|
APBRmetrics The statistical revolution will not be televised.
|
View previous topic :: View next topic |
Author |
Message |
DSMok1
Joined: 05 Aug 2009 Posts: 611 Location: Where the wind comes sweeping down the plains
|
Posted: Thu Dec 16, 2010 12:02 pm Post subject: Value of Contracts and Trade Value--ASPM |
|
|
As the next step in the progression of developing Advanced Statistical Plus/Minus, I have now reached the point of evaluating contracts.
The basic method:
1. Create a true-talent estimation for the current year (this was already done.
2. Apply aging to estimate true-talent for coming years.
3. Estimate inflation (I chose 5%--that could be high or low, depending on the CBA and actual real-world inflation rates)
4. Generate estimation of player VORP, based on future ASPM.
5. Generate dollar value of player in future years, adjusting for inflation.
6. Subtract contract price price from future estimated dollar value.
7. Convert future surplus value back into today's money.
8. Sum for each player.
As an example:
Marc Gasol's estimated true talent is +2.90. He's under contract for this year and next year. Including inflation and aging (he's 26, so aging is minor), he is expected to produce $18.2 million in value this year and $18.8 million next year. His contract is for $3.6 million this year and $4.5 million next year. This is a surplus of $14.6 million this year and $14.3 million next year. Remove the 31% of the season that has already been played this year, and adjust for the 5% inflation for next year and the numbers become $10 million and $13.6 million.
Thus Marc Gasol's current surplus value is $23.6 million dollars. That's a really good contract for the team; in fact, one of the top 10 or 12 over the next couple of years. Of course, it then ends and who knows what he'll get on the free agent market!
One thing I didn't account for (I'm not an accountant) is the uncertainty level of the forecast. This is the mean expectation; some players will have greater variability than others.
Here is a list of the top 30 most valuable contracts (a trade value table, if you will), and the 20 worst contracts:
Code: | The 30 Best Contracts
Tm Player Age Contract Surplus Value
MIA LeBron James 26 $108,189,910
MIA Dwyane Wade 29 $77,540,362
OKC Kevin Durant 22 $73,409,011 EDIT: Horford here
BOS Rajon Rondo 24 $65,222,619
NOH Chris Paul 25 $64,819,151
MIL Brandon Jennings 21 $58,938,806
WAS John Wall 20 $52,925,584
SAC Tyreke Evans 21 $49,889,419
OKC Russell Westbrook 22 $49,688,307
TOR Amir Johnson 23 $49,528,421
GSW Stephen Curry 22 $47,584,050
LAC Blake Griffin 21 $44,982,949
PHI Jrue Holiday 20 $43,956,754
SAS DeJuan Blair 21 $41,323,371
CHI Joakim Noah 25 $38,831,830
POR Nicolas Batum 22 $35,934,207
CHI Derrick Rose 22 $34,710,251
MIN Kevin Love 22 $34,543,821
NJN Derrick Favors 19 $34,312,179
SAS George Hill 24 $32,781,522
NJN Brook Lopez 22 $32,027,893
OKC James Harden 21 $30,109,460
UTA Paul Millsap 25 $28,851,162
ATL Josh Smith 25 $28,846,073
CHA Tyrus Thomas 24 $28,643,579
ATL Al Horford 24 $26,720,198
EDIT: extension not included, actually $68,000,000 or so.
