Replacement player value (Mike G, 2010)
Posted: Mon Apr 18, 2011 7:26 pm
Author Message
Mike G
Joined: 14 Jan 2005
Posts: 3612
Location: Hendersonville, NC
PostPosted: Thu May 20, 2010 5:05 pm Post subject: Replacement player value Reply with quote
If .80 of average efficiency differential (ED) is Replacement Level (RL) play, then a team of RL players, averaging a score of 80-100, would be expected to win about 5% of their games. They might go 4-78.
A team that averages 70-100 should only win 1% of their games. That's pretty close to 0-82.
To win 10 of 82 games, RL would be around .86 or .87 ED.
No team has ever been as bad as 4-78, and no team has ever been without any good player for a whole season. Of course, a team might have some players worse than RL and still average RL.
Here are 2010 league average per-36 rates, and .80 of those rates:
Code:
per36 Pts Reb Ast Stl Blk PF TO
avg 15.0 6.2 3.2 1.1 0.7 3.1 2.0
80% 12.0 5.0 2.5 0.9 0.6 2.5 1.6
No reason to suppose PF and TO should be lower than avg for an RP, so we do something else with those.
To the eye, though, the positives of line 1 look like a weak starter or decent bench guy. The 80% line is decidedly weak.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
BobboFitos
Joined: 21 Feb 2009
Posts: 201
Location: Cambridge, MA
PostPosted: Fri May 21, 2010 11:27 am Post subject: Reply with quote
Good stuff. The question though, focusing on scoring per 36, is are we assuming that .8 player is taking on the same offensive burden? Also, .8 TO of 2 per 36 should be 2.5 since it's the inverse of .8, no? If the shot attempts are the same, then their TS is falling from say an average value of ~54.3 to generate 15 per 36 (or 13.81 raw true shooting attempts) down to ~43.4 TS.*
*I basically figured to score 15 per 36 on the average NBA TS requires 13.81 TSA, so if you're scoring only 12 points on those same shot attempts, that's the relative drop in TS.
~~~
Do people view the worst historical record as a team full of replacement players? I don't. The 9 win 1973 76ers actually fell well below their pyth, but employed several players who had long-ish NBA careers. (Suggesting that they were not truly replaceable in the minimal sense) Granted I don't know anything about these players outside of skimming bref/looking at their career stats, but I think setting the door at a "10 win team" isn't really the hypothetical floor.
Does a team full of replacements - that is, a team that doesn't have ANYONE on the roster that would be deemed an improvement for any other team (maybe?) even win a game? There's a huge gap between a -20 and -30 PD.
_________________
http://pointsperpossession.com/
@PPPBasketball
Back to top
View user's profile Send private message
DSMok1
Joined: 05 Aug 2009
Posts: 611
Location: Where the wind comes sweeping down the plains
PostPosted: Fri May 21, 2010 12:04 pm Post subject: Reply with quote
A good start to studying replacement level would be to study baseball's accepted methodology. Hardball times has a few articles, for starters.
Basketball skill over the whole world's population is approximately a bell curve. We can all agree on that, right?
The players in the NBA are basically the right tail of that curve--so the curve should look fairly exponential, in terms of player quality from left to right.
However, once you get down to below average, there is a strong selective sampling issue--not every player of that quality actually is playing in the NBA. We end up seeing a slightly left-skewed graph of the quality of players in the NBA. Something like a log-normal distribution of sigma 1/2. Once the players get bad enough, none of them are in the NBA.
From a team's perspective, they are trying to get enough quality players on the court at once to win games. Is an NBA average player worth something? Yes, by definition. NBA-average players would win 41 games. What about somebody a little worse than average? Yes, he would help out also. You may need to play him, and players of his caliber (say an APM of -2) are not freely available.
So what is the threshold of replacement level? There is the rub. It is not the best player who isn't starting. That's way too high. It's probably above the 15th player on the bench--he could be old and bad but still under contract, or a youngster expected to develop.
Here's Tom Tango's definition:
Quote:
WAR is wins above replacement. Replacement is defined very specifically for my purposes: it’s the talent level for which you would pay the minimum salary on the open market, or for which you can obtain at minimal cost in a trade.
So we're not defining by how bad a team could be, but by what is available.
Which is rather hard to do, honestly.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
mtamada
Joined: 28 Jan 2005
Posts: 377
PostPosted: Fri May 21, 2010 4:39 pm Post subject: Reply with quote
DSMok1 wrote:
A good start to studying replacement level would be to study baseball's accepted methodology.
Yes, we do not know how many wins a replacement level player (or team of such players) would win, nor do we know what their stats would be.
But we do know the approximate description of a replacement level player: the freely available, marginal-on-or-off-the-roster player.
So I'd approach the estimation from three directions. First look at the stats of the 12th man (or these days, the 13th through 15th men) on NBA rosters. Defining the 12th man is a little tricky ... I'd basically use minutes played, but would make sure to exclude a Gred Oden-like player who may've played 12th man minutes due to injury not due to lack of ability.
That gives you an upper bound on how good the replacement level player is ... there might be some players who are below replacement level but manage to stick to a roster due to either a guaranteed contract or local popularity or for development or team sentimentality (Ken Griffey Jr. is like that right now for the Mariners) but on the whole, these are players who by definition made it onto an NBA roster, and are likely to be on average a bit better than the typical replacement player.
Next, look at the players who got called up to the NBA during the season. Again make sure to exclude good players who started the season injured or who were acquired as expensive free agents; include only minimum salary free agents and call-ups from the NBDL (again excluding players such as Hasheem Thabeet who may've been sent down for developmental reasons or whose ability level may've literally fluctuated throughout the season). This comes the closest I think to measuring replacement level players, but the small number of games and minutes that these players play may imply large standard errors in our estimates.
Finally, look literally at the minimum salary players. Some of these players will in fact be pretty good players who got belatedly discovered by some NBA team and who are truly better than replacement level, but are unlucky enough to be getting only a minimum salary contract. So this group again will tend to give you a bit of an overestimate of the replacement level.
None of these measures are perfect but between the three of them you could start really zeroing in on what the replacement level stats look like.
You'd probably want to do different stats for positions; a replacement level PG is going to have very different stats from a replacement level C.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3612
Location: Hendersonville, NC
PostPosted: Sat May 22, 2010 6:55 pm Post subject: Reply with quote
Are Win Shares (or WS/48) already above a replacement level?
This season, there were 368 players w >200 minutes and >0 WS/48. (Along with 18 who were negative.)
That's about 12.3 per team.
At 600 minutes, 310 positive and just 8 negative.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
DSMok1
Joined: 05 Aug 2009
Posts: 611
Location: Where the wind comes sweeping down the plains
PostPosted: Sat May 22, 2010 9:16 pm Post subject: Reply with quote
Mike G wrote:
Are Win Shares (or WS/4Cool already above a replacement level?
This season, there were 368 players w >200 minutes and >0 WS/48. (Along with 18 who were negative.)
That's about 12.3 per team.
At 600 minutes, 310 positive and just 8 negative.
I think not, because the total of all win shares in the league equals 41*30.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3612
Location: Hendersonville, NC
PostPosted: Sun May 23, 2010 6:14 am Post subject: Reply with quote
But if by definition only players above replacement level are adding wins, then only negative WS/48 players would be considered sub-replacement level.
If so, then .000 WS/48 = RL
Whether 12 per team is the 'right' number of >RL players, or whether the remaining 3 players x 30 rosters ('project' players, practice guys, etc) effectively ensures that there are no better players available, this might be close.
It's still possible for replacements to generate wins, as their distribution of performance varies from game to game. A whole team of RP's (and sub-RP's) can get it together against a merely bad team on a bad night, and pull off a win or 2, out of 82.
In the real world, though, the RP team is in a vacuum against relentless pressure. A Stackhouse or a Rodman will surely come out of exile or retirement and suit up.
Found this thread which discusses replacement level:
http://sonicscentral.com/apbrmetrics/vi ... php?t=2153
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
DSMok1
Joined: 05 Aug 2009
Posts: 611
Location: Where the wind comes sweeping down the plains
PostPosted: Sun May 23, 2010 8:24 am Post subject: Reply with quote
Mike G wrote:
But if by definition only players above replacement level are adding wins, then only negative WS/48 players would be considered sub-replacement level.
If so, then .000 WS/48 = RL
No.
The replacement level is set by what is available, not by how many wins are contributed. In baseball, we:
Quote:
figure out the offensive production that a team could expect from players not projected to be good enough to make a major league roster next year. These guys have fallen into that Four-A category, where they show more ability than your average Triple-A veteran but not enough to hold down a major league job. They’re usually available every winter as minor league free agents, via the Rule 5 draft, or as cheap trade acquisitions where a team can acquire one of these players without giving up any real talent in return.
As Sean showed in his article, and has been shown elsewhere, the expected value of a replacement level player is about negative 20 runs per 600 PA. Or, to phrase it a bit differently, if you lost a league average player and replaced him with a freely available guy, you’d lose about two wins.
(source)
In baseball, a "replacement level" team is still at about a .400 winning percentage--the scatter isn't as big.
In basketball, we need to figure out how good the players are that are "freely available" and consider that replacement level.
That could be done by looking at the distribution of players in terms of their APM, and see where the distribution transitions from "the tail of a bell curve" to more of a log-normal distribution--which would imply players of that quality are available.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
gabefarkas
Joined: 31 Dec 2004
Posts: 1313
Location: Durham, NC
PostPosted: Sun May 23, 2010 12:32 pm Post subject: Reply with quote
Instead of trying to theoretically come up with how many wins produced a replacement player would provide, why not try to come up with it empirically? What I mean is to try to identify actual replacement players, and then look at what they actually did on the court.
If we wanted to, from there we could even do permutation distributions, bootstrapping, jackknife, or any other resampling technique to come up with an empirical distribution for what would be expected from a replacement player.
So, what if we came up with a list of who we believe are actual replacement players? A starting point might be something along the lines of:
Aaron Gray
Aaron Miles
Alex Acker
Allan Ray
Andre Emmett
Awvee Storey
Bernard Robinson
Britton Johnsen
Casey Jacobsen
Chase Budinger
Chris Jefferies
Chris Quinn
Chris Taft
Daniel Ewing
Deng Gai
Dijon Thompson
Frank Williams
Gabe Pruitt
Josh McRoberts
etc etc etc
Does that make sense? Come up with a list of what you would consider replacement players, and then derive an empirical distribution from there.
