 |
APBRmetrics
The statistical revolution will not be televised.
|
| View previous topic :: View next topic |
| Author |
Message |
mathayus
Joined: 15 Aug 2005
Posts: 213
|
 Posted: Sat Mar 05, 2011 11:46 pm Post subject: |
 |
|
| bchaikin wrote: | You're surprised that a 4 time all-star who never made all-NBA didn't crack the top 100?
no, just the cheeks vs. nash comparison, since the list had nash rated so high...
Care to expound?
---g------min--ScFG%--ast-----st------to----player
1101--34845--56.2--7392--2310--2268---cheeks
1073--33606--59.7--9055--0813--3038---nash
each has played 15 years, close to the same number of minutes. nash has 1663 more assists, cheeks had 1497 more steals with 770 less turnovers, or 2267 more zero point team possessions. each team possession is worth approximately 1 point, so that's 2267 less team possession points for nash. from a team perspective an assist is certainly not worth anywhere near a point...
nash may have scored more but when cheeks did not score does not mean his team did not score. but a zero point team possession costs your team a possession/point, and cheeks is up on nash 2267 team zero point possessions, or about 2.1 per game over their entire careers...
this not to mention that outside of steals, blocked shots, and def rebs cheeks was the vastly superior defender (4 times all-D 1st team, 1 time all-D 2nd team), and this was during the days of great defensive guards like dennis johnson, michael cooper, and t.r. dunn in their prime...
for the 8 year period of 78-79 to 85-86, the 76ers were the league's 2nd best defensive team (lowest pts/poss allowed, milwaukee was 1st), and cheeks played 1/8 to 1/7 of the team's total minutes (2nd most minutes to erving). i'd say he was a key reason why they were so dominant defensively for so long...
nash was in the league for 8 seasons (96-97 to 03-04, ages 22-30) before he won his first mvp. how many people would consider him an all-time great based on what he did his first 8 years in the league?...
also he got a ton of votes for mvp from 04-05 to 07-08. but from 08-09 to 10-11 he got few. yet his stats for both 3 year stretchs were virtually identical... |
Interesting. Good to hear your perspective.
_________________
http://asubstituteforwar.wordpress.com/ |
|
| Back to top |
|
 |
Mike G
Joined: 14 Jan 2005
Posts: 3618
Location: Hendersonville, NC
|
| Posted: Sun Mar 06, 2011 7:05 am Post subject: |
|
|
I don't know if there's any way to do it, but I wonder if we might at least attempt to :
1) think about current careers, or at least (recent) players who played with and against current players.
2) judge players by what they've done, not by what they might do eventually.
3) judge players by their individual merits as players, not by how well they might fit into your Dream Team lineup.
#3 may mean you judge their effectiveness in the lineups they've actually served, not in some hypothetical. This should make it easier?
_________________
`
36% of all statistics are wrong |
|
| Back to top |
|
|
Mike G
Joined: 14 Jan 2005
Posts: 3618
Location: Hendersonville, NC
|
| Posted: Fri Mar 11, 2011 11:33 am Post subject: |
|
|
Referring back to the opening post, does it seem about right that there would be 12 current players among the all-time top 50?
And 24 of the top 100 ?
Debate about where current players are ranked relative to one another?
_________________
`
36% of all statistics are wrong |
|
| Back to top |
|
|
mathayus
Joined: 15 Aug 2005
Posts: 213
|
| Posted: Sat Mar 12, 2011 5:56 pm Post subject: |
|
|
| Mike G wrote: | Referring back to the opening post, does it seem about right that there would be 12 current players among the all-time top 50?
And 24 of the top 100 ?
Debate about where current players are ranked relative to one another? |
Active players at this point essentially cover 20 years out of a 60 year old league. We should expect that this group would make up about 1/3rd of the top players by career's end if we simply go by dominance over contemporaries, so for them to make up 1/4th of the top players at this point doesn't seem at all unreasonable.
Debate about relative ranking?
Well the most obvious disagreement for me is Nash so low. From experience though, I'm guessing most people here don't see it like I do.
On a methodology level, a fundamental part of how I rate players' careers is the idea that 1+1 != 2+0. Meaning, a top 20 level player who played for 40 years is not going to end up #1 on my list. At the most quantifiable level this means considering value above replacement.
I'm guessing you aren't doing this, am I wrong?