SAS Manu Ginobili 33 $25,887,887
CHI Ronnie Brewer 25 $24,432,905
OKC Serge Ibaka 21 $24,376,784
IND Roy Hibbert 24 $23,972,281
The 20 Worst Contracts:
Tm Player Age Contract Surplus Value
CHA DeSagana Diop 29 -$17,464,616
PHI Andres Nocioni 31 -$17,601,735
LAC Chris Kaman 28 -$17,941,262
DET Charlie Villanueva 26 -$20,714,162
MIL John Salmons 31 -$22,470,183
PHO Josh Childress 27 -$23,865,463
DEN Al Harrington 30 -$23,931,006
MIL Corey Maggette 31 -$24,190,972
PHI Elton Brand 31 -$24,320,238
MIL Drew Gooden 29 -$26,342,167
NJN Travis Outlaw 26 -$28,461,611
DET Richard Hamilton 32 -$28,518,019
ORL Rashard Lewis 31 -$28,967,734
PHO Hedo Turkoglu 31 -$29,014,244
TOR Andrea Bargnani 25 -$30,376,112
WAS Gilbert Arenas 29 -$35,442,228
NOH Emeka Okafor 28 -$36,946,912
DAL Brendan Haywood 31 -$37,715,177
DET Ben Gordon 27 -$38,336,260
ATL Joe Johnson 29 -$57,894,148 |
Note: all contract data is from ShamSports.com; some of the newest contracts have not been added in yet. _________________ GodismyJudgeOK.com/DStats
Twitter.com/DSMok1
Last edited by DSMok1 on Thu Dec 16, 2010 12:57 pm; edited 2 times in total |
|
Back to top |
|
|
DSMok1
Joined: 05 Aug 2009 Posts: 611 Location: Where the wind comes sweeping down the plains
|
Posted: Thu Dec 16, 2010 12:25 pm Post subject: |
|
|
Here's quick analysis of the recent 3-player trade:
Terrence Williams: Surplus value of: -$8.5 million
First Round pick, around 17th: +$7.7 Million
Sasha Vujacic: -$2.6 million
First Round pick, around 28th: +$4 million
Joe Smith: -$0.6 million
Basically, ASPM hates T-Will. Morey sees upside. I have T-Will as a -2.8 player right now (replacement level is -3.5) and since he's already 23 don't like his aging enough to make up for it.
Draft Pick Value from: http://sonicscentral.com/apbrmetrics/viewtopic.php?p=31603#31603 _________________ GodismyJudgeOK.com/DStats
Twitter.com/DSMok1 |
|
Back to top |
|
|
DSMok1
Joined: 05 Aug 2009 Posts: 611 Location: Where the wind comes sweeping down the plains
|
Posted: Thu Dec 16, 2010 1:03 pm Post subject: |
|
|
Team total surplus contract values:
Code: | Team Value Total Player-Years
MIA $171,568,781 41
OKC $159,082,891 50
SAS $ 85,063,998 39
CHI $ 81,588,271 42
POR $ 72,076,223 38
BOS $ 68,583,329 32
SAC $ 48,818,312 38
UTA $ 43,774,277 26
MIL $ 40,929,336 40
PHI $ 38,031,895 40
IND $ 32,429,646 36
LAC $ 26,349,453 36
GSW $ 25,646,201 42
NYK $ 23,413,377 33
MEM $ 20,667,007 40
ATL $ 19,602,337 38
NOH $ 17,515,800 36
DEN $ 15,995,233 30
HOU $ 15,865,222 40
NJN $ 15,488,568 42
ORL $ 13,027,084 36
TOR $ (1,412,880) 39
LAL $ (1,534,494) 38
CLE $(12,104,601) 29
WAS $(18,780,112) 44
CHA $(19,287,683) 38
PHO $(28,032,087) 32
MIN $(32,690,510) 49
DAL $(50,799,294) 32
DET $(60,783,166) 35 |
Miami and OKC are in incredible shape. _________________ GodismyJudgeOK.com/DStats
Twitter.com/DSMok1 |
|
Back to top |
|
|
Ilardi
Joined: 15 May 2008 Posts: 265 Location: Lawrence, KS
|
Posted: Thu Dec 16, 2010 1:25 pm Post subject: |
|
|
This is really interesting work!
I may have missed it before, but can you provide a bit of quick elaboration on how you determined player value vis-a-vis SPM? What, for example, would be your estimated dollar value of a player who has a one-year contract at an SPM level of 0.0? What about at, say, +3.5 or -3.5? |
|
Back to top |
|
|
back2newbelf
Joined: 21 Jun 2005 Posts: 279
|
Posted: Thu Dec 16, 2010 1:28 pm Post subject: |
|
|
How are Dallas and San Antonio that far from each other on that list. What's so different about their situation? |
|
Back to top |
|
|
DSMok1
Joined: 05 Aug 2009 Posts: 611 Location: Where the wind comes sweeping down the plains
|
Posted: Thu Dec 16, 2010 1:43 pm Post subject: |
|
|
Ilardi wrote: | This is really interesting work!
I may have missed it before, but can you provide a bit of quick elaboration on how you determined player value vis-a-vis SPM? What, for example, would be your estimated dollar value of a player who has a one-year contract at an SPM level of 0.0? What about at, say, +3.5 or -3.5? |
It will all be on the spreadsheet, which I will release to the public when I polish it a little.