Back to top
View user's profile Send private message Send e-mail AIM Address
Carlos
Joined: 21 Jan 2005
Posts: 64
Location: Montevideo, Uruguay
PostPosted: Sun May 23, 2010 12:38 pm Post subject: Reply with quote
What about looking at guys picked from the NBDL and undrafted players or midseason additions? It's relatively rare, but each year you see a few guys enter the league who were truly obtained at minimum costs. I remember the Grizziles a few years ago had all of their PGs injured and picked Elliot Perry basically off the street to play for a couple of games. It was unusual, sure, but I'd bet that every year there are a few cases like that, of teams playing truly marginal guys because of health issues.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3612
Location: Hendersonville, NC
PostPosted: Sun May 23, 2010 12:46 pm Post subject: Reply with quote
DSMok1 wrote:
The replacement level is set by what is available, not by how many wins are contributed. ..
And are these mutually exclusive?
If teams are actually utilizing players who are worse than available players, they aren't optimizing their wins.
You might even say they're putting guys on the floor who are removing wins, relative to the default position of 'best available replacements'.
Quote:
In baseball, a "replacement level" team is still at about a .400 winning percentage ...
Winning 40% means winning 80% of average.
This really is a far cry from 80% of average efficiency differential -- which in NBA translates to about 5% wins. This is an artifact of 48-minute games (actually, of 90-95 possession games). In a one-possession game, your W/L = your ED.
How do you explain baseball teams that win <40% ? They're trying to lose?
If baseball statisticians have made serious logical errors, do we wish to perpetuate them?
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
DLew
Joined: 13 Nov 2006
Posts: 224
PostPosted: Sun May 23, 2010 1:45 pm Post subject: Reply with quote
Mike,
You always seem hung up by anything that suggests teams are not optimizing. Why is this? It seems clear that many teams are not. Some teams play guys who are worse than the freely available players. Do you disagree?
Back to top
View user's profile Send private message
schtevie
Joined: 18 Apr 2005
Posts: 413
PostPosted: Sun May 23, 2010 2:34 pm Post subject: Reply with quote
Can someone say, slowly and clearly, why one should be impressed with the concept of a replacement level player? In particular, why is it useful at all in discussing basketball?
I kind of can see in baseball, where the production function of the game is relatively simple, defining some inherently arbitrary below-average standard as a benchmark for comparing player values and this becoming a convention over time. But basketball is different. The production function is very complicated; it is not clear what the below-average standard should be; and it is difficult to measure player contributions precisely.
It is far more intuitive to at least eliminate one of these problems and have the benchmark of comparison be the average player, whether this value be zero (in terms of net points) or 8.2 (for per game WS).
The futility can be illustrated with 2007-08 data, chosen to utilize Ilardi & Barizlai's multi-year APM numbers as estimates of value. Let's take as an arbitrary starting point, the 360th ranked player in minutes per game (so, kind of the 12th man on the 30th team): Malik Rose, playing 10.1 minutes per game. Finding the 15 players above and below this mark (again, arbitrary) that meet the minimum minutes cut-off for I & B's estimates, the average multi-year APM is -3.0.
Might this replacement level definition be consistent over time? Perhaps. Ten minutes seems plausible. But even if the contribution of ten minute players hurt their teams to the tune of 3 points per 100 possessions on average, there is a great deal of dispersion around the mean. The standard deviation in the instance is 4.1, the estimates ranging from Ronnie Price's 7.17 to Dominic McGuire's -12.07. As such, what is the point of the exercise? What does the concept of a replacement level player gain us?
Back to top
View user's profile Send private message
Kevin Pelton
Site Admin
Joined: 30 Dec 2004
Posts: 979
Location: Seattle
PostPosted: Sun May 23, 2010 3:07 pm Post subject: Reply with quote
http://web.archive.org/web/200404040408 ... 7557.shtml
Back to top
View user's profile Send private message Send e-mail Visit poster's website
DSMok1
Joined: 05 Aug 2009
Posts: 611
Location: Where the wind comes sweeping down the plains
PostPosted: Sun May 23, 2010 4:56 pm Post subject: Reply with quote
Kevin Pelton wrote:
http://web.archive.org/web/200404040408 ... 7557.shtml
There we go. I knew you had that around somewhere, but I couldn't find it.
That article is the baseline for A) Why replacement l level is important and B) a first effort at how to arrive at that level.
schtevie wrote:
Can someone say, slowly and clearly, why one should be impressed with the concept of a replacement level player? In particular, why is it useful at all in discussing basketball?
Replacement level is required to determine the actual value of a player to the team. Since the players are all pulled from the tail end of a bell curve, at a certain level, there are lots of players available--and thus basically freely available. The value of any player in the NBA is actually their marginal value above the players that are freely available.
schtevie wrote:
I kind of can see in baseball, where the production function of the game is relatively simple, defining some inherently arbitrary below-average standard as a benchmark for comparing player values and this becoming a convention over time. But basketball is different. The production function is very complicated; it is not clear what the below-average standard should be; and it is difficult to measure player contributions precisely.
It is far more intuitive to at least eliminate one of these problems and have the benchmark of comparison be the average player, whether this value be zero (in terms of net points) or 8.2 (for per game WS).
The futility can be illustrated with 2007-08 data, chosen to utilize Ilardi & Barizlai's multi-year APM numbers as estimates of value. Let's take as an arbitrary starting point, the 360th ranked player in minutes per game (so, kind of the 12th man on the 30th team): Malik Rose, playing 10.1 minutes per game. Finding the 15 players above and below this mark (again, arbitrary) that meet the minimum minutes cut-off for I & B's estimates, the average multi-year APM is -3.0.
Might this replacement level definition be consistent over time? Perhaps. Ten minutes seems plausible. But even if the contribution of ten minute players hurt their teams to the tune of 3 points per 100 possessions on average, there is a great deal of dispersion around the mean. The standard deviation in the instance is 4.1, the estimates ranging from Ronnie Price's 7.17 to Dominic McGuire's -12.07. As such, what is the point of the exercise? What does the concept of a replacement level player gain us?
One issue of looking at it purely in a MPG frame is that there is a ton of error associated in looking at players of that level. Another is that a player that is playing 10 MPG on the Lakers is a lot better than a player that is getting 10 MPG on the Nets.
I, however, also came to the same number (-3 pts/100 pos.) when evaluating players using Statistical +/-. It appears that a player of that caliber is nearly freely available. Interestingly, translating that -3 number into a team efficiency differential and running the Pythagorean formula yields the 10 win number yet again.
Author Message
Kevin Pelton
Site Admin
Joined: 30 Dec 2004
Posts: 979
Location: Seattle
PostPosted: Sun May 23, 2010 5:22 pm Post subject: Reply with quote
DSMok1 wrote:
There we go. I knew you had that around somewhere, but I couldn't find it.
Unfortunately, the Hoopsworld archive is nonexistent, so that stuff is tough to find unless you know exactly where you're looking.
The big problem with using that ancient linear-weights stat is there was no defined "average," so I wasn't able to look at replacement level as a percentage of average.
Were I to do it again, I'd want to do some sort of cross-metric comparison like AaronB did with the draft a few years back: http://www.82games.com/barzilai1.htm
That way, you ensure biases in the specific metric aren't affecting your conclusion.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
schtevie
Joined: 18 Apr 2005
Posts: 411
PostPosted: Sun May 23, 2010 5:40 pm Post subject: Reply with quote
DSMok1 wrote:
Kevin Pelton wrote:
http://web.archive.org/web/200404040408 ... 7557.shtml
There we go. I knew you had that around somewhere, but I couldn't find it.
That article is the baseline for A) Why replacement l level is important and B) a first effort at how to arrive at that level.
Having (re)read the article, I remain at my initial conclusion. The best I can tell is that defining replacement level is important because some guys in baseball do it. And the institutional/economic conditions for using it there do not obtain in basketball
DSMok1 wrote:
schtevie wrote:
Can someone say, slowly and clearly, why one should be impressed with the concept of a replacement level player? In particular, why is it useful at all in discussing basketball?
Replacement level is required to determine the actual value of a player to the team. Since the players are all pulled from the tail end of a bell curve, at a certain level, there are lots of players available--and thus basically freely available. The value of any player in the NBA is actually their marginal value above the players that are freely available.
Let's be clear. What is written in bold is at best misleading but more basically simply untrue. First let's define value. Never mind economic value, because that adds an unnecessary layer of complication. Let's stick with competitive value, which ultimately is a function of a player's influence on the scoreboard on a per possession basis. Such value, to a first approximation is completely independent of the value of players that are freely available. One can try to come up with a meaningful (as in accurate and precise) estimate of the marginal value of this representative replacement player (which is the entire exercise in futility). But stipulating success in that exercise, what does one have or gain? Nothing but an arbitrary benchmark. So LBJ is valued at 19 points per 100 possessions (or whatever) above the replacement player, instead of 16 points better than average. And?
DSMok1 wrote:
schtevie wrote:
I kind of can see in baseball, where the production function of the game is relatively simple, defining some inherently arbitrary below-average standard as a benchmark for comparing player values and this becoming a convention over time. But basketball is different. The production function is very complicated; it is not clear what the below-average standard should be; and it is difficult to measure player contributions precisely.
It is far more intuitive to at least eliminate one of these problems and have the benchmark of comparison be the average player, whether this value be zero (in terms of net points) or 8.2 (for per game WS).
The futility can be illustrated with 2007-08 data, chosen to utilize Ilardi & Barizlai's multi-year APM numbers as estimates of value. Let's take as an arbitrary starting point, the 360th ranked player in minutes per game (so, kind of the 12th man on the 30th team): Malik Rose, playing 10.1 minutes per game. Finding the 15 players above and below this mark (again, arbitrary) that meet the minimum minutes cut-off for I & B's estimates, the average multi-year APM is -3.0.
Might this replacement level definition be consistent over time? Perhaps. Ten minutes seems plausible. But even if the contribution of ten minute players hurt their teams to the tune of 3 points per 100 possessions on average, there is a great deal of dispersion around the mean. The standard deviation in the instance is 4.1, the estimates ranging from Ronnie Price's 7.17 to Dominic McGuire's -12.07. As such, what is the point of the exercise? What does the concept of a replacement level player gain us?