_________________
http://asubstituteforwar.wordpress.com/ |
|
| Back to top |
|
|
BobboFitos
Joined: 21 Feb 2009
Posts: 201
Location: Cambridge, MA
|
| Posted: Sat Mar 12, 2011 6:16 pm Post subject: |
|
|
| mathayus wrote: | | Mike G wrote: | Referring back to the opening post, does it seem about right that there would be 12 current players among the all-time top 50?
And 24 of the top 100 ?
Debate about where current players are ranked relative to one another? |
Active players at this point essentially cover 20 years out of a 60 year old league. We should expect that this group would make up about 1/3rd of the top players by career's end if we simply go by dominance over contemporaries, so for them to make up 1/4th of the top players at this point doesn't seem at all unreasonable.
Debate about relative ranking?
Well the most obvious disagreement for me is Nash so low. From experience though, I'm guessing most people here don't see it like I do.
On a methodology level, a fundamental part of how I rate players' careers is the idea that 1+1 != 2+0. Meaning, a top 20 level player who played for 40 years is not going to end up #1 on my list. At the most quantifiable level this means considering value above replacement.
I'm guessing you aren't doing this, am I wrong? |
I don't know if this is a fair way to do things. The NBA currently has 30 teams and draws from a significantly broader population. (International game etc.) In the early 60s, for example, there were 8 teams. Being the 5th best player now is far more impressive then being the 5th best player in 1960, using that same logic.
It probably works better to shift the NBA into eras, similar to how Bref has it - pre shotclock, when basketball was truly a different game. pre 3pt line. and you could maybe divide the past 30 years in terms of pre lockout and post lockout.
Pre shotclock, you have a small group of guys who deserve to be on a top 100 list; something like Neil Johnston, Dolph Schayes, George Mikan. I don't know NBA history that well so maybe there are a few other greats who would make a fringe 100.
Current players "should" probably make up 40%+ of a top 100 list...
_________________
http://pointsperpossession.com/
@PPPBasketball |
|
| Back to top |
|
|
mathayus
Joined: 15 Aug 2005
Posts: 213
|
| Posted: Sat Mar 12, 2011 6:36 pm Post subject: |
|
|
@BobboFitos. Oh, I'm not saying it is the perfect estimate, I just don't see anything ridiculous about it. I personally would only use pure statistical measurements as a baseline anyway, and would adjust further based on my perceptions of the change in competition level over eras among other things - though I realize that's an unsatisfactory approach to anyone trying to achieve a complete model through mathematics.
_________________
http://asubstituteforwar.wordpress.com/ |
|
| Back to top |
|
|
Mike G
Joined: 14 Jan 2005
Posts: 3618
Location: Hendersonville, NC
|
| Posted: Mon Mar 14, 2011 7:53 am Post subject: |
|
|
| mathayus wrote: | ... a top 20 level player who played for 40 years is not going to end up #1 on my list. At the most quantifiable level this means considering value above replacement.
I'm guessing you aren't doing this, am I wrong? |
Not exactly. I generate career 'equivalent totals' and multiply the square root of eTotals by the T rate.
This product has no units, so they're just generic 'credits', or 'brownie points'.
If 2 players have identical productivity (T) and one player did it for twice as long (minutes), he won't have twice as many credits, but 1.414 times as many (Sqrt of 2).
He'd have to have 4 times as long a career to have twice as many career 'credits'.
Then there are playoff sub-careers. The norm is about 6-7% of minutes in playoffs; better players tend to 10%. That's a 9:1 ratio of RS to PO min/eTotals.
But when you use square roots, it becomes a 3:1 ratio -- playoffs then are 1/4 of a player's career, rather than 1/10.
For these reasons, LeBron ranks over Kidd, despite Kidd having more than twice as many games.
If both careers ended now, I'd rank LeBron's as the more significant career.
Using Win Shares, not counting playoffs, Nash ranks 9th among active players. If playoff WS are 1 or 2 times as important, he ranks 10th. If you weigh playoffs 3 or 4 times as heavily, he's 11th.
My system has him 14th, as less reward is given for high shooting%. I also don't think Nowitzki has been better than Garnett, nor Manu better than Kobe.
_________________
`
36% of all statistics are wrong |
|
| Back to top |
|
|
Mike G
Joined: 14 Jan 2005
Posts: 3618
Location: Hendersonville, NC
|
| Posted: Mon Mar 14, 2011 8:13 am Post subject: |
|
|
| BobboFitos wrote: | ... The NBA currently has 30 teams and draws from a significantly broader population. (International game etc.) ...
Current players "should" probably make up 40%+ of a top 100 list... |
Would this be your assessment IF we know the extent of current players' ultimate careers? Which, of course, we don't know.