I set replacement level at -3.5 for ASPM.
I ran an estimate of how many minutes a player at each level of ASPM will play; the max allowed was about 2500.
I calculated (ASPM + 3.5)*Minutes played for the league, and took the total value of all contracts in the league and divided the ASPM VORP * Minutes by that total dollar value to get a dollar value for (ASPM + 3.5)*Minute.
Then, I projected each player's ASPM, generated an estimated minutes played (obviously, the minutes are for a league-average team), and calculated each player's VORP*Min; multiplied by the dollar per unit, this yields a league-wide estimated value.
+3.5 is projected to play 2477 minutes (the max) and is worth right at $20,000,000.
+0 is projected to play 1650 minutes and is worth right at $6,600,000
-3.5 is projected to play 106 minutes and is worth $0. (Replacement level). _________________ GodismyJudgeOK.com/DStats
Twitter.com/DSMok1 |
|
Back to top |
|
|
DSMok1
Joined: 05 Aug 2009 Posts: 611 Location: Where the wind comes sweeping down the plains
|
Posted: Thu Dec 16, 2010 2:06 pm Post subject: |
|
|
back2newbelf wrote: | How are Dallas and San Antonio that far from each other on that list. What's so different about their situation? |
Several things: Roddy Beaubois was missing because he hasn't played this year.
Dallas has only 1 good, cheap, young player. Dallas also has Brendan Haywood.
San Antonio has 2 really good cheap, young players and 1 superstar that isn't paid like one.
Code: | DAL Rodrigue Beaubois 22 $34,244,171
DAL Jason Kidd 37 $10,287,986
DAL Dirk Nowitzki 32 $8,288,214
DAL Brian Cardinal 33 ($523,779)
DAL Steve Novak 27 ($593,210)
DAL Tyson Chandler 28 ($624,878)
DAL Jose Barea 26 ($913,016)
DAL Alexis Ajinca 22 ($1,019,134)
DAL Ian Mahinmi 24 ($1,539,540)
DAL Caron Butler 30 ($1,639,930)
DAL Jason Terry 33 ($1,828,424)
DAL DeShawn Stevenson 29 ($2,813,191)
DAL Dominique Jones 22 ($7,704,900)
DAL Shawn Marion 32 ($12,460,313)
DAL Brendan Haywood 31 ($37,715,177)
($16,555,122)
SAS DeJuan Blair 21 $41,323,371
SAS George Hill 24 $32,781,522
SAS Manu Ginobili 33 $25,887,887
SAS Matt Bonner 30 $5,633,457
SAS Tony Parker 28 $3,263,252
SAS Tim Duncan 34 $2,033,743
SAS Ime Udoka 33 ($352,709)
SAS Chris Quinn 27 ($13,886)
SAS Gary Neal 26 ($195,725)
SAS Bobby Simmons 30 ($454,348)
SAS Tiago Splitter 26 ($596,621)
SAS Garrett Temple 24 ($734,359)
SAS Alonzo Gee 23 ($1,538,049)
SAS James Anderson 21 ($3,514,018)
SAS Antonio McDyess 36 ($4,439,251)
SAS Richard Jefferson 30 ($14,372,975)
$84,711,289 |
_________________ GodismyJudgeOK.com/DStats
Twitter.com/DSMok1 |
|
Back to top |
|
|
Ilardi
Joined: 15 May 2008 Posts: 265 Location: Lawrence, KS
|
Posted: Thu Dec 16, 2010 2:16 pm Post subject: |
|
|
Very helpful, thanks.
Next question: how did you determine "replacement level"? |
|
Back to top |
|
|
DSMok1
Joined: 05 Aug 2009 Posts: 611 Location: Where the wind comes sweeping down the plains
|
Posted: Thu Dec 16, 2010 2:35 pm Post subject: |
|
|
Ilardi wrote: | Very helpful, thanks.
Next question: how did you determine "replacement level"? |
That's tough. Baseball is the primary place to learn about replacement level...
I looked at the distribution of talent in the NBA. I looked at players that are available for cheap/min contract. I considered how bad a team of replacement-level players should be.