One issue of looking at it purely in a MPG frame is that there is a ton of error associated in looking at players of that level. Another is that a player that is playing 10 MPG on the Lakers is a lot better than a player that is getting 10 MPG on the Nets.
I, however, also came to the same number (-3 pts/100 pos.) when evaluating players using Statistical +/-. It appears that a player of that caliber is nearly freely available. Interestingly, translating that -3 number into a team efficiency differential and running the Pythagorean formula yields the 10 win number yet again.
Here again, I must disagree. First, I wasn't really advocating for an MPG frame, or any other frame. MPG has particular shortcomings I am sure. Primarily I am guessing in terms of biasing the "true" replacement value upward. (That is, within the low minute cohort, that there are up and coming players, earning their minutes who are actually more valuable than "true" replacement players.) My point in picking an arbitrary standard that seemed basically fine was to illustrate the futility of the exercise.
And second, to quibble, I don't think there is necessarily any reason to believe, a priori, that a Lakers 10 MPG player would be better than a Nets 10 MPG player. The Lakers are clearly a better team, but this doesn't necessarily mean that their 12th man is better.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3597
Location: Hendersonville, NC
PostPosted: Mon May 24, 2010 8:18 am Post subject: Reply with quote
Schtevie, I believe the thought process is this:
NBA teams are forced to use below-average players for many minutes per season. While we all agree certain guys are below average, the debate is about whether they're actually hurting their teams' ability to win.
If teams are using players who are below the value of available replacements, then it's arguable that the use of these players actually does hurt them. If they're the best available, and you have to play someone, then they're as good as it gets.
In another thread, DSMok1 chose to evaluate the wins contributed by players at various draft positions, using [WS/48 - .025] . In other words, wins above replacement, defined as 25% of average win-production rate.
Since we find players such as Monta Ellis getting 41 mpg, at a rate of .023 WS/48, it's a wonder that legendary coach Don Nelson would be inclined to 'hurt' his team by playing a guy who adds nothing to the team's win probability, and who could be easily replaced.
If we have an educated estimate of replacement value -- perhaps .50 , .20 , or -.10 of average -- it helps us form statements about a player's real 'value'.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
DLew
Joined: 13 Nov 2006
Posts: 224
PostPosted: Mon May 24, 2010 9:56 am Post subject: Reply with quote
Mike,
I don't think anyone would argue Monta Ellis adds nothing to his team's win probability in absolute terms. Clearly having him out there is better than having only 4 players on the court. I agree only with the second half of your statement, that (according to Win Shares) he adds nothing above what a freely available replacement player (assuming the 0.025 threshold) would add. Interestingly enough, the Warriors employed several freely available replacement players this season (Anthony Tolliver, Coby Karl, Reggie Williams). Tolliver and Williams rated somewhat higher than Ellis, Karl not so much.
Back to top
View user's profile Send private message
DSMok1
Joined: 05 Aug 2009
Posts: 610
Location: Where the wind comes sweeping down the plains
PostPosted: Mon May 24, 2010 10:04 am Post subject: Reply with quote
Here's what a graph of Win Shares over Replacement vs. MPG vs. Team Efficiency Differential looks like, using the 0.025 level, historically:
As you can see, a league average team basically doesn't play anybody that's below replacement level. However, as soon as you get below league average, a few players of that type start to play. Once you get down into BAD territory, then a lot of the replacement-level players start to appear.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3597
Location: Hendersonville, NC
PostPosted: Mon May 24, 2010 10:53 am Post subject: Reply with quote
DLew wrote:
I don't think anyone would argue Monta Ellis adds nothing to his team's win probability in absolute terms. ..
And yet, this thread was prompted by another in which DSMok1 stated:
Quote:
They were at or below replacement level, so they contributed nothing to the team, other than taking up space.
(which was in reply to this list --
http://www.basketball-reference.com/pla ... i?id=NG7tG
-- of players with 4-yr WS/48 <.030)
Monta Ellis was .023 WS/48 this year.
It seems to me that the draft pick value study is potentially a monumental work, but it's based on a couple of questionable assumptions:
- that WS/48 is a true measure of player value
- that .025 WS/48 = replaceable value.
Most of the time, when you have such large samples (rookies through 4 years), the idiosyncrasies of a metric will average out. But, given that the better rookies tend to be drafted by the weaker teams, such a study could be skewed by the 2 assumptions above.
Indeed, by PER, Ellis is well above NBA average (16.7) -- RWilliams (16.0), Tolliver (13.7), Karl (7.1) not so much.
By my own metric, Ellis was a bit over par, at .106 eW/48; the others .88, .79, and .13 respectively.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
DSMok1
Joined: 05 Aug 2009
Posts: 610
Location: Where the wind comes sweeping down the plains
PostPosted: Mon May 24, 2010 2:24 pm Post subject: Reply with quote
Mike G wrote:
DLew wrote:
I don't think anyone would argue Monta Ellis adds nothing to his team's win probability in absolute terms. ..
And yet, this thread was prompted by another in which DSMok1 stated:
Quote:
They were at or below replacement level, so they contributed nothing to the team, other than taking up space.
(which was in reply to this list --
http://www.basketball-reference.com/pla ... i?id=NG7tG
-- of players with 4-yr WS/48 <.030)
Monta Ellis was .023 WS/48 this year.
It seems to me that the draft pick value study is potentially a monumental work, but it's based on a couple of questionable assumptions:
- that WS/48 is a true measure of player value
- that .025 WS/48 = replaceable value.
Most of the time, when you have such large samples (rookies through 4 years), the idiosyncrasies of a metric will average out. But, given that the better rookies tend to be drafted by the weaker teams, such a study could be skewed by the 2 assumptions above.
Indeed, by PER, Ellis is well above NBA average (16.7) -- RWilliams (16.0), Tolliver (13.7), Karl (7.1) not so much.
By my own metric, Ellis was a bit over par, at .106 eW/48; the others .88, .79, and .13 respectively.
I think "absolute" and "marginal" are being confused. Sure, Monta Ellis adds something to the team. He's better than air; he's better than me. However, he is not necessarily better than a "replacement level" player.
In fact, using wins as an absolute measure is one of the problems. Wins are nonlinear in this regard. I would use points/possession as the measure, since that is linear. I could say that I am "Absolutely" -25 pts/100poss, and Monta Ellis is -3 pts/100poss. How would we measure wins, though?
In fact, there is no "floor" for NBA players in terms of value added. There is no easy way to quantify absolute value, accept in terms of a rate--not like baseball, where I would simply never get a hit. In the NBA, it is all part of 1 large continuum.
Me < Monta < Russel Westbrook < Lebron. That we can say. We can even quantify, in terms of points, a level for the players. Using RAPM, I would maybe be -25 pts, Ellis -5, Westbrook +1, Lebron +6 pts/100 poss. How much would each of us be worth to an NBA team? Or more precisely, to an average NBA team?
Well, I think we all could agree that I would be worth nothing. Truly 0. Somewhere, value starts to appear. Where? Where the level of talent starts to become scarce.
Suppose the whole population's basketball scoring value is basically one huge bell curve, centered at -44, with a standard deviation of 9.7, and there are 50,000,000 in the population. (That fits what we see decently well). Well, there are only 143 players above 0 pts/100. (The placement of 0 is actually arbitrary and dependent on the league, but the point stands). So the replacement level would be below that--there's a lot more than 143 players.
Illustrating the above paragraph
However, there are a few players that don't end up in the NBA that are below 0 pts/100 poss. In fact, quite a few that could have been filling in that part of the bell curve from -4 to 0. Whether they exist, I don't know--the NBA players are a subset that received specialized training AFTER demonstrating they were in this upper echelon. Sort of a selective sampling issue.
However, there are enough out there that at, say, -3 (to pick a number) a team can find such a player with a little effort but minimal expenditure. European players, NBDL, draft, free agents--players are available at that level for a minimum contract.
So, any player of that level is worth basically 0 to the team--they are all interchangeable. If one gets hurt, the team does not suffer at all--just pick somebody else up.
The value of a player to the team would then be quantified relative to the replacement level. I don't know that the relation is linear (in fact, I'm sure it's not--as opposed to baseball).
Back to top
View user's profile Send private message Send e-mail Visit poster's website
DLew
Joined: 13 Nov 2006
Posts: 224
PostPosted: Mon May 24, 2010 4:02 pm Post subject: Reply with quote
I should elaborate on the point I was trying to make: Replacement level is a useful construct for assessing player value because absolute zero is so difficult to assess. Value over replacement is much easier to determine than just 'value' because replacement level performance exists in the NBA, absolute zero level performance has never existed in the NBA so it's extremely difficult to calculate where it might fall.
Generally I think value over replacement is useful to explicitly calculate because every rating system is a value over something, it's just those that don't explicitly define zero level to some meaningful mark are much harder to interpret.
Back to top
View user's profile Send private message
schtevie
Joined: 18 Apr 2005
Posts: 411
PostPosted: Mon May 24, 2010 4:18 pm Post subject: Reply with quote
DSMok1, I was hoping for more persuasive arguments as to the utility of the concept of a replacement player in NBA basketball.
It surely is possible that there is a potential supply of such players numbering in the hundreds or thousands. But this is ultimately irrelevant. The demand for such players is tiny, and ever shall it remain so. And because the market is so thin, there is no real reason to reify this concept. To the contrary, it only creates confusion with no attendant benefits.
Who are replacement players? Basically, they are practice bodies and insurance policies. They play when there is injury, garbage time, and for a few minutes here and there for the benefit of team cohesion.
Why is it desirable to establish the "average" performance of such players? Surely not as a benchmark for those mostly outside this group. This isn't baseball, where because of the rules of the game, the worst players basically get as many at bats as the best. And, this aside, it isn't meaningful to speak of a single replacement level anyway. Eli W's clearly showed that each position has its own average.
To beat a horse that should die, basketball is not baseball. Basketball needs improved estimates of player values, not arbitrary referential standards for the worst players.
Back to top
View user's profile Send private message
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
PostPosted: Mon May 24, 2010 4:56 pm Post subject: Reply with quote
Replacement player value is useful because everything we do is about making a decision, usually a cost vs a benefit. Costs are usually straightforward - a player salary, though negotiation of that at times is the issue. Benefits are harder - the value of a win and the responsibility of a player for those. If you're going to pay a lot of money for a player, you should want to know their value, especially relative to the lowest price player out there. Do you buy generic if you think it's just as good as Frosted Flakes because it costs less? Or do you assume Frosted Flakes are better because they cost more? When our decisions value a lot more than Frosted Flakes, we want to know how good both the original and the generic are. In our case, there are a lot of essentially "generics" available every year. Assigning a general replacement value saves you from having to really value them all, which can be tricky with their limited minutes.