We can look back at a distant season, say 1979, and count how many Hall of Famers or top50s or whatever were active. But we can't do that now. The uncertainty isn't just in some arbitrary future voting, it's in the fact that the future isn't here yet.
Grant Hill, McGrady, Yao, didn't/won't have the future they might have had.
If you include players in an 'eventual' top100 that you wouldn't include if their careers ended today, you're speculating. Boozer, Carmelo, Paul, DWilliams, Bosh, Josh Smith ... may or may not eventually meet your standards.
_________________
`
36% of all statistics are wrong |
|
| Back to top |
|
|
mathayus
Joined: 15 Aug 2005
Posts: 213
|
| Posted: Mon Mar 14, 2011 1:20 pm Post subject: |
|
|
| Mike G wrote: |
Not exactly. I generate career 'equivalent totals' and multiply the square root of eTotals by the T rate.
This product has no units, so they're just generic 'credits', or 'brownie points'.
If 2 players have identical productivity (T) and one player did it for twice as long (minutes), he won't have twice as many credits, but 1.414 times as many (Sqrt of 2).
He'd have to have 4 times as long a career to have twice as many career 'credits'.
Then there are playoff sub-careers. The norm is about 6-7% of minutes in playoffs; better players tend to 10%. That's a 9:1 ratio of RS to PO min/eTotals.
But when you use square roots, it becomes a 3:1 ratio -- playoffs then are 1/4 of a player's career, rather than 1/10.
For these reasons, LeBron ranks over Kidd, despite Kidd having more than twice as many games.
If both careers ended now, I'd rank LeBron's as the more significant career.
Using Win Shares, not counting playoffs, Nash ranks 9th among active players. If playoff WS are 1 or 2 times as important, he ranks 10th. If you weigh playoffs 3 or 4 times as heavily, he's 11th.
My system has him 14th, as less reward is given for high shooting%. I also don't think Nowitzki has been better than Garnett, nor Manu better than Kobe. |
Ah okay. I do think the square root method is a good compromise between avoiding too much emphasis on longevity and too much complexity which kills intuitive interpretation of the model.
_________________
http://asubstituteforwar.wordpress.com/ |
|
| Back to top |
|
|
BobboFitos
Joined: 21 Feb 2009
Posts: 201
Location: Cambridge, MA
|
| Posted: Mon Mar 14, 2011 1:42 pm Post subject: |
|
|
| Mike G wrote: | | BobboFitos wrote: | ... The NBA currently has 30 teams and draws from a significantly broader population. (International game etc.) ...
Current players "should" probably make up 40%+ of a top 100 list... |
Would this be your assessment IF we know the extent of current players' ultimate careers? Which, of course, we don't know.
We can look back at a distant season, say 1979, and count how many Hall of Famers or top50s or whatever were active. But we can't do that now. The uncertainty isn't just in some arbitrary future voting, it's in the fact that the future isn't here yet.
Grant Hill, McGrady, Yao, didn't/won't have the future they might have had.
If you include players in an 'eventual' top100 that you wouldn't include if their careers ended today, you're speculating. Boozer, Carmelo, Paul, DWilliams, Bosh, Josh Smith ... may or may not eventually meet your standards. |
True/good point about speculation with regard to the fringier guys - we can assume they will eventually be on a top 100 list, but probably aren't there yet.