And then I just picked the number. -3 could be a better number. Looking at my true-talent projections for this year, there are a significant number of players better than -3 that are not on rosters (most retired, others are out because NBA teams would rather have a below-replacement level rookie with potential upside.) Players like
Josh Boone
Kyle Weaver
Sergio Rodriguez
Larry Hughes
Rafer Alston
James Singleton
Ronald Murray
Rasho Nesterovic
Rashad McCants
Malik Hairston
I have above -3 estimated.
A team of 5 -3.5 players would be expected to win how many games? -3.5*5 =-17.5 efficiency differential. That's 6.5 wins. -3*5 yields -15, or 9 wins. _________________ GodismyJudgeOK.com/DStats
Twitter.com/DSMok1 |
|
Back to top |
|
|
DSMok1
Joined: 05 Aug 2009 Posts: 611 Location: Where the wind comes sweeping down the plains
|
Posted: Thu Dec 16, 2010 3:30 pm Post subject: |
|
|
Spreadsheet available HERE:
Advanced SPM Salary Analysis
This spreadsheet has everything related to ASPM outside of the preseason projection system, including some other things I've been working on. For salaries and analysis, go to the Value & Contracts tab.
EDIT: Link Updated _________________ GodismyJudgeOK.com/DStats
Twitter.com/DSMok1
Last edited by DSMok1 on Thu Dec 16, 2010 4:01 pm; edited 1 time in total |
|
Back to top |
|
|
Crow
Joined: 20 Jan 2009 Posts: 826
|
Posted: Thu Dec 16, 2010 3:41 pm Post subject: |
|
|
Average contract length about 2.4 to 2.5 seasons? |
|
Back to top |
|
|
DSMok1
Joined: 05 Aug 2009 Posts: 611 Location: Where the wind comes sweeping down the plains
|
Posted: Thu Dec 16, 2010 4:10 pm Post subject: |
|
|
Here are the teams, broken down by year. Be aware--a few contracts are still missing like Beaubois from Dallas.
This gives a good idea of when teams have the biggest competitive advantage. Look at it this way: to get the same quality of team, the given team will have to spend this much more (if they're negative) or less (if they're positive).
Code: | Team Total Value 2011 2012 2013 2014 2015 2016
MIA $171,568,781 $30,604,806 $39,040,369 $30,943,343 $26,647,204 $21,361,305 $22,971,754
OKC $159,082,096 $16,114,033 $46,666,176 $47,759,447 $28,122,628 $8,919,631 $11,500,181
SAS $85,660,619 $31,446,965 $31,591,798 $31,849,322 -$8,433,460 -$794,006 $0
CHI $81,588,271 $22,206,524 $25,217,717 $21,671,041 $3,501,074 $4,975,119 $4,016,797
POR $72,076,223 $23,318,833 $30,323,507 $20,799,645 $816,710 -$3,182,471 $0
BOS $68,583,329 $19,663,798 $11,382,551 $12,495,381 $11,156,770 $13,884,829 $0
SAC $48,818,312 -$6,451,989 $9,973,235 $14,863,163 $21,891,627 $8,542,276 $0
UTA $43,774,277 $6,824,637 $16,974,415 $13,736,605 $2,610,757 $3,627,863 $0
MIL $40,929,336 $16,738,576 $25,502,082 $2,608,260 $9,126,207 -$13,045,789 $0
PHI $38,031,895 -$1,956,184 $18,398,488 $4,961,406 $16,993,609 -$365,424 $0
IND $32,429,646 -$2,457,063 $7,099,599 $12,778,103 $8,260,413 $6,748,593 $0
LAC $26,349,453 -$6,892,679 -$2,745,610 $13,198,458 $17,207,909 $5,581,375 $0
NYK $23,413,377 $11,589,508 $22,310,454 $5,152,669 -$7,598,946 -$8,040,309 $0
GSW $25,646,201 -$4,749,108 $11,416,179 $19,304,945 $15,919,642 -$8,385,437 -$7,860,021
MEM $20,667,007 $5,823,957 $15,889,949 -$1,931,581 -$1,939,069 $2,823,751 $0
NOH $17,180,724 $9,552,400 $15,110,112 $7,373,786 -$12,201,706 -$2,653,866 $0
DEN $15,995,233 $14,140,781 $9,690,229 $1,182,441 -$2,833,712 -$6,184,505 $0
HOU $15,865,222 $10,400,914 $8,841,312 $3,313,749 -$1,500,391 -$5,190,362 $0
NJN $15,488,568 -$3,703,210 $3,971,621 $12,129,825 $110,606 $2,979,726 $0
ATL $19,602,337 $6,866,576 $10,986,451 $12,144,520 $1,867,965 -$5,599,769 -$6,663,407
ORL $13,027,084 $7,638,506 $6,228,263 $5,712,735 -$6,552,420 $0 $0
TOR -$18,584,261 -$20,489,794 -$6,613,322 -$472,781 $5,456,539 $3,535,097 $0
LAL -$1,534,494 $8,816,002 $7,547,086 -$5,068,169 -$12,829,413 $0 $0