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers.
Back to top
View user's profile Send private message Visit poster's website
mtamada
Joined: 28 Jan 2005
Posts: 377
PostPosted: Mon May 24, 2010 6:44 pm Post subject: Reply with quote
Yup, replacement level is crucial in making roster decisions. Even if we ignore salaries, we still have to compare player qualities somehow, which means making tradeoffs between high performance, low minute players versus lower performance but higher minute players -- KevinP's 2004 comparison of Shaq vs McGrady is an excellent one. Ignoring for the moment salary and what the teams' needs are, i.e. team fit, would you trade one of those players for the other? Which was more valuable, McGrady's higher points, assists, etc. in 2003, or Shaq's higher per-minute productivity? The correct baseline to compare those players against is the replacement level player, not the average player (even the Lakers, Celtics, and Magic don't have a marginal player who is at league average level, rather their marginal player is way below that level).
People who ignore replacement level will overvalue the longer minutes but lower performing player. We see this all the time in Hall of Fame debates, where all too often the arguments are made based on the accumulation of stats such as points and rebounds, and even per game statistics, and too little attention is paid to the player's performance level. (The problem is even worse in baseball where some people will point at Pete Rose's 4,256 hits, and fail to take into account how many outs he made in accumulating those hits.)
Back to top
View user's profile Send private message
gabefarkas
Joined: 31 Dec 2004
Posts: 1313
Location: Durham, NC
PostPosted: Mon May 24, 2010 8:41 pm Post subject: Reply with quote
schtevie wrote:
This isn't baseball, where because of the rules of the game, the worst players basically get as many at bats as the best.
Umm, just as an aside, that's not actually true. Batting order dictates that some players will consistently get more plate appearances than others. Worse hitters are at the back of the lineup, just like worse shooters in basketball tend to take less FGA.
Back to top
View user's profile Send private message Send e-mail AIM Address
schtevie
Joined: 18 Apr 2005
Posts: 411
PostPosted: Mon May 24, 2010 9:57 pm Post subject: Reply with quote
HoopStudies wrote:
Replacement player value is useful because everything we do is about making a decision, usually a cost vs a benefit. Costs are usually straightforward - a player salary, though negotiation of that at times is the issue. Benefits are harder - the value of a win and the responsibility of a player for those. If you're going to pay a lot of money for a player, you should want to know their value, especially relative to the lowest price player out there. Do you buy generic if you think it's just as good as Frosted Flakes because it costs less? Or do you assume Frosted Flakes are better because they cost more? When our decisions value a lot more than Frosted Flakes, we want to know how good both the original and the generic are. In our case, there are a lot of essentially "generics" available every year. Assigning a general replacement value saves you from having to really value them all, which can be tricky with their limited minutes.
"Replacement player value" is not and cannot be a synonym for "relevant opportunity cost", except by mere coincidence.
First of all, Dean, I don't know why you are mixing on-court productivity with financial compensation here. Things, apparently, are confused enough already. And breakfast cereal analogies are not illuminating.
My point is that there is no relevant, informative replacement value. If one is considering replacing Zydrunas Ilgauskas, say, with Shaquille O'Neal. The analysis ultimately centers on how many net points one produces in comparison to the other, and the manner in which they are assumed to do so. One can use as a bench mark the notional productivity average of seven foot stiffs in the D-League, in Europe, and/or parts unknown (or if one is really going to have an aggregate replacement value benchmark, throw in the weighted contributions of point guards who can't shoot, or unathletic small forwards) but why? The comparison doesn't require this, and it is not only a completely uninformative detour on the face of it, but the replacement level estimate is necessarily imprecise.
If however one is looking for a new bench warmer to replace a bench warmer, fine, by all means compare him to a referential bench warmer. But at least compare to the relevant position and not some non-existent composite.
mtamada wrote:
Yup, replacement level is crucial in making roster decisions. Even if we ignore salaries, we still have to compare player qualities somehow, which means making tradeoffs between high performance, low minute players versus lower performance but higher minute players -- KevinP's 2004 comparison of Shaq vs McGrady is an excellent one. Ignoring for the moment salary and what the teams' needs are, i.e. team fit, would you trade one of those players for the other? Which was more valuable, McGrady's higher points, assists, etc. in 2003, or Shaq's higher per-minute productivity? The correct baseline to compare those players against is the replacement level player, not the average player (even the Lakers, Celtics, and Magic don't have a marginal player who is at league average level, rather their marginal player is way below that level).
People who ignore replacement level will overvalue the longer minutes but lower performing player. We see this all the time in Hall of Fame debates, where all too often the arguments are made based on the accumulation of stats such as points and rebounds, and even per game statistics, and too little attention is paid to the player's performance level. (The problem is even worse in baseball where some people will point at Pete Rose's 4,256 hits, and fail to take into account how many outs he made in accumulating those hits.)
Mike, I do not quite get your point or your disagreement with my position. I think we can agree that it is good to compare like to like. And that is precisely what I am advocating. In the form of not putting forth useless benchmarks. Every player should be analyzed in terms of what he can produce in terms of his own skill set in a hypothetical situation. The only bench mark that matters is the expected effect on the scoreboard (and/or resulting expected win totals). A replacement player - whatever he is - is only a useful point of comparison in the coincidental case that you are looking for a replacement player.
It is simply not the case that ignoring replacement level (whatever that is - and again, the problem in basketball is estimating value, which obtains to determining any notional benchmark. Assuming an accurate and precise benchmark - even if of dubious comparative value - isn't a solution.) overvalues or undervalues any type of player. Any such analysis must be coherent and stand on its own.
And just to emphasize the point, the motivation of this detour was the discussion of VORP in terms of the draft. Such an exercise perfectly illustrates the shortcomings. Every player in the draft, from the highly skilled lottery picks to the workhorses of the late second round are compared against a common imprecise composite. Again, to what gain?
Finally,
gabefarkas wrote:
schtevie wrote:
This isn't baseball, where because of the rules of the game, the worst players basically get as many at bats as the best.
Umm, just as an aside, that's not actually true. Batting order dictates that some players will consistently get more plate appearances than others. Worse hitters are at the back of the lineup, just like worse shooters in basketball tend to take less FGA.
Let me restate: basically, the worst players get as many at bats as the best.
Back to top
View user's profile Send private message
Neil Paine
Joined: 13 Oct 2005
Posts: 774
Location: Atlanta, GA
PostPosted: Tue May 25, 2010 3:32 am Post subject: Reply with quote
schtevie wrote:
I don't know why you are mixing on-court productivity with financial compensation here.
And herein lies your barrier to understanding VORP.
_________________
http://www.basketball-reference.com/blog/
Back to top
View user's profile Send private message Visit poster's website
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
PostPosted: Tue May 25, 2010 8:12 am Post subject: Reply with quote
schtevie wrote:
"Replacement player value" is not and cannot be a synonym for "relevant opportunity cost", except by mere coincidence.
Would you like to explain this or are you stating it's true by decree?
schtevie wrote:
First of all, Dean, I don't know why you are mixing on-court productivity with financial compensation here. Things, apparently, are confused enough already. And breakfast cereal analogies are not illuminating.
As Neil says, if you don't understand this, then you aren't going to understand any of the arguments here. Cost and benefit. There is a minimal cost for any player. If all you are interested in is rating players, then replacement level is of lesser interest. But if you have to make decisions involving money, it matters.
schtevie wrote:
My point is that there is no relevant, informative replacement value.
I would agree that there is no unique replacement value. But no relevant?
schtevie wrote:
If one is considering replacing Zydrunas Ilgauskas, say, with Shaquille O'Neal. The analysis ultimately centers on how many net points one produces in comparison to the other, and the manner in which they are assumed to do so.
Not completely true. If you know guys get injured, how much they don't play and get replaced by some replacement level guy, then you need that replacement level to really understand the deal.
schtevie wrote:
If however one is looking for a new bench warmer to replace a bench warmer, fine, by all means compare him to a referential bench warmer. But at least compare to the relevant position and not some non-existent composite.
This is purely semantics. Replacement level is non-unique, so yes, you can do one for varying positions (or a number of reasons). But people here have generally looked for a league standard and that's fine. Certainly far better than arguing there is no relevant or informative value.
schtevie wrote:
mtamada wrote:
...
People who ignore replacement level will overvalue the longer minutes but lower performing player. We see this all the time in Hall of Fame debates, where all too often the arguments are made based on the accumulation of stats such as points and rebounds, and even per game statistics, and too little attention is paid to the player's performance level. (The problem is even worse in baseball where some people will point at Pete Rose's 4,256 hits, and fail to take into account how many outs he made in accumulating those hits.)
...
It is simply not the case that ignoring replacement level (whatever that is - and again, the problem in basketball is estimating value, which obtains to determining any notional benchmark. Assuming an accurate and precise benchmark - even if of dubious comparative value - isn't a solution.) overvalues or undervalues any type of player. Any such analysis must be coherent and stand on its own.
Huh? MikeT is completely on target here.
schtevie wrote:
And just to emphasize the point, the motivation of this detour was the discussion of VORP in terms of the draft. Such an exercise perfectly illustrates the shortcomings. Every player in the draft, from the highly skilled lottery picks to the workhorses of the late second round are compared against a common imprecise composite. Again, to what gain?
Draft picks, in particular, need to be evaluated against a replacement level because many are not that different from the easy-to-obtain free agent at replacement level.
schtevie wrote:
Finally,
gabefarkas wrote:
schtevie wrote:
This isn't baseball, where because of the rules of the game, the worst players basically get as many at bats as the best.
Umm, just as an aside, that's not actually true. Batting order dictates that some players will consistently get more plate appearances than others. Worse hitters are at the back of the lineup, just like worse shooters in basketball tend to take less FGA.
Let me restate: basically, the worst players get as many at bats as the best.
Again, huh? For guys on a roster all season long, you can see guys with 700 plate appearances and guys with 100. And, as Gabe says, guys at the top of a lineup inherently can get 100+ extra plate appearances a season over guys at the bottom. Don't get ridiculous.