_________________
http://pointsperpossession.com/
@PPPBasketball |
|
| Back to top |
|
|
Mike G
Joined: 14 Jan 2005
Posts: 3618
Location: Hendersonville, NC
|
| Posted: Tue Mar 15, 2011 3:18 am Post subject: |
|
|
http://www.slamonline.com/online/the-magazine/toc/2011/02/500-greatest-on-sale-now/
All-time top 500 NBA players list, put together by Ben Osborne and | Quote: | | Russ Bengtson, Michael Bradley, Franklyn Calle, David Cassilo, Adam Figman, Ryne Nelson, Alan Paul, Khalid Salaam, Al Stark, Tzvi Twersky and Lang Whitaker. |
Here are players current as of last season, with their Slam ranking, and then mine for comparison. | Code: | slam alltime ranks MG slam alltime ranks MG
4 Shaquille O'Neal 4 152 Shawn Marion 82
8 Tim Duncan 6 154 Yao Ming 133
10 Kobe Bryant 15 156 Rasheed Wallace 61
28 Jason Kidd 28 171 Chris Bosh 232
30 Kevin Garnett 14 174 Jermaine O'Neal 110
31 LeBron James 20 177 Elton Brand 86
40 Allen Iverson 40 183 Gilbert Arenas 219
49 Dwyane Wade 41 184 Joe Johnson 287
50 Steve Nash 60 199 Baron Davis 81
55 Dirk Nowitzki 25 201 Lamar Odom 84
77 Paul Pierce 33 206 Carlos Boozer 111
79 Ray Allen 71 208 Marcus Camby 140
86 Dwight Howard 78 223 Rip Hamilton 117
96 Pau Gasol 54 224 Michael Redd 365
97 Tracy McGrady 39 225 Peja Stojakovic 239
98 Grant Hill 75 230 Jason Richardson 240
99 Vince Carter 50 236 Antawn Jamison 168
105 Carmelo Anthony 142 244 Zydrunas Ilgauskas 154
107 Chris Paul 143 251 Mike Bibby 173
123 Ben Wallace 131 255 Andre Miller 149
124 Chauncey Billups 55 261 David West 327
128 Deron Williams 226 273 Zach Randolph 265
139 Tony Parker 104 287 Michael Finley 171
140 Ron Artest 215 295 Jerry Stackhouse 213
143 Manu Ginobili 70 299 Kenyon Martin 184
144 Amar'e Stoudemire 101 302 Rashard Lewis 174
|
I've also got Brad Miller, Josh Smith, and Mehmet Okur in my top 300, and Slam's 500 missed them.
_________________
`
36% of all statistics are wrong |
|
| Back to top |
|
|
Mike G
Joined: 14 Jan 2005
Posts: 3618
Location: Hendersonville, NC
|
| Posted: Tue Mar 15, 2011 2:34 pm Post subject: |
|
|
It may be that some of the big differences in my 'career' rank and Slam's 'best player' rank is explained by their arbitrary cutoff of 5 NBA seasons, vs my arbitrary cutoff of 10,000 equivalent points+rebounds+etc.
Slam experts say Michael Redd is a 'better player' than Stojakovic, Jamison, Ilgauskas, Bibby, Andre Miller, etc. And I say those players all have had better careers.
Carmelo surely can't have had almost as great a career as Carter; nor can Deron Williams' be that close to Billups'. If they haven't produced the career, how can they be as good?
And no clue how Artest > Manu, Amar'e, Marion, Brand, etc. Not in peak value, not in consistency, not anything.
_________________
`
36% of all statistics are wrong |
|
| Back to top |
|
|
Mike G
Joined: 14 Jan 2005
Posts: 3618
Location: Hendersonville, NC
|
| Posted: Fri Mar 18, 2011 10:51 am Post subject: |
|
|
Wade #49 vs McGrady #97, according to Slam.
TMac is twice as far from #1 ?
Win Shares: McGrady 93 (plus 4.6 playoff WS), Wade 80 (10 PO).
That's 72nd all-time (RS) vs #125 ; mitigated by playoff disparity.
Of the top 100 in WS, McGrady ranks 21st in PER, just behind Oscar, West, Baylor; ahead of Moses, Gasol, Erving (NBA), Schayes, Stockton, Gervin, Lanier...
Wade heads up the 2nd 100, followed by Chris Paul.
I have to wonder if a few bad years negates the good years, in these opinions.
http://bkref.com/tiny/8WYUX
McGrady actually has the higher playoff PER.
_________________
`
36% of all statistics are wrong |
|
| Back to top |
|
|
Mike G
Joined: 14 Jan 2005
Posts: 3618
Location: Hendersonville, NC
|
| Posted: Fri Mar 18, 2011 11:11 am Post subject: |
|
|
Kobe #10 and Garnett #30 ?
RS Win Shares: KG 173 (#11), Kobe 155 (19)
Playoffs to Kobe, 26-12
http://bkref.com/tiny/wVmK8
Identical RS PER (23.5), tied at #16.
Playoffs to Kobe, by a whisker (PER and WS/48).
Of course, KG's 'over the hill' years are over-represented in playoffs.
Career RS WS/48 : edge to KG, .191-.187, #20 vs #23.
_________________
`
36% of all statistics are wrong |
|
| Back to top |
|
|
bchaikin
Joined: 27 Jan 2005
Posts: 690
Location: cleveland, ohio
|
| Posted: Fri Mar 18, 2011 12:38 pm Post subject: |
|
|
Here are players current as of last season, with their Slam ranking, and then mine for comparison.
where does mo cheeks rank in your list vs theirs? |
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
Powered by phpBB © 2001, 2005 phpBB Group
|
FAQ
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
