8225

On that list, I'd have serious doubts about Carter and Hill, and maybe McGrady.  Even noting that NCAA performances count towards HOF credentials, I don't think those guys deserve enshrinement.  All the rest belong, in my mind.

The next group is interesting:

15.   Tony Parker   0.3976
16.   Carmelo Anthony   0.2653
17.   Elton Brand   0.1955
18.   Shawn Marion   0.1673
19.   Chauncey Billups0.1486
20.   Amare Stoudemire0.1414
21.   Chris Bosh   0.0946
22.   Baron Davis   0.0898
23.   Jermaine O'Neal   0.0554
24.   Antawn Jamison   0.0539
25.   Manu Ginobili   0.0538
26.   Yao Ming   0.0538
27.   Richard Hamilton0.0513
28.   Carlos Boozer   0.0364
29.   Rasheed Wallace   0.0245
30.   Peja Stojakovic   0.0225

There are some potential Hall of Famers on there, but I would agree:  as of now, none of them have the credentials to get in.

Also of note:  Chris Webber has a listed probability of 0.9128.  I can't imagine he gets in, but even the biggest Webber fan couldn't imagine him having a 91% chance at enshrinement.  Similarly, Alonzo Mourning is 0.8850.  In terms of eligible guys who aren't in yet, I think the highest are Bernard King at 0.7910 and Jo Jo White at 0.7844.  DJ is only at 0.4983.

Interesting stuff.



All the negativity in this town sucks. It sucks, and it stinks, and it sucks. - Rick Pitino

Portland CrotoNats:  2009 CB Draft Champions

Re: Hall of Fame probability
« Reply #3 on: January 27, 2009, 03:45:16 PM »

Offline EarthBall

  • Jrue Holiday
  • Posts: 361
  • Tommy Points: 110
I hate hate hate this.

First, the author is clearly using a technique called data mining. Just looking for the best fitting model with little mechanistic understanding. By doing a regression on every combination of variables, he's finding those variables that correlate best with HOF entry. Correlation does not equal causation. This is why height is a covariate. I don't really believe that HOF voters weigh height of the players that heavily.

Second, MVPs get too much weight. Yes, you could argue that every MVP winner is in the HOF. But is that because he won an MVP or because his other stats and championships won are higher?

Third, all star game selections? Really? Considering Yi was almost voted in by the (Chinese) fans and that T-Mac and AI continue to start the games doesn't give much credence to this predictor.

Fourth, I love statistics (as you know), but there are some things they can't show. Defense (besides steals and blocks - which are used in this formula), intangibles and leadership abilities, for example. That's why trying to predict a player's HOF chances using this formula is just stupid.

Looking at some of the rankings gives me a heart attack:

Reggie Miller (185)
Artis Gilmore (147)
Chris Mullen (145)

are all lower than

Antoine Walker (133)
Jack Sikma (105)
Gilbert Arenas (92)

And the all-time assists and steals leader John Stockton at 75 with an 81% chance of HOF entry? Are you freaking kidding me?
« Last Edit: January 27, 2009, 03:56:14 PM by EarthBall »

Re: Hall of Fame probability
« Reply #4 on: January 27, 2009, 03:45:37 PM »

Offline CelticsWhat35

  • Jim Loscutoff
  • **
  • Posts: 2954
  • Tommy Points: 356
McGrady is waaaaaaaay too high.

Re: Hall of Fame probability
« Reply #5 on: January 27, 2009, 04:01:24 PM »

Offline xmuscularghandix

  • Tiny Archibald
  • *******
  • Posts: 7620
  • Tommy Points: 280
Dennis Rodman works out to .2656

What does that say to you about this system?

Re: Hall of Fame probability
« Reply #6 on: January 27, 2009, 04:04:21 PM »

Offline fairweatherfan

  • Johnny Most
  • ********************
  • Posts: 20738
  • Tommy Points: 2365
  • Be the posts you wish to see in the world.
I hate hate hate this.

First, the author is clearly using a technique called data mining. Just looking for the best fitting model with little mechanistic understanding. By doing a regression on every combination of variables, he's finding those variables that correlate best with HOF entry. Correlation does not equal causation. This is why height is a covariate. I don't really believe that HOF voters weigh height of the players that heavily.

Second, MVPs get too much weight. Yes, you could argue that every MVP winner is in the HOF. But is that because he won an MVP or because his other stats and championships won are higher?

Third, all star game selections? Really? Considering Yi was almost voted in by the (Chinese) fans and that T-Mac and AI continue to start the games doesn't give much credence to this predictor.

Fourth, I love statistics (as you know), but there are some things they can't show. Defense (besides steals and blocks - which are used in this formula), intangibles and leadership abilities, for example. That's why trying to predict a player's HOF chances using this formula is just stupid.

I 2nd this, TP - this is a blind model that just looks how closely a player fits the typical profile of existing Hall of Famers, not what actually causes them to get in.  It needs at the very least several more playoff factors - TMac is nowhere near in because he's never made it past the first round, but give him a Finals trip or a couple conference finals, and he's a lot closer even without a title.  Hard to equalize that across the decades though.

The All-Star thing is a good predictor though - even though sometimes players get voted in who don't deserve to be in the game at all, they are rare exceptions.  Being selected by fans is a good indicator of overall popularity, and by coaches a good indicator of league-wide respect, and both of those are HOF factors. 

Re: Hall of Fame probability
« Reply #7 on: January 27, 2009, 04:12:29 PM »

Offline NoraG1

  • Bill Walton
  • *
  • Posts: 1350
  • Tommy Points: 108
I hate hate hate this.