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers.
Mike G
Joined: 14 Jan 2005
Posts: 3612
Location: Hendersonville, NC
PostPosted: Thu May 20, 2010 5:05 pm Post subject: Replacement player value Reply with quote
If .80 of average efficiency differential (ED) is Replacement Level (RL) play, then a team of RL players, averaging a score of 80-100, would be expected to win about 5% of their games. They might go 4-78.
A team that averages 70-100 should only win 1% of their games. That's pretty close to 0-82.
To win 10 of 82 games, RL would be around .86 or .87 ED.
No team has ever been as bad as 4-78, and no team has ever been without any good player for a whole season. Of course, a team might have some players worse than RL and still average RL.
Here are 2010 league average per-36 rates, and .80 of those rates:
Code:
per36 Pts Reb Ast Stl Blk PF TO
avg 15.0 6.2 3.2 1.1 0.7 3.1 2.0
80% 12.0 5.0 2.5 0.9 0.6 2.5 1.6
No reason to suppose PF and TO should be lower than avg for an RP, so we do something else with those.
To the eye, though, the positives of line 1 look like a weak starter or decent bench guy. The 80% line is decidedly weak.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
BobboFitos
Joined: 21 Feb 2009
Posts: 201
Location: Cambridge, MA
PostPosted: Fri May 21, 2010 11:27 am Post subject: Reply with quote
Good stuff. The question though, focusing on scoring per 36, is are we assuming that .8 player is taking on the same offensive burden? Also, .8 TO of 2 per 36 should be 2.5 since it's the inverse of .8, no? If the shot attempts are the same, then their TS is falling from say an average value of ~54.3 to generate 15 per 36 (or 13.81 raw true shooting attempts) down to ~43.4 TS.*
*I basically figured to score 15 per 36 on the average NBA TS requires 13.81 TSA, so if you're scoring only 12 points on those same shot attempts, that's the relative drop in TS.
~~~
Do people view the worst historical record as a team full of replacement players? I don't. The 9 win 1973 76ers actually fell well below their pyth, but employed several players who had long-ish NBA careers. (Suggesting that they were not truly replaceable in the minimal sense) Granted I don't know anything about these players outside of skimming bref/looking at their career stats, but I think setting the door at a "10 win team" isn't really the hypothetical floor.
Does a team full of replacements - that is, a team that doesn't have ANYONE on the roster that would be deemed an improvement for any other team (maybe?) even win a game? There's a huge gap between a -20 and -30 PD.
_________________
http://pointsperpossession.com/
@PPPBasketball
Back to top
View user's profile Send private message
DSMok1
Joined: 05 Aug 2009
Posts: 611
Location: Where the wind comes sweeping down the plains
PostPosted: Fri May 21, 2010 12:04 pm Post subject: Reply with quote
A good start to studying replacement level would be to study baseball's accepted methodology. Hardball times has a few articles, for starters.
Basketball skill over the whole world's population is approximately a bell curve. We can all agree on that, right?
The players in the NBA are basically the right tail of that curve--so the curve should look fairly exponential, in terms of player quality from left to right.
However, once you get down to below average, there is a strong selective sampling issue--not every player of that quality actually is playing in the NBA. We end up seeing a slightly left-skewed graph of the quality of players in the NBA. Something like a log-normal distribution of sigma 1/2. Once the players get bad enough, none of them are in the NBA.
From a team's perspective, they are trying to get enough quality players on the court at once to win games. Is an NBA average player worth something? Yes, by definition. NBA-average players would win 41 games. What about somebody a little worse than average? Yes, he would help out also. You may need to play him, and players of his caliber (say an APM of -2) are not freely available.
So what is the threshold of replacement level? There is the rub. It is not the best player who isn't starting. That's way too high. It's probably above the 15th player on the bench--he could be old and bad but still under contract, or a youngster expected to develop.
Here's Tom Tango's definition:
Quote:
WAR is wins above replacement. Replacement is defined very specifically for my purposes: it’s the talent level for which you would pay the minimum salary on the open market, or for which you can obtain at minimal cost in a trade.
So we're not defining by how bad a team could be, but by what is available.
Which is rather hard to do, honestly.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
mtamada
Joined: 28 Jan 2005
Posts: 377
PostPosted: Fri May 21, 2010 4:39 pm Post subject: Reply with quote
DSMok1 wrote:
A good start to studying replacement level would be to study baseball's accepted methodology.
Yes, we do not know how many wins a replacement level player (or team of such players) would win, nor do we know what their stats would be.
But we do know the approximate description of a replacement level player: the freely available, marginal-on-or-off-the-roster player.
So I'd approach the estimation from three directions. First look at the stats of the 12th man (or these days, the 13th through 15th men) on NBA rosters. Defining the 12th man is a little tricky ... I'd basically use minutes played, but would make sure to exclude a Gred Oden-like player who may've played 12th man minutes due to injury not due to lack of ability.
That gives you an upper bound on how good the replacement level player is ... there might be some players who are below replacement level but manage to stick to a roster due to either a guaranteed contract or local popularity or for development or team sentimentality (Ken Griffey Jr. is like that right now for the Mariners) but on the whole, these are players who by definition made it onto an NBA roster, and are likely to be on average a bit better than the typical replacement player.
Next, look at the players who got called up to the NBA during the season. Again make sure to exclude good players who started the season injured or who were acquired as expensive free agents; include only minimum salary free agents and call-ups from the NBDL (again excluding players such as Hasheem Thabeet who may've been sent down for developmental reasons or whose ability level may've literally fluctuated throughout the season). This comes the closest I think to measuring replacement level players, but the small number of games and minutes that these players play may imply large standard errors in our estimates.
Finally, look literally at the minimum salary players. Some of these players will in fact be pretty good players who got belatedly discovered by some NBA team and who are truly better than replacement level, but are unlucky enough to be getting only a minimum salary contract. So this group again will tend to give you a bit of an overestimate of the replacement level.
None of these measures are perfect but between the three of them you could start really zeroing in on what the replacement level stats look like.
You'd probably want to do different stats for positions; a replacement level PG is going to have very different stats from a replacement level C.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3612
Location: Hendersonville, NC
PostPosted: Sat May 22, 2010 6:55 pm Post subject: Reply with quote
Are Win Shares (or WS/48) already above a replacement level?
This season, there were 368 players w >200 minutes and >0 WS/48. (Along with 18 who were negative.)
That's about 12.3 per team.
At 600 minutes, 310 positive and just 8 negative.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
DSMok1
Joined: 05 Aug 2009
Posts: 611
Location: Where the wind comes sweeping down the plains
PostPosted: Sat May 22, 2010 9:16 pm Post subject: Reply with quote
Mike G wrote:
Are Win Shares (or WS/4Cool already above a replacement level?
This season, there were 368 players w >200 minutes and >0 WS/48. (Along with 18 who were negative.)
That's about 12.3 per team.
At 600 minutes, 310 positive and just 8 negative.
I think not, because the total of all win shares in the league equals 41*30.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3612
Location: Hendersonville, NC
PostPosted: Sun May 23, 2010 6:14 am Post subject: Reply with quote
But if by definition only players above replacement level are adding wins, then only negative WS/48 players would be considered sub-replacement level.
If so, then .000 WS/48 = RL
Whether 12 per team is the 'right' number of >RL players, or whether the remaining 3 players x 30 rosters ('project' players, practice guys, etc) effectively ensures that there are no better players available, this might be close.
It's still possible for replacements to generate wins, as their distribution of performance varies from game to game. A whole team of RP's (and sub-RP's) can get it together against a merely bad team on a bad night, and pull off a win or 2, out of 82.
In the real world, though, the RP team is in a vacuum against relentless pressure. A Stackhouse or a Rodman will surely come out of exile or retirement and suit up.
Found this thread which discusses replacement level:
http://sonicscentral.com/apbrmetrics/vi ... php?t=2153
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
DSMok1
Joined: 05 Aug 2009
Posts: 611
Location: Where the wind comes sweeping down the plains
PostPosted: Sun May 23, 2010 8:24 am Post subject: Reply with quote
Mike G wrote:
But if by definition only players above replacement level are adding wins, then only negative WS/48 players would be considered sub-replacement level.
If so, then .000 WS/48 = RL
No.
The replacement level is set by what is available, not by how many wins are contributed. In baseball, we:
Quote:
figure out the offensive production that a team could expect from players not projected to be good enough to make a major league roster next year. These guys have fallen into that Four-A category, where they show more ability than your average Triple-A veteran but not enough to hold down a major league job. They’re usually available every winter as minor league free agents, via the Rule 5 draft, or as cheap trade acquisitions where a team can acquire one of these players without giving up any real talent in return.
As Sean showed in his article, and has been shown elsewhere, the expected value of a replacement level player is about negative 20 runs per 600 PA. Or, to phrase it a bit differently, if you lost a league average player and replaced him with a freely available guy, you’d lose about two wins.
(source)
In baseball, a "replacement level" team is still at about a .400 winning percentage--the scatter isn't as big.
In basketball, we need to figure out how good the players are that are "freely available" and consider that replacement level.
That could be done by looking at the distribution of players in terms of their APM, and see where the distribution transitions from "the tail of a bell curve" to more of a log-normal distribution--which would imply players of that quality are available.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
gabefarkas
Joined: 31 Dec 2004
Posts: 1313
Location: Durham, NC
PostPosted: Sun May 23, 2010 12:32 pm Post subject: Reply with quote
Instead of trying to theoretically come up with how many wins produced a replacement player would provide, why not try to come up with it empirically? What I mean is to try to identify actual replacement players, and then look at what they actually did on the court.
If we wanted to, from there we could even do permutation distributions, bootstrapping, jackknife, or any other resampling technique to come up with an empirical distribution for what would be expected from a replacement player.
So, what if we came up with a list of who we believe are actual replacement players? A starting point might be something along the lines of:
Aaron Gray
Aaron Miles
Alex Acker
Allan Ray
Andre Emmett
Awvee Storey
Bernard Robinson
Britton Johnsen
Casey Jacobsen
Chase Budinger
Chris Jefferies
Chris Quinn
Chris Taft
Daniel Ewing
Deng Gai
Dijon Thompson
Frank Williams
Gabe Pruitt
Josh McRoberts
etc etc etc
Does that make sense? Come up with a list of what you would consider replacement players, and then derive an empirical distribution from there.