CLE -$15,913,925 -$5,033,687 -$11,519,488 $256,083 $1,668,571 -$1,285,404 $0
CHA -$18,694,473 -$9,832,559 -$9,938,468 -$6,105,041 $2,488,067 $4,693,528 $0
WAS -$18,780,112 -$12,462,718 -$17,193,217 $4,895,497 -$2,190,533 $8,170,858 $0
PHO -$27,530,843 $6,073,826 $2,024,467 -$13,536,156 -$16,335,602 -$5,757,379 $0
MIN -$32,690,510 -$6,110,183 -$2,526,238 -$2,844,152 -$15,656,538 -$5,553,398 $0
DAL -$50,799,294 -$1,748,424 -$2,917,609 -$10,754,939 -$16,473,021 -$10,648,751 -$8,256,550
DET -$60,783,166 -$14,361,246 -$14,150,093 -$30,160,326 -$11,242,822 $9,131,321 $0 |
The biggest advantage any 1 team has (and it's not close) is OKC in 2012 and 2013. _________________ GodismyJudgeOK.com/DStats
Twitter.com/DSMok1 |
|
Back to top |
|
|
acollard
Joined: 22 Sep 2010 Posts: 56 Location: MA
|
Posted: Fri Dec 17, 2010 5:23 pm Post subject: |
|
|
DSMok1 wrote: | One thing I didn't account for (I'm not an accountant) is the uncertainty level of the forecast. This is the mean expectation; some players will have greater variability than others.
|
I'm not an accountant, but here are some things I remember from financial theory:
Because people are risk averse, you have to pay a premium for larger uncertainty in returns, holding all else equal. There are a lot of ways to quantify this risk premium, but the Sharpe Ratio is probably the easiest and most well known. It seems like a good place to start.However, this isn't a perfect market, and its not as if GMs actually have an estimated expected value or standard deviation in front of them. The Sharpe Ratio assumes a normal distribution of expected outcomes, and I'm not really sure if this is the case for player performance.
I could imagine a setup going like this, though.
1) Verifying that NBA teams are risk averse. Compare samples of players with similar expected values for a time period, are the players with a higher standard devation less valued?
2) Create a modified Sharpe Ratio for ASPM. Because you don't have a better way to quantify it: using the ASPM prediction model with only the data before a contract is signed, find the expected value and the standard deviation for a given player. Calculate the ratio for every contract over a period of years and average it. You could arbitarily set the risk-free rate, I don't see how else you would determine it.
3) Use the average Sharpe Ratio determined in 2 for the valuation model to lower the value of players with high risk.
You would also want to disclude any imposed market inefficiencies, like max contracts, rookie-scale contracts, or minimum contracts, because these seem would be unable to properly value the total expected value and risk.
This method assumes a lot of things, I'm not sure it would even work and it would require a lot of time/ effort. But I think its interesting to think about.
Another thing to think about: using values below replacement level for a player might not make sense, and weighting those at all in an average of possible expectations might undervalue the player. Values below replacement aren't relevant, because at that level a player just won't see the court. For example, if player A has an expected ASPM of -3 with a sigma of 1, he is a much better bet than player B with the same expected ASPM with a sigma close to 0. It means there is a chance the first player will be productive, and if not, he's at replacement level, but there is no chance the second player will be better than replacement. |
|
Back to top |
|
|
greyberger
Joined: 27 Sep 2010 Posts: 54
|
Posted: Sat Dec 18, 2010 1:26 pm Post subject: |
|
|
Thanks for posting this, DSM1. |
|
Back to top |
|
|
DSMok1
Joined: 05 Aug 2009 Posts: 611 Location: Where the wind comes sweeping down the plains
|
|
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
|