First, the author is clearly using a technique called data mining. Just looking for the best fitting model with little mechanistic understanding. By doing a regression on every combination of variables, he's finding those variables that correlate best with HOF entry. Correlation does not equal causation. This is why height is a covariate. I don't really believe that HOF voters weigh height of the players that heavily.

Second, MVPs get too much weight. Yes, you could argue that every MVP winner is in the HOF. But is that because he won an MVP or because his other stats and championships won are higher?

Third, all star game selections? Really? Considering Yi was almost voted in by the (Chinese) fans and that T-Mac and AI continue to start the games doesn't give much credence to this predictor.

Fourth, I love statistics (as you know), but there are some things they can't show. Defense (besides steals and blocks - which are used in this formula), intangibles and leadership abilities, for example. That's why trying to predict a player's HOF chances using this formula is just stupid.

I 2nd this, TP - this is a blind model that just looks how closely a player fits the typical profile of existing Hall of Famers, not what actually causes them to get in.  It needs at the very least several more playoff factors - TMac is nowhere near in because he's never made it past the first round, but give him a Finals trip or a couple conference finals, and he's a lot closer even without a title.  Hard to equalize that across the decades though.

The All-Star thing is a good predictor though - even though sometimes players get voted in who don't deserve to be in the game at all, they are rare exceptions.  Being selected by fans is a good indicator of overall popularity, and by coaches a good indicator of league-wide respect, and both of those are HOF factors. 

All Star game players that do not deserve to go are becoming less and less rare. Yi almost got to be a starter. That alone says that all star game selections should not be much of a factor. Fans don't always vote for the best players.

Re: Hall of Fame probability
« Reply #8 on: January 27, 2009, 04:25:58 PM »

Offline Roy Hobbs

  • In The Rafters
  • The Natural
  • *********************************
  • Posts: 33333
  • Tommy Points: 6430
  • Doc could learn a thing or two from Norman Dale
All Star game players that do not deserve to go are becoming less and less rare. Yi almost got to be a starter. That alone says that all star game selections should not be much of a factor. Fans don't always vote for the best players.

While all this is true, when HOF voters are weighing Player A with 5 all-star appearances, and Player B with 12, in all likelihood they're going to go with Player B.  There's a strong correlation between perennial all-star appearances and Hall of Fame voting, and I suspect voters will use this particular short cut long into the future.

All the negativity in this town sucks. It sucks, and it stinks, and it sucks. - Rick Pitino

Portland CrotoNats:  2009 CB Draft Champions

Re: Hall of Fame probability
« Reply #9 on: January 27, 2009, 04:33:11 PM »

Offline papa shuttlesworth

  • Jrue Holiday
  • Posts: 333
  • Tommy Points: 46
I agree about the limitations of this site, but it is pretty interesting.  It can also help us define what makes someone HOF worthy or not, such as WHY is Vince not worthy when he, Pierce and Allen all have approximately the same rating.  Or at least fuel discussion.

But for all of its limitations, the site can be accurate.  Scalabrine, for example, is ranked 147th among active players.  So there is a chance...

Re: Hall of Fame probability
« Reply #10 on: January 27, 2009, 04:51:10 PM »

Offline Big Ticket

  • Bailey Howell
  • **
  • Posts: 2356
  • Tommy Points: 561
  • The good ole days.
First, the author is clearly using a technique called data mining. Just looking for the best fitting model with little mechanistic understanding. By doing a regression on every combination of variables, he's finding those variables that correlate best with HOF entry. Correlation does not equal causation. This is why height is a covariate. I don't really believe that HOF voters weigh height of the players that heavily.

EarthBall... you just dropped way too many buzz words from my CFA Statistics review.  Darn you!! (TP for agreeing on your points though!)


"It ain't about me.  It's about us."  - KG, interview with John Thompson, 2005 All Star Game.

Re: Hall of Fame probability
« Reply #11 on: January 27, 2009, 04:57:19 PM »

Offline KungPoweChicken

  • Bailey Howell
  • **
  • Posts: 2102
  • Tommy Points: 228
Why is Dirk so high on that list? It must be because he was gifted the MVP one year.

Re: Hall of Fame probability
« Reply #12 on: January 27, 2009, 04:59:54 PM »

Offline Brendan

  • Jim Loscutoff
  • **
  • Posts: 2990
  • Tommy Points: 72
He's pretty clear in his write up (at least last time I wrote it) that he's simply predicting based on how HoF voters have voted in the past. All the points against his method, probably are better proof that HoF voters are idiots, rather than proof that his method has little predictive value.

Or what Hobbs said - HoF voters will be using short cuts like these for a long time to go.

Re: Hall of Fame probability
« Reply #13 on: January 27, 2009, 05:05:22 PM »

Offline BballTim

  • Dave Cowens
  • ***********************
  • Posts: 23724
  • Tommy Points: 1123
Why is Dirk so high on that list? It must be because he was gifted the MVP one year.

  He also got to the finals which probably helps.

Re: Hall of Fame probability
« Reply #14 on: January 27, 2009, 05:29:19 PM »

Offline Redz

  • Punner
  • Global Moderator
  • Red Auerbach
  • *******************************
  • Posts: 31741
  • Tommy Points: 3845
  • Yup
Why is Dirk so high on that list? It must be because he was gifted the MVP one year.

  He also got to the finals which probably helps.

Then choked his face off!
Yup