Back to top
View user's profile Send private message Send e-mail AIM Address
Carlos
Joined: 21 Jan 2005
Posts: 64
Location: Montevideo, Uruguay
PostPosted: Sun May 23, 2010 12:38 pm Post subject: Reply with quote
What about looking at guys picked from the NBDL and undrafted players or midseason additions? It's relatively rare, but each year you see a few guys enter the league who were truly obtained at minimum costs. I remember the Grizziles a few years ago had all of their PGs injured and picked Elliot Perry basically off the street to play for a couple of games. It was unusual, sure, but I'd bet that every year there are a few cases like that, of teams playing truly marginal guys because of health issues.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3612
Location: Hendersonville, NC
PostPosted: Sun May 23, 2010 12:46 pm Post subject: Reply with quote
DSMok1 wrote:
The replacement level is set by what is available, not by how many wins are contributed. ..
And are these mutually exclusive?
If teams are actually utilizing players who are worse than available players, they aren't optimizing their wins.
You might even say they're putting guys on the floor who are removing wins, relative to the default position of 'best available replacements'.
Quote:
In baseball, a "replacement level" team is still at about a .400 winning percentage ...
Winning 40% means winning 80% of average.
This really is a far cry from 80% of average efficiency differential -- which in NBA translates to about 5% wins. This is an artifact of 48-minute games (actually, of 90-95 possession games). In a one-possession game, your W/L = your ED.
How do you explain baseball teams that win <40% ? They're trying to lose?
If baseball statisticians have made serious logical errors, do we wish to perpetuate them?
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
DLew
Joined: 13 Nov 2006
Posts: 224
PostPosted: Sun May 23, 2010 1:45 pm Post subject: Reply with quote
Mike,
You always seem hung up by anything that suggests teams are not optimizing. Why is this? It seems clear that many teams are not. Some teams play guys who are worse than the freely available players. Do you disagree?
Back to top
View user's profile Send private message
schtevie
Joined: 18 Apr 2005
Posts: 413
PostPosted: Sun May 23, 2010 2:34 pm Post subject: Reply with quote
Can someone say, slowly and clearly, why one should be impressed with the concept of a replacement level player? In particular, why is it useful at all in discussing basketball?
I kind of can see in baseball, where the production function of the game is relatively simple, defining some inherently arbitrary below-average standard as a benchmark for comparing player values and this becoming a convention over time. But basketball is different. The production function is very complicated; it is not clear what the below-average standard should be; and it is difficult to measure player contributions precisely.
It is far more intuitive to at least eliminate one of these problems and have the benchmark of comparison be the average player, whether this value be zero (in terms of net points) or 8.2 (for per game WS).
The futility can be illustrated with 2007-08 data, chosen to utilize Ilardi & Barizlai's multi-year APM numbers as estimates of value. Let's take as an arbitrary starting point, the 360th ranked player in minutes per game (so, kind of the 12th man on the 30th team): Malik Rose, playing 10.1 minutes per game. Finding the 15 players above and below this mark (again, arbitrary) that meet the minimum minutes cut-off for I & B's estimates, the average multi-year APM is -3.0.
Might this replacement level definition be consistent over time? Perhaps. Ten minutes seems plausible. But even if the contribution of ten minute players hurt their teams to the tune of 3 points per 100 possessions on average, there is a great deal of dispersion around the mean. The standard deviation in the instance is 4.1, the estimates ranging from Ronnie Price's 7.17 to Dominic McGuire's -12.07. As such, what is the point of the exercise? What does the concept of a replacement level player gain us?
Back to top
View user's profile Send private message
Kevin Pelton
Site Admin
Joined: 30 Dec 2004
Posts: 979
Location: Seattle
PostPosted: Sun May 23, 2010 3:07 pm Post subject: Reply with quote
http://web.archive.org/web/200404040408 ... 7557.shtml
Back to top
View user's profile Send private message Send e-mail Visit poster's website
DSMok1
Joined: 05 Aug 2009
Posts: 611
Location: Where the wind comes sweeping down the plains
PostPosted: Sun May 23, 2010 4:56 pm Post subject: Reply with quote
Kevin Pelton wrote:
http://web.archive.org/web/200404040408 ... 7557.shtml
There we go. I knew you had that around somewhere, but I couldn't find it.
That article is the baseline for A) Why replacement l level is important and B) a first effort at how to arrive at that level.
schtevie wrote:
Can someone say, slowly and clearly, why one should be impressed with the concept of a replacement level player? In particular, why is it useful at all in discussing basketball?
Replacement level is required to determine the actual value of a player to the team. Since the players are all pulled from the tail end of a bell curve, at a certain level, there are lots of players available--and thus basically freely available. The value of any player in the NBA is actually their marginal value above the players that are freely available.
schtevie wrote:
I kind of can see in baseball, where the production function of the game is relatively simple, defining some inherently arbitrary below-average standard as a benchmark for comparing player values and this becoming a convention over time. But basketball is different. The production function is very complicated; it is not clear what the below-average standard should be; and it is difficult to measure player contributions precisely.
It is far more intuitive to at least eliminate one of these problems and have the benchmark of comparison be the average player, whether this value be zero (in terms of net points) or 8.2 (for per game WS).
The futility can be illustrated with 2007-08 data, chosen to utilize Ilardi & Barizlai's multi-year APM numbers as estimates of value. Let's take as an arbitrary starting point, the 360th ranked player in minutes per game (so, kind of the 12th man on the 30th team): Malik Rose, playing 10.1 minutes per game. Finding the 15 players above and below this mark (again, arbitrary) that meet the minimum minutes cut-off for I & B's estimates, the average multi-year APM is -3.0.
Might this replacement level definition be consistent over time? Perhaps. Ten minutes seems plausible. But even if the contribution of ten minute players hurt their teams to the tune of 3 points per 100 possessions on average, there is a great deal of dispersion around the mean. The standard deviation in the instance is 4.1, the estimates ranging from Ronnie Price's 7.17 to Dominic McGuire's -12.07. As such, what is the point of the exercise? What does the concept of a replacement level player gain us?
One issue of looking at it purely in a MPG frame is that there is a ton of error associated in looking at players of that level. Another is that a player that is playing 10 MPG on the Lakers is a lot better than a player that is getting 10 MPG on the Nets.
I, however, also came to the same number (-3 pts/100 pos.) when evaluating players using Statistical +/-. It appears that a player of that caliber is nearly freely available. Interestingly, translating that -3 number into a team efficiency differential and running the Pythagorean formula yields the 10 win number yet again.
Author Message
Kevin Pelton
Site Admin
Joined: 30 Dec 2004
Posts: 979
Location: Seattle
PostPosted: Sun May 23, 2010 5:22 pm Post subject: Reply with quote
DSMok1 wrote:
There we go. I knew you had that around somewhere, but I couldn't find it.
Unfortunately, the Hoopsworld archive is nonexistent, so that stuff is tough to find unless you know exactly where you're looking.
The big problem with using that ancient linear-weights stat is there was no defined "average," so I wasn't able to look at replacement level as a percentage of average.
Were I to do it again, I'd want to do some sort of cross-metric comparison like AaronB did with the draft a few years back: http://www.82games.com/barzilai1.htm
That way, you ensure biases in the specific metric aren't affecting your conclusion.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
schtevie
Joined: 18 Apr 2005
Posts: 411
PostPosted: Sun May 23, 2010 5:40 pm Post subject: Reply with quote
DSMok1 wrote:
Kevin Pelton wrote:
http://web.archive.org/web/200404040408 ... 7557.shtml
There we go. I knew you had that around somewhere, but I couldn't find it.
That article is the baseline for A) Why replacement l level is important and B) a first effort at how to arrive at that level.
Having (re)read the article, I remain at my initial conclusion. The best I can tell is that defining replacement level is important because some guys in baseball do it. And the institutional/economic conditions for using it there do not obtain in basketball
DSMok1 wrote:
schtevie wrote:
Can someone say, slowly and clearly, why one should be impressed with the concept of a replacement level player? In particular, why is it useful at all in discussing basketball?
Replacement level is required to determine the actual value of a player to the team. Since the players are all pulled from the tail end of a bell curve, at a certain level, there are lots of players available--and thus basically freely available. The value of any player in the NBA is actually their marginal value above the players that are freely available.
Let's be clear. What is written in bold is at best misleading but more basically simply untrue. First let's define value. Never mind economic value, because that adds an unnecessary layer of complication. Let's stick with competitive value, which ultimately is a function of a player's influence on the scoreboard on a per possession basis. Such value, to a first approximation is completely independent of the value of players that are freely available. One can try to come up with a meaningful (as in accurate and precise) estimate of the marginal value of this representative replacement player (which is the entire exercise in futility). But stipulating success in that exercise, what does one have or gain? Nothing but an arbitrary benchmark. So LBJ is valued at 19 points per 100 possessions (or whatever) above the replacement player, instead of 16 points better than average. And?
DSMok1 wrote:
schtevie wrote:
I kind of can see in baseball, where the production function of the game is relatively simple, defining some inherently arbitrary below-average standard as a benchmark for comparing player values and this becoming a convention over time. But basketball is different. The production function is very complicated; it is not clear what the below-average standard should be; and it is difficult to measure player contributions precisely.
It is far more intuitive to at least eliminate one of these problems and have the benchmark of comparison be the average player, whether this value be zero (in terms of net points) or 8.2 (for per game WS).
The futility can be illustrated with 2007-08 data, chosen to utilize Ilardi & Barizlai's multi-year APM numbers as estimates of value. Let's take as an arbitrary starting point, the 360th ranked player in minutes per game (so, kind of the 12th man on the 30th team): Malik Rose, playing 10.1 minutes per game. Finding the 15 players above and below this mark (again, arbitrary) that meet the minimum minutes cut-off for I & B's estimates, the average multi-year APM is -3.0.
Might this replacement level definition be consistent over time? Perhaps. Ten minutes seems plausible. But even if the contribution of ten minute players hurt their teams to the tune of 3 points per 100 possessions on average, there is a great deal of dispersion around the mean. The standard deviation in the instance is 4.1, the estimates ranging from Ronnie Price's 7.17 to Dominic McGuire's -12.07. As such, what is the point of the exercise? What does the concept of a replacement level player gain us?
One issue of looking at it purely in a MPG frame is that there is a ton of error associated in looking at players of that level. Another is that a player that is playing 10 MPG on the Lakers is a lot better than a player that is getting 10 MPG on the Nets.
I, however, also came to the same number (-3 pts/100 pos.) when evaluating players using Statistical +/-. It appears that a player of that caliber is nearly freely available. Interestingly, translating that -3 number into a team efficiency differential and running the Pythagorean formula yields the 10 win number yet again.
Here again, I must disagree. First, I wasn't really advocating for an MPG frame, or any other frame. MPG has particular shortcomings I am sure. Primarily I am guessing in terms of biasing the "true" replacement value upward. (That is, within the low minute cohort, that there are up and coming players, earning their minutes who are actually more valuable than "true" replacement players.) My point in picking an arbitrary standard that seemed basically fine was to illustrate the futility of the exercise.
And second, to quibble, I don't think there is necessarily any reason to believe, a priori, that a Lakers 10 MPG player would be better than a Nets 10 MPG player. The Lakers are clearly a better team, but this doesn't necessarily mean that their 12th man is better.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3597
Location: Hendersonville, NC
PostPosted: Mon May 24, 2010 8:18 am Post subject: Reply with quote
Schtevie, I believe the thought process is this:
NBA teams are forced to use below-average players for many minutes per season. While we all agree certain guys are below average, the debate is about whether they're actually hurting their teams' ability to win.
If teams are using players who are below the value of available replacements, then it's arguable that the use of these players actually does hurt them. If they're the best available, and you have to play someone, then they're as good as it gets.
In another thread, DSMok1 chose to evaluate the wins contributed by players at various draft positions, using [WS/48 - .025] . In other words, wins above replacement, defined as 25% of average win-production rate.
Since we find players such as Monta Ellis getting 41 mpg, at a rate of .023 WS/48, it's a wonder that legendary coach Don Nelson would be inclined to 'hurt' his team by playing a guy who adds nothing to the team's win probability, and who could be easily replaced.
If we have an educated estimate of replacement value -- perhaps .50 , .20 , or -.10 of average -- it helps us form statements about a player's real 'value'.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
DLew
Joined: 13 Nov 2006
Posts: 224
PostPosted: Mon May 24, 2010 9:56 am Post subject: Reply with quote
Mike,
I don't think anyone would argue Monta Ellis adds nothing to his team's win probability in absolute terms. Clearly having him out there is better than having only 4 players on the court. I agree only with the second half of your statement, that (according to Win Shares) he adds nothing above what a freely available replacement player (assuming the 0.025 threshold) would add. Interestingly enough, the Warriors employed several freely available replacement players this season (Anthony Tolliver, Coby Karl, Reggie Williams). Tolliver and Williams rated somewhat higher than Ellis, Karl not so much.
Back to top
View user's profile Send private message
DSMok1
Joined: 05 Aug 2009
Posts: 610
Location: Where the wind comes sweeping down the plains
PostPosted: Mon May 24, 2010 10:04 am Post subject: Reply with quote
Here's what a graph of Win Shares over Replacement vs. MPG vs. Team Efficiency Differential looks like, using the 0.025 level, historically:
As you can see, a league average team basically doesn't play anybody that's below replacement level. However, as soon as you get below league average, a few players of that type start to play. Once you get down into BAD territory, then a lot of the replacement-level players start to appear.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3597
Location: Hendersonville, NC
PostPosted: Mon May 24, 2010 10:53 am Post subject: Reply with quote
DLew wrote:
I don't think anyone would argue Monta Ellis adds nothing to his team's win probability in absolute terms. ..
And yet, this thread was prompted by another in which DSMok1 stated:
Quote:
They were at or below replacement level, so they contributed nothing to the team, other than taking up space.
(which was in reply to this list --
http://www.basketball-reference.com/pla ... i?id=NG7tG
-- of players with 4-yr WS/48 <.030)
Monta Ellis was .023 WS/48 this year.
It seems to me that the draft pick value study is potentially a monumental work, but it's based on a couple of questionable assumptions:
- that WS/48 is a true measure of player value
- that .025 WS/48 = replaceable value.
Most of the time, when you have such large samples (rookies through 4 years), the idiosyncrasies of a metric will average out. But, given that the better rookies tend to be drafted by the weaker teams, such a study could be skewed by the 2 assumptions above.
Indeed, by PER, Ellis is well above NBA average (16.7) -- RWilliams (16.0), Tolliver (13.7), Karl (7.1) not so much.
By my own metric, Ellis was a bit over par, at .106 eW/48; the others .88, .79, and .13 respectively.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
DSMok1
Joined: 05 Aug 2009
Posts: 610
Location: Where the wind comes sweeping down the plains
PostPosted: Mon May 24, 2010 2:24 pm Post subject: Reply with quote
Mike G wrote:
DLew wrote:
I don't think anyone would argue Monta Ellis adds nothing to his team's win probability in absolute terms. ..
And yet, this thread was prompted by another in which DSMok1 stated:
Quote:
They were at or below replacement level, so they contributed nothing to the team, other than taking up space.
(which was in reply to this list --
http://www.basketball-reference.com/pla ... i?id=NG7tG
-- of players with 4-yr WS/48 <.030)
Monta Ellis was .023 WS/48 this year.
It seems to me that the draft pick value study is potentially a monumental work, but it's based on a couple of questionable assumptions:
- that WS/48 is a true measure of player value
- that .025 WS/48 = replaceable value.
Most of the time, when you have such large samples (rookies through 4 years), the idiosyncrasies of a metric will average out. But, given that the better rookies tend to be drafted by the weaker teams, such a study could be skewed by the 2 assumptions above.
Indeed, by PER, Ellis is well above NBA average (16.7) -- RWilliams (16.0), Tolliver (13.7), Karl (7.1) not so much.
By my own metric, Ellis was a bit over par, at .106 eW/48; the others .88, .79, and .13 respectively.
I think "absolute" and "marginal" are being confused. Sure, Monta Ellis adds something to the team. He's better than air; he's better than me. However, he is not necessarily better than a "replacement level" player.
In fact, using wins as an absolute measure is one of the problems. Wins are nonlinear in this regard. I would use points/possession as the measure, since that is linear. I could say that I am "Absolutely" -25 pts/100poss, and Monta Ellis is -3 pts/100poss. How would we measure wins, though?
In fact, there is no "floor" for NBA players in terms of value added. There is no easy way to quantify absolute value, accept in terms of a rate--not like baseball, where I would simply never get a hit. In the NBA, it is all part of 1 large continuum.
Me < Monta < Russel Westbrook < Lebron. That we can say. We can even quantify, in terms of points, a level for the players. Using RAPM, I would maybe be -25 pts, Ellis -5, Westbrook +1, Lebron +6 pts/100 poss. How much would each of us be worth to an NBA team? Or more precisely, to an average NBA team?
Well, I think we all could agree that I would be worth nothing. Truly 0. Somewhere, value starts to appear. Where? Where the level of talent starts to become scarce.
Suppose the whole population's basketball scoring value is basically one huge bell curve, centered at -44, with a standard deviation of 9.7, and there are 50,000,000 in the population. (That fits what we see decently well). Well, there are only 143 players above 0 pts/100. (The placement of 0 is actually arbitrary and dependent on the league, but the point stands). So the replacement level would be below that--there's a lot more than 143 players.
Illustrating the above paragraph
However, there are a few players that don't end up in the NBA that are below 0 pts/100 poss. In fact, quite a few that could have been filling in that part of the bell curve from -4 to 0. Whether they exist, I don't know--the NBA players are a subset that received specialized training AFTER demonstrating they were in this upper echelon. Sort of a selective sampling issue.
However, there are enough out there that at, say, -3 (to pick a number) a team can find such a player with a little effort but minimal expenditure. European players, NBDL, draft, free agents--players are available at that level for a minimum contract.
So, any player of that level is worth basically 0 to the team--they are all interchangeable. If one gets hurt, the team does not suffer at all--just pick somebody else up.
The value of a player to the team would then be quantified relative to the replacement level. I don't know that the relation is linear (in fact, I'm sure it's not--as opposed to baseball).
Back to top
View user's profile Send private message Send e-mail Visit poster's website
DLew
Joined: 13 Nov 2006
Posts: 224
PostPosted: Mon May 24, 2010 4:02 pm Post subject: Reply with quote
I should elaborate on the point I was trying to make: Replacement level is a useful construct for assessing player value because absolute zero is so difficult to assess. Value over replacement is much easier to determine than just 'value' because replacement level performance exists in the NBA, absolute zero level performance has never existed in the NBA so it's extremely difficult to calculate where it might fall.
Generally I think value over replacement is useful to explicitly calculate because every rating system is a value over something, it's just those that don't explicitly define zero level to some meaningful mark are much harder to interpret.
Back to top
View user's profile Send private message
schtevie
Joined: 18 Apr 2005
Posts: 411
PostPosted: Mon May 24, 2010 4:18 pm Post subject: Reply with quote
DSMok1, I was hoping for more persuasive arguments as to the utility of the concept of a replacement player in NBA basketball.
It surely is possible that there is a potential supply of such players numbering in the hundreds or thousands. But this is ultimately irrelevant. The demand for such players is tiny, and ever shall it remain so. And because the market is so thin, there is no real reason to reify this concept. To the contrary, it only creates confusion with no attendant benefits.
Who are replacement players? Basically, they are practice bodies and insurance policies. They play when there is injury, garbage time, and for a few minutes here and there for the benefit of team cohesion.
Why is it desirable to establish the "average" performance of such players? Surely not as a benchmark for those mostly outside this group. This isn't baseball, where because of the rules of the game, the worst players basically get as many at bats as the best. And, this aside, it isn't meaningful to speak of a single replacement level anyway. Eli W's clearly showed that each position has its own average.
To beat a horse that should die, basketball is not baseball. Basketball needs improved estimates of player values, not arbitrary referential standards for the worst players.
Back to top
View user's profile Send private message
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
PostPosted: Mon May 24, 2010 4:56 pm Post subject: Reply with quote
Replacement player value is useful because everything we do is about making a decision, usually a cost vs a benefit. Costs are usually straightforward - a player salary, though negotiation of that at times is the issue. Benefits are harder - the value of a win and the responsibility of a player for those. If you're going to pay a lot of money for a player, you should want to know their value, especially relative to the lowest price player out there. Do you buy generic if you think it's just as good as Frosted Flakes because it costs less? Or do you assume Frosted Flakes are better because they cost more? When our decisions value a lot more than Frosted Flakes, we want to know how good both the original and the generic are. In our case, there are a lot of essentially "generics" available every year. Assigning a general replacement value saves you from having to really value them all, which can be tricky with their limited minutes.
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers.
Back to top
View user's profile Send private message Visit poster's website
mtamada
Joined: 28 Jan 2005
Posts: 377
PostPosted: Mon May 24, 2010 6:44 pm Post subject: Reply with quote
Yup, replacement level is crucial in making roster decisions. Even if we ignore salaries, we still have to compare player qualities somehow, which means making tradeoffs between high performance, low minute players versus lower performance but higher minute players -- KevinP's 2004 comparison of Shaq vs McGrady is an excellent one. Ignoring for the moment salary and what the teams' needs are, i.e. team fit, would you trade one of those players for the other? Which was more valuable, McGrady's higher points, assists, etc. in 2003, or Shaq's higher per-minute productivity? The correct baseline to compare those players against is the replacement level player, not the average player (even the Lakers, Celtics, and Magic don't have a marginal player who is at league average level, rather their marginal player is way below that level).
People who ignore replacement level will overvalue the longer minutes but lower performing player. We see this all the time in Hall of Fame debates, where all too often the arguments are made based on the accumulation of stats such as points and rebounds, and even per game statistics, and too little attention is paid to the player's performance level. (The problem is even worse in baseball where some people will point at Pete Rose's 4,256 hits, and fail to take into account how many outs he made in accumulating those hits.)
Back to top
View user's profile Send private message
gabefarkas
Joined: 31 Dec 2004
Posts: 1313
Location: Durham, NC
PostPosted: Mon May 24, 2010 8:41 pm Post subject: Reply with quote
schtevie wrote:
This isn't baseball, where because of the rules of the game, the worst players basically get as many at bats as the best.
Umm, just as an aside, that's not actually true. Batting order dictates that some players will consistently get more plate appearances than others. Worse hitters are at the back of the lineup, just like worse shooters in basketball tend to take less FGA.
Back to top
View user's profile Send private message Send e-mail AIM Address
schtevie
Joined: 18 Apr 2005
Posts: 411
PostPosted: Mon May 24, 2010 9:57 pm Post subject: Reply with quote
HoopStudies wrote:
Replacement player value is useful because everything we do is about making a decision, usually a cost vs a benefit. Costs are usually straightforward - a player salary, though negotiation of that at times is the issue. Benefits are harder - the value of a win and the responsibility of a player for those. If you're going to pay a lot of money for a player, you should want to know their value, especially relative to the lowest price player out there. Do you buy generic if you think it's just as good as Frosted Flakes because it costs less? Or do you assume Frosted Flakes are better because they cost more? When our decisions value a lot more than Frosted Flakes, we want to know how good both the original and the generic are. In our case, there are a lot of essentially "generics" available every year. Assigning a general replacement value saves you from having to really value them all, which can be tricky with their limited minutes.
"Replacement player value" is not and cannot be a synonym for "relevant opportunity cost", except by mere coincidence.
First of all, Dean, I don't know why you are mixing on-court productivity with financial compensation here. Things, apparently, are confused enough already. And breakfast cereal analogies are not illuminating.
My point is that there is no relevant, informative replacement value. If one is considering replacing Zydrunas Ilgauskas, say, with Shaquille O'Neal. The analysis ultimately centers on how many net points one produces in comparison to the other, and the manner in which they are assumed to do so. One can use as a bench mark the notional productivity average of seven foot stiffs in the D-League, in Europe, and/or parts unknown (or if one is really going to have an aggregate replacement value benchmark, throw in the weighted contributions of point guards who can't shoot, or unathletic small forwards) but why? The comparison doesn't require this, and it is not only a completely uninformative detour on the face of it, but the replacement level estimate is necessarily imprecise.
If however one is looking for a new bench warmer to replace a bench warmer, fine, by all means compare him to a referential bench warmer. But at least compare to the relevant position and not some non-existent composite.
mtamada wrote:
Yup, replacement level is crucial in making roster decisions. Even if we ignore salaries, we still have to compare player qualities somehow, which means making tradeoffs between high performance, low minute players versus lower performance but higher minute players -- KevinP's 2004 comparison of Shaq vs McGrady is an excellent one. Ignoring for the moment salary and what the teams' needs are, i.e. team fit, would you trade one of those players for the other? Which was more valuable, McGrady's higher points, assists, etc. in 2003, or Shaq's higher per-minute productivity? The correct baseline to compare those players against is the replacement level player, not the average player (even the Lakers, Celtics, and Magic don't have a marginal player who is at league average level, rather their marginal player is way below that level).
People who ignore replacement level will overvalue the longer minutes but lower performing player. We see this all the time in Hall of Fame debates, where all too often the arguments are made based on the accumulation of stats such as points and rebounds, and even per game statistics, and too little attention is paid to the player's performance level. (The problem is even worse in baseball where some people will point at Pete Rose's 4,256 hits, and fail to take into account how many outs he made in accumulating those hits.)
Mike, I do not quite get your point or your disagreement with my position. I think we can agree that it is good to compare like to like. And that is precisely what I am advocating. In the form of not putting forth useless benchmarks. Every player should be analyzed in terms of what he can produce in terms of his own skill set in a hypothetical situation. The only bench mark that matters is the expected effect on the scoreboard (and/or resulting expected win totals). A replacement player - whatever he is - is only a useful point of comparison in the coincidental case that you are looking for a replacement player.
It is simply not the case that ignoring replacement level (whatever that is - and again, the problem in basketball is estimating value, which obtains to determining any notional benchmark. Assuming an accurate and precise benchmark - even if of dubious comparative value - isn't a solution.) overvalues or undervalues any type of player. Any such analysis must be coherent and stand on its own.
And just to emphasize the point, the motivation of this detour was the discussion of VORP in terms of the draft. Such an exercise perfectly illustrates the shortcomings. Every player in the draft, from the highly skilled lottery picks to the workhorses of the late second round are compared against a common imprecise composite. Again, to what gain?
Finally,
gabefarkas wrote:
schtevie wrote:
This isn't baseball, where because of the rules of the game, the worst players basically get as many at bats as the best.
Umm, just as an aside, that's not actually true. Batting order dictates that some players will consistently get more plate appearances than others. Worse hitters are at the back of the lineup, just like worse shooters in basketball tend to take less FGA.
Let me restate: basically, the worst players get as many at bats as the best.
Back to top
View user's profile Send private message
Neil Paine
Joined: 13 Oct 2005
Posts: 774
Location: Atlanta, GA
PostPosted: Tue May 25, 2010 3:32 am Post subject: Reply with quote
schtevie wrote:
I don't know why you are mixing on-court productivity with financial compensation here.
And herein lies your barrier to understanding VORP.
_________________
http://www.basketball-reference.com/blog/
Back to top
View user's profile Send private message Visit poster's website
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
PostPosted: Tue May 25, 2010 8:12 am Post subject: Reply with quote
schtevie wrote:
"Replacement player value" is not and cannot be a synonym for "relevant opportunity cost", except by mere coincidence.
Would you like to explain this or are you stating it's true by decree?
schtevie wrote:
First of all, Dean, I don't know why you are mixing on-court productivity with financial compensation here. Things, apparently, are confused enough already. And breakfast cereal analogies are not illuminating.
As Neil says, if you don't understand this, then you aren't going to understand any of the arguments here. Cost and benefit. There is a minimal cost for any player. If all you are interested in is rating players, then replacement level is of lesser interest. But if you have to make decisions involving money, it matters.
schtevie wrote:
My point is that there is no relevant, informative replacement value.
I would agree that there is no unique replacement value. But no relevant?
schtevie wrote:
If one is considering replacing Zydrunas Ilgauskas, say, with Shaquille O'Neal. The analysis ultimately centers on how many net points one produces in comparison to the other, and the manner in which they are assumed to do so.
Not completely true. If you know guys get injured, how much they don't play and get replaced by some replacement level guy, then you need that replacement level to really understand the deal.
schtevie wrote:
If however one is looking for a new bench warmer to replace a bench warmer, fine, by all means compare him to a referential bench warmer. But at least compare to the relevant position and not some non-existent composite.
This is purely semantics. Replacement level is non-unique, so yes, you can do one for varying positions (or a number of reasons). But people here have generally looked for a league standard and that's fine. Certainly far better than arguing there is no relevant or informative value.
schtevie wrote:
mtamada wrote:
...
People who ignore replacement level will overvalue the longer minutes but lower performing player. We see this all the time in Hall of Fame debates, where all too often the arguments are made based on the accumulation of stats such as points and rebounds, and even per game statistics, and too little attention is paid to the player's performance level. (The problem is even worse in baseball where some people will point at Pete Rose's 4,256 hits, and fail to take into account how many outs he made in accumulating those hits.)
...
It is simply not the case that ignoring replacement level (whatever that is - and again, the problem in basketball is estimating value, which obtains to determining any notional benchmark. Assuming an accurate and precise benchmark - even if of dubious comparative value - isn't a solution.) overvalues or undervalues any type of player. Any such analysis must be coherent and stand on its own.
Huh? MikeT is completely on target here.
schtevie wrote:
And just to emphasize the point, the motivation of this detour was the discussion of VORP in terms of the draft. Such an exercise perfectly illustrates the shortcomings. Every player in the draft, from the highly skilled lottery picks to the workhorses of the late second round are compared against a common imprecise composite. Again, to what gain?
Draft picks, in particular, need to be evaluated against a replacement level because many are not that different from the easy-to-obtain free agent at replacement level.
schtevie wrote:
Finally,
gabefarkas wrote:
schtevie wrote:
This isn't baseball, where because of the rules of the game, the worst players basically get as many at bats as the best.
Umm, just as an aside, that's not actually true. Batting order dictates that some players will consistently get more plate appearances than others. Worse hitters are at the back of the lineup, just like worse shooters in basketball tend to take less FGA.
Let me restate: basically, the worst players get as many at bats as the best.
Again, huh? For guys on a roster all season long, you can see guys with 700 plate appearances and guys with 100. And, as Gabe says, guys at the top of a lineup inherently can get 100+ extra plate appearances a season over guys at the bottom. Don't get ridiculous.
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers.