Ben F.
PostPosted: Fri Aug 12, 2005 11:36 am Post subject: Usage vs. Efficiency Reply with quote
I'd like to raise this question again, after we've had some time off from talking about it, because I think that it is a very fundamental question that we have not yet answered - Does efficiency drop as usage rises? And if so, how much?
Obviously it's player specific - the players we traditionally think of as "stars" are those that can (usually) maintain high levels of efficiency at high usage rates. The best "role players" are those that are efficient at low usage rates.
The question is, if you gave that role player a much larger role, so that his usage climbed drastically, would he be able to maintain his efficiency? And if you limited the usage of the star, would his efficiency rise?
I looked into this a little bit, and looked at the 100 most efficient players of the last 5 years, according to ORating at b-r.com. I don't know that much statistically, but I found a very slight correlation (-.3) between usage and efficiency that suggests that as usage rises efficiency decreases.
But that's only very slight. I believe if you just take the top 100 from this last year you actually find the opposite trend, although again, it's VERY slight.
It seems to me that theoretically, the reason "stars" help the team so much are for the reasons Dean O layed out in BoP, in his chapter on skill curves - that their usage at a fairly high efficiency helps the other players on the team maintain their ideal usage, and therefore their highest efficiency. That was his theory on Iverson - that although he didn't seem that efficient, he could use very large numbers of possessions at a slightly below average efficiency, but in turn the players around him could maximize their efficiency because they didn't have to force anything. In this way the team was built almost perfectly around him, and it carried them to the championship.
That all makes very logical sense - but as Dean also said, it's very hard to prove statistically. There's really not a large sample size to be able to view what happens when role players take on possessions and stars lose them. And it also seems that you will almost never see a "role player" take on a lot of possessions for extended periods of time - they default to what is most comfortable, I suppose. So it's very hard to study.
I suppose this is all a plea to say: Prove it to me! (I also suppose this is confirmation bias, but hell, it seems so damn logical.) Where is the proof that Efficiency drops as usage rises? How do we go about studying this? How can we find that point that Dean described, where all the players are using right around the perfect amount of possessions?
Mike G
PostPosted: Fri Aug 12, 2005 11:52 am Post subject: Re: Usage vs. Efficiency Reply with quote
FFSBasketball wrote:
... Where is the proof that Efficiency drops as usage rises? How do we go about studying this? ...
Get into a game and force yourself to shoot more than you are comfortable with; handle the ball more than you really want to.
Back to top
View user's profile Send private message Send e-mail
Kevin Pelton
PostPosted: Fri Aug 12, 2005 12:10 pm Post subject: Reply with quote
I think the biggest obstacle to any study is the natural optimization. When a role player does take on a lot more possessions in a given game, it's probably because he's hot or has a favorable matchup or something that should improve his efficiency as well.
Now maybe if we got to be coaches and were willing to sacrifice some regular-season games so we could force Brent Barry to shoot 20 times a night that would be different, but otherwise I don't think this will ever be "provable." That's the nature of studying something you can't control.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Dan Rosenbaum
PostPosted: Fri Aug 12, 2005 3:13 pm Post subject: Reply with quote
KevinP's post is very much on the mark. This relationship between usage and efficiency is very hard to estimate empirically. Well, here is a strategy and a fairly advanced econometrics lesson.
We basically have the regression equation.
EFF = B1 + B2*USAGE + e,
where EFF is efficiency, USAGE is possession usage, e is a random error term, and B1 and B2 are regression parameters.
But the problem is that in games where players are more efficient, the coach will encourage them to use more possessions. Thus, there is positive correlation between USAGE and the error term. This will result in an estimate of B2 using data from games during a season to get an estimate of B2 that is too high, i.e. that is upwardly biased.
One trick commonly used in economics is to use a technique called instrumental variables. The idea is to find an instrument - Z - that would be correlated with USAGE (either positive or negative correlation will do), but would not be correlated with the error term, i.e. the part of efficiency not explained by possession usage.
Here is a candidate for an instrument - the average possession usage in the previous season of a player's current season teammates. In computing this average I would weight by the minutes played in the current season.
The logic behind this methodology is that if a player gets stuck with a bunch of teammates who in the past did not use many possessions, he will be forced to up his own possession usage. It allows us to use "natural experiments" in the data where players are forced into new roles.
That is better than just looking at players who change a lot in possession usage, because that change could be caused by the player getting a lot better. Using the previous season possession usage of the teammates gets around that issue.
Puting the data together and running this instrumental variables regression is quite a bit of work, so I am not quite sure when I will be able to do this. (This upcoming semester is going to be very busy for me.) So if anyone wants to try this, the results are likely to be pretty interesting.
Back to top
View user's profile Send private message Send e-mail Visit poster's website Yahoo Messenger
Kevin Pelton
Site Admin
PostPosted: Fri Aug 12, 2005 3:39 pm Post subject: Reply with quote
Dan points out a way to do it season-to-season. We could also look game-to-game at situations when a high-possession player was out of the lineup, for example.
Hopefully, these kinds of situations can provide some solid evidence (whether it proves our theory or not). I don't wish to be overly pessimistic on the issue, but I do think that often the notion of "proof" we're left with from the lab sciences (and social sciences) is something of a burden at times.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Ben F.
PostPosted: Fri Aug 12, 2005 3:57 pm Post subject: Reply with quote
Well, then, fine. We don't have to "prove" it, but I think it's something very important to study. Ideally, if Dean is right about this, it gives a very interesting lesson on chemistry, how a group of players might function together on the offensive end.
I think it's something that we should look into more, and thus was asking perhaps for ways to look at it (if it's even there). Dan's suggesting was very interesting, and although I understood the logical reasoning, I don't know the nuts and bolts of the statistics, otherwise I'd do the work.
Back to top
View user's profile Send private message
bchaikin
Location: cleveland, ohio
PostPosted: Fri Aug 12, 2005 6:20 pm Post subject: Reply with quote
Does efficiency drop as usage rises?
we tried this once before, and one of the problems was what defines usage? touches/min? possessions as defined in BoP?....
There's really not a large sample size to be able to view what happens when role players take on possessions and stars lose them.
there's tons of data using touches/min, all you need do is look at players who play significant minutes annually and who's touches/min change alot from one season to the next...
again use iverson as an example. he had his highest touches/min in 04-05 and also one of his most efficient seasons. i just don't know if his "possessions" as defined by BoP were the highest of his career....
Back to top
View user's profile Send private message Send e-mail Visit poster's website
HoopStudies
Location: Near Philadelphia, PA
PostPosted: Sat Aug 13, 2005 11:26 am Post subject: Reply with quote
Quick comment as I'm out of the country and my girlfriend didn't allow me to bring my computer and only gives me a couple minutes per day to digest the outside world. And it is good that way. Trust me. I needed this vacation.
I think the burden of proof should be on those saying that such an effect doesn't exist. If you've played hoops, you have some sense that you can't be efficient if you try to score all the time. Or if someone really good tries to score all the time.
I haven't tried to "prove" it myself, but I've seen the effect in numerous game level studies. I haven't looked across seasons, which is more what Bob is suggesting. I have run skill curves for different player seasons and seen differences, but usually they are similar from season to season. I remember the break out Elton Brand season (first year with Clips) really bugging me because it was sooo different than other things he'd done. But really that was an anomaly that he hasn't duplicated. Most players don't show big differences.
It is tough to get at because of reasons mentioned by KP and Dan. But there are a lot of different game level studies that see it. Getting skill curves, though, is tougher and obviously something I'm keeping to myself until I can put food on the table doing this job (those were a part of BoP that proved frequently useful this past season).
Uh, gotta go have fun.
_________________
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
Dan Rosenbaum
PostPosted: Sat Aug 13, 2005 2:04 pm Post subject: Reply with quote
I think the instrumental variables approach that I am suggesting could work at the game-level or the season-level. It would be interesting to do both.
I had an idea that might be fun. There is the Journal of Quantitative Analysis in Sports that DeanO and Roland are helping start. Done right, I think this would be a nice article for that journal.
http://www.bepress.com/jqas/
I am very busy this Fall and probably only have time to serve as the organizer of this, but I was wondering who might be interested in being part of this project. We probably can't have 10 co-authors, but I suspect that even four or five would not be completely unworkable. Here is what I think needs to be done.
(1) putting together the season-level data
(2) putting together the game-level data
(3) devising the empirical strategy
(4) running the regressions
(5) putting together the tables
(6) writing up the results/writing the paper
(7) revising the paper
(8) organizing the process
(9) handling the submission
With the limited time I have this Fall, I think that I am best employed doing (3), (4), part of (5), some guidance on (6), (7), (8), and (9). If there were folks who were interested in helping with (1), (2), (5), (6), and (7), that would be great.
Let me know either through e-mail, PM, or in this thread if you might be interested. If we dozens of people interested in helping out, we may need to make some choices about whose name goes on the paper, but it would be good to get many people involved in this. (But it would look bad to have a list of co-authors longer than the article itself - although I have seen that done with medical journals.)
Back to top
View user's profile Send private message Send e-mail Visit poster's website Yahoo Messenger
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Sat Aug 20, 2005 11:25 pm Post subject: Reply with quote
What response, if any, Dan, have you gotten from this?
Back to top
View user's profile Send private message
bchaikin
PostPosted: Sun Aug 21, 2005 12:38 am Post subject: Reply with quote
I think the burden of proof should be on those saying that such an effect doesn't exist.
this is the 2nd time this arguement of "...i don't have the statistical evidence, so guess what? the burden of proof is on someone else who doesn't agree with it..." has been used concerning this supposed tenet of efficiency-falls-off-as-usage-increases. if you are going to make a statement and declare it as fact, shouldn't you then substantiate it with statistical evidence, as again this is a stats analysis group?....
i for one have seen the statistical evidence go both ways on this issue. i can show stats for players whose efficiency decreases with usage (touches/min), and for those whose efficiency increases with usage. but if one is going to claim something as fact, then it is up to that who claims it to provide the evidence for it....
i broached this topic earlier when i mentioned players like allen iverson and derek fisher (both used as examples in BoP), who had increased touches/min in 04-05 and with greater efficiency than they had in recent seasons with less touches/min, but never got a response - just a discussion of how possessions were different than touches/min, but not anything specific on iverson and fisher...
but in using touches/min there are numerous examples that can be looked at...
If you've played hoops, you have some sense that you can't be efficient if you try to score all the time. Or if someone really good tries to score all the time.
i don't doubt this for a second - but to quantify it with stats or statistical analysis is another story. statements like "...well it must be true otherwise team A would give the ball to player B all of the time..." isn't enough imho to quantify it...
This relationship between usage and efficiency is very hard to estimate empirically.
exactly - that is why we use stats, and there are plenty of those to look at historically. allen iverson and dwyane wade and dirk nowitzki and larry hughes increased their touches/min from 03-04 to 04-05 and were more efficient, damon jones and jalen rose and jamal crawford and jason kidd decreased their touches/min and were more efficient. there are multiple examples for both. i just don't know if their possessions as defined by BoP reflect this...
I think it's something that we should look into more, and thus was asking perhaps for ways to look at it (if it's even there). Dan's suggesting was very interesting, and although I understood the logical reasoning, I don't know the nuts and bolts of the statistics, otherwise I'd do the work.
again all player touches/min stats for the past 28 seasons are in the free stats database at http://www.apbr.org or http://www.bballsports.com. the evidence is there for anyone who wishes to look at it....
Back to top
View user's profile Send private message Send e-mail Visit poster's website
jkubatko
PostPosted: Sun Aug 21, 2005 1:50 pm Post subject: Reply with quote
First, let me state that this argument is somewhat similar to the argument in baseball that there is no such thing as a clutch hitter. You're going to have people on both sides who refuse to budge one bit, and at the end of the day no progress will be made.
Getting back to basketball, I believe that some players will show little or no loss in efficiency (and perhaps even a gain) with increased usage, while others will see a loss in efficiency. Determining a priori who those players are, however, is another matter.
I decided to examine this issuse by doing a small study. Using true shooting attempts (TSA = FGA + 0.44*FTA) as my measure of usage and true shooting percentage (TS% = 0.5*PTS / TSA) as my measure of efficiency, I found players who met the following crtieria:
1) Were at least 28 years old in the target season. (I thought that players in this age range would be less likely to see an improvement in TS% due to improvement of skills.)
2) Saw an increase of at least 1/3 in their TSA per 40 minutes played from the prior season to the target season.
3) Played at least 1000 minutes in the prior season and target season.
A total of 43 players met the criteria above. Here they are:
Code:
--- TSA -- ------ TSP ------
Name Year Yr-1 Yr Yr-1 Yr Diff
+----------------------+----+-----+-----+-----+-----+------+
Michael Adams 1991 16.8 28.2 54.2 53.0 -1.2
Danny Ainge 1989 14.1 19.5 59.8 55.4 -4.5
Nick Anderson 1998 14.2 19.8 49.4 52.9 3.5
Shandon Anderson 2003 10.5 14.3 48.9 55.3 6.4
John Bagley 1992 8.6 12.6 48.9 47.8 -1.1
James Bailey 1986 12.2 16.6 48.5 51.6 3.1
Henry Bibby 1979 12.0 16.0 49.2 49.3 0.1
Frank Brickowski 1991 12.5 17.1 57.6 57.8 0.2
Mark Bryant 1999 9.6 13.2 54.5 51.3 -3.2
Antoine Carr 1990 14.5 19.7 54.1 55.7 1.6
Darwin Cook 1987 13.8 18.8 46.5 46.0 -0.5
Chuck Cooper 1955 11.3 15.5 37.5 42.2 4.7
Freddie Crawford 1970 12.3 17.2 48.4 51.4 3.0
Michael Curry 2000 8.4 11.2 50.7 55.9 5.2
Mike Evans 1985 14.7 20.0 50.6 56.8 6.2
Sleepy Floyd 1991 13.8 22.6 55.2 48.0 -7.3
Chris Ford 1979 13.1 18.1 51.0 50.2 -0.7
Kendall Gill 1997 14.3 21.1 54.5 53.0 -1.6
Gail Goodrich 1972 18.9 25.5 52.1 54.8 2.7
Robert Horry 2004 9.1 12.3 48.8 49.6 0.8
Bobby Jackson 2002 13.7 19.1 50.1 53.7 3.5
Jim Jackson 2000 14.1 19.2 49.9 49.6 -0.4
Mike James 2005 13.0 17.5 52.8 53.4 0.6
Avery Johnson 2002 10.1 14.9 47.5 53.1 5.6
Caldwell Jones 1983 9.2 12.6 57.1 50.5 -6.5
Jon Koncak 1994 5.9 8.4 47.6 44.4 -3.1
Freddie Lewis 1975 15.2 20.3 47.0 54.0 7.0
Alton Lister 1989 9.4 13.8 54.1 52.9 -1.3
Danny Manning 2001 11.9 17.2 46.0 53.7 7.7
Reggie Miller 2005 11.8 15.9 60.0 58.2 -1.8
Calvin Natt 1985 15.3 22.6 63.3 60.4 -2.9
Mike Newlin 1980 15.1 24.1 56.5 54.0 -2.5
Johnny Newman 1998 12.7 18.3 54.5 54.7 0.2
Charles Oakley 1994 9.1 12.2 55.6 54.0 -1.6
Guy Rodgers 1969 12.5 17.9 41.6 43.3 1.7
Dennis Rodman 1995 4.5 7.4 54.9 60.4 5.5
Ray Scott 1971 19.6 26.7 47.9 49.6 1.6
Rory Sparrow 1989 11.8 15.8 42.3 48.6 6.3
Wes Unseld 1979 8.6 11.7 53.5 59.9 6.4
Foots Walker 1982 7.4 10.0 53.7 49.2 -4.5
Ben Wallace 2004 7.2 11.5 48.6 44.1 -4.5
Ben Warley 1968 14.9 20.1 49.6 53.4 3.8
Orlando Woolridge 1991 18.5 25.8 60.1 56.5 -3.6
Let me interpret Orlando Woolridge's line as an example. In 1990 he averaged 18.5 TSA/40, which increased to 25.8 in 19991. His TS% dropped from 60.1% in 1990 to 56.5% in 1991, a decrease of 3.6%.
Of the 43 players above, 19 showed no improvement in their TS% and 24 showed improvement in their TS%. The average change was 0.8%, with a standard deviation of 3.91%.
I'm not suggesting that the information above proves anything one way or the other. However, I do think it is interesting.
_________________
Regards,
Justin Kubatko
Basketball-Reference.com
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Mike G
PostPosted: Mon Aug 22, 2005 5:35 am Post subject: Reply with quote
Studies of consecutive player-seasons that don't incorporate some knowledge of Player X's "need to shoot" don't address the causality question: Did he shoot more because he was shooting better? or did he shoot better because he shot more?
Offhand, I'd say unless 80-90% of all major (+25%, say) increases in TSA/min are accompanied by improved TS%, the indication is that just increasing your shots tends to make you miss More of them (all else being equal). In other words, 80-90% of those TSA/min jumps would be the Result of (perceived) better shooting habits/effectiveness, rather than the Cause.
Of course, the opposite would also be true: the great majority of players whose shots and TS% both Drop is explained by their taking fewer shots Because their shot went south, and not vise versa.
Back to top
View user's profile Send private message Send e-mail
badgerbucco
Joined: 18 Jul 2005
Posts: 3
PostPosted: Mon Aug 22, 2005 7:50 pm Post subject: Burden of proof Reply with quote
Since I'm still in the process of teaching myself statistics, so I can contribute more fully to the analysis here, I'll just stick to what I do know.
I agree with Dean on the burden of proof. In a legal sense, the burden is on the person trying to prove an presumption wrong. For example, in a criminal case, the state has the burden of proof of guilt since the presumption is innocence.
In the law, the presumptions are generally based on policy, such as that we REALLY want to make sure that a person is guilty before we take away his/her liberty. Dean's presumption is based not on policy but on the conventional wisdom that efficiency declines as usage increases. Although there isn't great empirical evidence of this, I think anyone who has played the game feels that it is true. This is as good a reason as any to come up with a presumption.
Without the presumption and burden of proof, you have nothing against which to push. The burden of proof in this case is therefore similar to the scientific method since it essentially sets something up as a hypothesis which the analyst community must work to prove or disprove.
We believe that statistical analysis can disprove conventional wisdom but that fact doesn't mean conventional wisdom has no basis to be believed. Conventional wisdom is based on experience that should be given its due. For every piece of conventional wisdom that is disproven with analysis, there are 10s pieces that are shown to be true. Its track record is better than the credit it is given.
Back to top
View user's profile Send private message
marc200
PostPosted: Tue Aug 23, 2005 4:43 pm Post subject: New member...*might* be able to do this Reply with quote
Just joined...I am another economics professor who does applied micro, I actually teach undergrad metrics. This sounds like an interesting idea and I'm familiar with the methods Dan is proposing to use. I don't have that much time this semester but might be able to fit in some work on this. Especially as we move into 06, when I should free up more. I'll email Dan directly. I might also be able to find some undergrad sports fans who would be interested in working on this as a metrics project.
It seems to me using injuries to major high-usage stars as an instrumental variable would also work here. Unless people are being put on injured reserve to give them rest for unimportant games, injuries should be exogenous to the expected skills, "hot hands", or convenient matchups of role players on the team. Injuries should also result in increased usage for other players.
I actually think this could be a better IV than the one Dan proposes, as it seems to me that the usage patterns of new teammates the management brings in could well be endogenous to expected changes in usage by current players. For example, if you think someone is ready to take a step from being a role player to a more central player on the team who gets more touches, you might surround him with people who have more role-type skills. Conversely, if you think someone should be playing a more limited role than they are you might go out and search for a star-type player to replace them as the team's offensive motor. I'm not saying Dan's IV wouldn't work, just that one must be careful.
As Dan knows, applied micro people can argue about endogeniety of instruments all day (or all seminar) long, it's very boring for those outside the tribe...sorry to those who had to suffer through the post...
Last edited by marc200 on Tue Aug 23, 2005 5:01 pm; edited 1 time in total
Author Message
Dan Rosenbaum
PostPosted: Tue Aug 23, 2005 4:50 pm Post subject: Reply with quote
Hey guys, I have gotten a number of folks interested in this, so we will defininitely be going forward on this project. I have been struggling to keep my head above water with the new baby and the start of classes last week, but I am planning to send e-mail around to those who have expressed interest in this. For those of you who sent me PMs, please send me an e-mail, so that I have your e-mail address.
We will keep the board apprised of our results and so in a way this paper is likely to become an APBRmetrics project.
Best wishes,
Dan
MDC
PostPosted: Wed Aug 24, 2005 9:26 pm Post subject: Reply with quote
jkubatko wrote:
I decided to examine this issuse by doing a small study. Using true shooting attempts (TSA = FGA + 0.44*FTA) as my measure of usage and true shooting percentage (TS% = 0.5*PTS / TSA) as my measure of efficiency, I found players who met the following crtieria:
The results are interesting, but it seems to me that turnovers are a big part of the picture that are not being calculated here. Instinctively, I've felt that my turnovers rise disproportionately to my touches if I have a heavier burden in the offense. I would be surprised if this wasn't true.
_________________
_______________________________________
Nicholas Hunt
Back to top
View user's profile Send private message
NickS
PostPosted: Thu Aug 25, 2005 11:55 am Post subject: Reply with quote
I'm strongly in the camp that believes that, at some point, efficiency will fall as usage increases.
In fact I would hope that we can all agree that at some point that will be true.
If you had a team of Kevin Ollie, Fred Hoiberg, Bruce Bown, Reggie Evans, & Theo Ratliff something would have to give. Even though all those players have ORtgs above 100 I suspect that as a team they would be terrible offensively.
I think part of what makes the belief that efficiency and usage are inversely related so strong is that we all watch the game and see that some players can only score when the defense gives them an opening and others can create openings for themselves.
But I worry that the fact that we have such a strong narrative on one side of the question makes it difficult to be properly analytical so let me play Devil's advocate and offer an explanation (that I find partially convincing) for how efficiency could increase as usage increases within limits.
I know that, for myself (as a mediocre ball player), my efficiency does not necessarily increase in games where I'm not expected to create much offense. I get out of rhythm, I begin to doubt myself, I get frustrated with the fact that I haven't touched the ball on 10 trips down the floor and I force a shot when I do get the ball . . .
The games in which I shoot the best are the games in which I am not asked to do too much and in which I have the ball in my hands a reasonably amount and I feel that my teammates trust me to make decisions with the ball in my hands.
In terms of NBA players any of these psychological factors may be ato work, but consider another idea.
The theory that efficiency increases as usage decreases argues that players responsibility decreases they have the opportunity to have improved shot selection. This implies that the player sees the ball the same amount and has the opportunity to assess the situation and give up the ball when they can't get the shot that they want.
But this may not describe the typical game situation. It is at least as plausible that when a player has a smaller role on offense that they touch the ball less frequently and have less leeway to create their own shot.
[certainly this explanation was offered for why shammond williams was not a better offensive player on the Sonics. Of course he didn't improve when he went elsewhere]
It's also possible that an offensive player has more ability to make adjustments in their matchup with a defensive player than the defensive player does. If this is true a player that gets more shot attempts will have more chances to improve in their one on one matchup with the defender. We might expect to see efficiency rise with usage until the other team responds by allocating more help defense.
If any of these explanations are true it's important to realize that it doesn't imply that teams could improve just by running more of the offense through fred hoiberg (or Mike Miller). This would suggest that an offense is a dynamic function based on the role that each player takes and that you can't shift players between roles without affecting everyone else on the court.
If you believe that players offensive efficiency depends not just on their own shot selection but on their role in the offense (how many touches they get and whether those touches come in places on the floor that are well suited to their offensive skills) and that running the offense through, say, Dale Davis (oRtg > 115) might make everyone else on the team suffer because it means running the offense through places on the court that aren't well suited to the other players.
In other words even if high efficiency, low usage players might be capable of maintaining their efficiency while increasing their usage it's possible that the team offense isn't capable of giving them a role that will allow them to do that.
Back to top
View user's profile Send private message
bchaikin
PostPosted: Sun Aug 28, 2005 12:15 am Post subject: Reply with quote
I'm strongly in the camp that believes that, at some point, efficiency will fall as usage increases.
yep - the question is can we show this statistically, since this is a statistical analysis group. simply accepting this as dogma without the statistical evidence sets a poor precedence for a group claiming to be statistically oriented...
If you had a team of Kevin Ollie, Fred Hoiberg, Bruce Bown, Reggie Evans, & Theo Ratliff something would have to give. Even though all those players have ORtgs above 100 I suspect that as a team they would be terrible offensively.
last season the atlanta hawks went just 13-69. they had a per game point differential of -9.7 pts/g, which translates historically to a team that should have gone about 16-66 or 17-65. when i run a simulation of the 04-05 hawks against their actual 04-05 schedule with each player playing approximately the same total minutes they played in real life, they go on average between 19-63 to 20-62...
if i change their computer played sub pattern so that each of the hawks starters (drobjnak, a.harrington, j.smith, j.childress, t.lue, i left a.walker out) plays 36 min/g, that W-L record improves to about 21-61 to 22-60. if i replace those 5 starters with the phoenix suns starters (a.stoudamire, marion, j.johnson, q.richardson, nash) and have each play 36 min/g, the team averages a W-L record of 56-26 (the suns actually went 62-20, and the hawks bench was worse than the suns bench)...
replace that lineup with the miami heat starters (shaq, haslem, e.jones, wade, d.jones) each playing 36 min/g and the team goes 54-28 (the heat actually went 59-23). use the spurs starters (nesterovic, duncan, bowen, ginobili, t.parker) and the team goes 55.5-26.5. use the charlotte bobcats starters (brezec, okafor, g.wallace, bogans, knight) and the team goes 25-57, the new orleans hornets starters (c.andersen, p.j.brown, nailon, j.r.smith, dickau) and the team goes 25.5-56.5....
so you can see the scale of how easy it is to change the fortunes of the team based solely on who the starting players on it are...
replace the hawks starters with ratliff, evans, bowen, hoiberg, and ollie (i used ollie's 03-04 stats as he played < 200 min in 04-05) and the team goes 40-42. their off FG% as a team was .432, on defense .441. their Scoring FG% as a team was .517, on defense .515. they committed 15.0 TO/g, gave up 13.4 TO/g. they got 43.4 reb/g and allowed 42.2 reb/g. here are some of their other stats:
ratliff - 9.9 pts/g, 260 BS, 152 TO
evans - 10.9 pts/g, 1010 reb (12 reb/g), 236 TO
bowen - 14.6 pts/g, 182/443 3pters (41%), 103 TO
hoiberg - 19.0 pts/g, 65% ScFG%, 243/506 3pters (48%), 101 TO
ollie - 12.3 pts/g, 630 ast, 220 TO
here is how their touches/min changed (real life compared to the simulation), in parenthesis are their career high touches/min and the year that occurred:
real-simulated
0.35-0.58 ratliff (0.62-200001)
0.49-0.75 evans (0.49-200405)
0.57-0.93 bowen (0.70-199798)
0.70-1.13 hoiberg (0.98-200001)
1.34-1.70 ollie (1.34-200304)
so the question is, could these players realistically play as they have in real life with the increase in touches/min?...
ratliff has had higher touches/min in a single season than in this simulation so obviously he could. evans' touches/min were 35% higher than what he has ever done, but he's only been in the league for 3 seasons (and chris dudley has had a change of 30% in his touches/min between consecutive seasons where he played 1400 minutes or more in each). bowen's is 25% higher than his highest touches/min (juwan howard had a 25% increase from his 1st to his 2nd season playing 2000+ min in each season), and hoiberg's is 13% higher than its ever been....
ollie's is 21% higher than his career best. ollie got 193 assists playing 1146 minutes with the bulls in 01-02. that works out to about 500 assists playing 36 min/g and 82 games. i'm guessing that most PGs that can get 500 assists as a starter can also get 630 as a starter in perhaps a different situation. andre miller got 882 assists playing for the 01-02 cavs, and has never gotten anywhere near that again despite playing similar minutes on other teams (only one other time has he gotten more than 600 assists). he got those high assists because he played on a team with a couple of starters that had much lower touches/min than the average player for that position that season, and thus he got alot of the remaining touches...
so my question is is there any statistical evidence that shows that these players could not play with an increase in touches/min at the same efficiency as they did at their career best/high touches/min?....
also for fun i ran this same sub pattern but with tim duncan at PF rather than reggie evans for 36 min/g. the team's W-L record improved to 52-30 as duncan scored 28.5 pts/g and got 1.45 touches/min. i can just hear the smirks and snickers now - duncan has never come close to these numbers (his career best pts/g is 25.5 pts/g and career highest touches/min 1.27), and that he could never handle the ball this much nor score this much consistently (that would be a 13% increase in touches/min over his career best)....
but guess what? looking at the historical stats there is precedence for just this kind of situation, and coincidentally from the same team, the spurs, no less...
in 1991-92 and 1992-93 david robinson scored just 23-24 pts/g in each season (his 3rd and 4th seasons in the league, he had scored 25.6 pts/g in his 2nd season). his touches per minute were 1.01 and 1.14 in the two seasons. then in 93-94 dennis rodman is traded to the spurs, and he plays almost every game and 38 min/g at PF but with just 0.47 touches/min, the lowest touches/min of any player in the league that season to play at least 2000 minutes....
on top of this dale ellis was also on that team at SF and he got just 0.6 touches/min (he was mostly a spot-up shooter all season, shot with almost 2/3 of his touches, and 1/3 of all his shots were 3pt attempts). that was the lowest touches/min of any SF in the league that season to play at least 2000 minutes, and he played in almost every game and 34 min/g....
so here's a team where two starters had the lowest touches/min in the league at their respective positions - on the same team. willie anderson and vinny del negro were the starting guards (neither got more than 1.25 touches/min), and thus there were alot of touches left over - and robinson got alot of them. he averaged 1.40 touches/min and scored 30 pts/g....
so here is a real world situation where a player had scored just 23/24 pts/g in two previous seasons, but because two players who were starters alongside him had very low touches/min, much lower than the average PF and SF that season, robinson ends up getting the bulk of the remaining touches. that 1.40 touches/min of his was 18% more than his previous highest touches/min, and 25%-28% higher than his 3 other seasons in the league...
i wonder how many people in this discussion group would have said before the 93-94 season that based on the statistical evidence of robinson's first 4 seasons in the league that he'd never be a scoring champ, nor score anywhere near 30 pts/g over the span of a whole season, nor get more touches/min at C than any center in 16 years. but he did - because of who his teammates were...
so i don't see much difference between robinson getting 1.40 touches/min and 30 pts/g in real life and a simulated duncan getting 1.45 touches/min and 28.5 pts/g (another thread in this discussion group has shown how statistically similar the careers of these two players were when looking at robinson's first 6 seasons in the league and duncan's best 6 year sequence). it all depends on who your teammates are...
The theory that efficiency increases as usage decreases argues that (as) players responsibility decreases they have the opportunity to have improved shot selection. This implies that the player sees the ball the same amount and has the opportunity to assess the situation and give up the ball when they can't get the shot that they want. But this may not describe the typical game situation. It is at least as plausible that when a player has a smaller role on offense that they touch the ball less frequently and have less leeway to create their own shot.
exactly....
Back to top
View user's profile Send private message Send e-mail Visit poster's website
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
PostPosted: Sun Aug 28, 2005 9:17 am Post subject: Reply with quote
Still not home, but a couple things:
- The way I've built skill curves is within a season. If you go across seasons, you have to then account for both teammates (the Ron Harper impact) and player growth (are they at the prime of their career or not, are they likely to change in general). Possible, but painful. I do it in detailed analysis, but not for surveys.
- At one point in the past, I did something along the lines of what Ed posted. I didn't constrain things as much as he had (I looked over a career, not just big changes), but I used offensive ratings adjusted for league avg, not just shooting %. I looked for how many players showed inverse relationship between usage and offensive ratings through their careers. I can try to dig this up when I get home, but I know that most players showed the inverse relationship (or a negative correlation).
- Some guys who come to mind who have very well replicated skill curves as they've changed responsibility level: Antoine Walker to Dallas and to Atlanta, even Allen Iverson this year when you account for the league change in efficiency (his skill curve suggests he doesn't change much anyway), Nick Van Exel frequently in his career, Ben Wallace, Ron Harper. There were a lot more, but I just don't remember.
- Something I have never done but realize I can do is get a skill curve for David Robinson prior to that '93-'94 season. As BoP showed, some players can ramp up their performance without major hits on efficiency. Others cannot. It is possible that Robinson's skill curve suggested the ramp up capability. He was at the prime of his career, but that season still sticks out as a very good year for him.
- implying that all these high percentage low usage shooters can ramp up their usage without penalty implies that the people running the NBA are not just a little wrong. It implies also that the fundamental nature of basketball is poorly understood. It implies that any sort of linear weights rating is wrong -- you should normalize a player's shots to 20% (or something) and turnovers (to something else) then put it together with rebounds, etc. It implies that pretty much every rating method is wrong because the context in which players are being used is incorrect. Dan's method which is totally different from others here has to be wrong because it is flawed by the decision to not let Fred Hoiberg shoot 25x per game. This is not just a matter of a tiny little assumption that has to be proven. This is a principle that really underlies the game of basketball. It very much distinguishes it from baseball, where players take turns being on offense. That's why OBP and SLG work in baseball -- players can only hit once per 9 players. If hoops is like baseball, let's just define the percentages that matter because usage clearly doesn't.
- Ultimately this is a principle I have used to make projections. It is just one of many tools to do that. How do we explain Hoiberg's extreme variation in TS%? Or actually predict it? The base assumption is to predict it based on what they did last year or in their career. Duh. But how do you improve that prediction? That's (in part) what this principle is designed to help with. (It is too bad that Hoiberg's health will keep him out this year and perhaps forever, not only for the obvious reason that he's a person with really bad luck, but because he has provided a lot of material for this group. Get well, Fred.)
_________________
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
Ben
Joined: 13 Jan 2005
Posts: 266
Location: Iowa City
PostPosted: Sun Aug 28, 2005 1:26 pm Post subject: Reply with quote
I can't provide "statistical proof", but I would venture to say that a team with Bruce Bowen as its second leading scorer would be unlikely to approach .500 ball.
Back to top
View user's profile Send private message
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Mon Aug 29, 2005 1:20 pm Post subject: Reply with quote
Had a long conversation online with Bob about this topic, thought I'd post it, see what response are. Perhaps it can illuminate some thinking of some people.
I suppose what this argument comes down to, in the end, is that Bob does not disagree with everyone on the premise - he just says that you can't really control players' usages. Sure, we could say, if Hoiberg gambled away all his money, and a bunch of mobsters held a gun to his head and told him he had to get 25 shots per game or they'd whack him, his efficiency would probably decrease. But short of that scenario, Hoiberg ain't taking 25 shots. We say, if he did, he be less efficient. Bob says, he never could, it's just not in his skill set.
-------------------------------------------------------------------------------
FFS (12:29:03 PM): out of curiousity, would you say that if you had fred hoiberg on your team, and were coaching, you'd direct players to let him have the ball more?
BobC (12:29:22 PM): no this is not the point, but this is...
BobC (12:29:30 PM): i ran a sim with the players someone else suggested
BobC (12:29:37 PM): all were low touches/min players
BobC (12:29:53 PM): so the question is what happens when you put then all together...
BobC (12:30:24 PM): well, of course, someone has to handle the ball, so each has their touches/min increase somewhat, and that increase depends on how often they pass the ball
FFS (12:30:36 PM): but your sim is not reality - that is what we're arguing about! the sim assumes efficiency does NOT decrease as usage rises
FFS (12:30:51 PM): you can't use that as evidence
BobC (12:31:07 PM): why do you say the sim is not reality?
FFS (12:31:34 PM): its awful close, I think, but its not the same as analyzing what actually happens
BobC (12:31:38 PM): did you read the david robinson example?
FFS (12:31:53 PM): david robinson never was a low usage player
BobC (12:31:57 PM): ah ye of little faith
FFS (12:31:58 PM): sure he increased it to his highest
FFS (12:32:05 PM): of his career
BobC (12:32:12 PM): but?
FFS (12:32:19 PM): but that's not saying having fred hoiberg do it would have the same results
FFS (12:32:22 PM): or kevin ollie
FFS (12:32:29 PM): or steven hunter
BobC (12:32:32 PM): i can give you multiple example
FFS (12:32:39 PM): the fact is
FFS (12:32:49 PM): I don't doubt the sim in normal situations
FFS (12:32:56 PM): but when it gets to the extremes
FFS (12:33:05 PM): like 5 low usage players, or 5 high usage players
FFS (12:33:09 PM): I can't assume it is truth
BobC (12:33:12 PM): anything that happened in real life IS a normal situation
BobC (12:33:18 PM): right?
FFS (12:33:23 PM): right - those extremes haven't happened
BobC (12:33:28 PM): guess again
BobC (12:33:33 PM): take the robinson example
FFS (12:33:45 PM): robinson was not 5 low usage players
BobC (12:33:58 PM): wait a sec
BobC (12:34:18 PM): robinson had traded to his team the league's lowest touches/min PF in rodman
FFS (12:34:30 PM): I read your post
BobC (12:34:34 PM): then the spurs also had dale ellis who was the lowest touches/min SF that year
FFS (12:34:37 PM): he had 4 low usage players around him
BobC (12:34:42 PM): right...
BobC (12:34:51 PM): so how is this not real?
FFS (12:34:52 PM): so his usage increased 6
FFS (12:34:57 PM): but that was from 26 alread!
FFS (12:35:14 PM): he averaged a usage of 25, which is star level, for his first 4 years
BobC (12:35:46 PM): wait a minute
BobC (12:36:05 PM): robinson's touches/min that season was the highest ever for a C in 16 years
BobC (12:36:16 PM): are you saying that's not unusual?
BobC (12:36:26 PM): that it was in fact predictable?
FFS (12:36:27 PM): I'm saying that a player like Robinson
FFS (12:36:47 PM): is much less likely to exhibit "normal" behavior
FFS (12:36:52 PM): since he was never a "normal" player
FFS (12:36:57 PM): in fact he was a star
FFS (12:37:02 PM): which by definition is an anomaly
FFS (12:37:11 PM): so we can't base our asssumptions off of him
FFS (12:37:24 PM): if you took someone like Dennis Rodman
FFS (12:37:29 PM): and he started doing that
FFS (12:37:38 PM): and his efficiency stayed relatively level
FFS (12:37:41 PM): THEN that's something
BobC (12:37:54 PM): and rodman started doing what?
FFS (12:38:36 PM): jumped his usage up 20%
FFS (12:39:37 PM): his first couple years in the NBA were his highest usages
FFS (12:39:44 PM): but we can't judge from that, since he was a rookie
BobC (12:39:46 PM): between 87-88 and 88-89 his touches/min dropped by 26%, this is not the same thing?
FFS (12:39:57 PM): thats what I was saying
FFS (12:40:02 PM): they dropped drastically
FFS (12:40:05 PM): but at the same time
FFS (12:40:24 PM): his initial inefficiency could be attributed to his relatively low experience
BobC (12:41:14 PM): well tell me this, between 84-85 and 85-86 sleepy floyd's touches/min increased by 26% in one season
BobC (12:41:25 PM): and he played major minutes both seasons
BobC (12:41:35 PM): this jumps are NOT unusual
BobC (12:41:51 PM): there are many many examples
FFS (12:42:07 PM): but I'm talking about a jump in usage from somewhere in the low teens to perhaps the 20s
FFS (12:42:21 PM): like a Fred Hoiberg/Bruce Bowen getting the touches of a star
BobC (12:42:40 PM): now what a minute, did you look at the sim results of hoiberg? his scoring went to 19 pts/g
FFS (12:42:45 PM): floyd's usage in '84 was 23, his usage in '85 was 27
FFS (12:42:59 PM): That's what I'm saying about the sim! I don't believe that would happen in real life
BobC (12:43:05 PM): what does that 23/27 number mean?
FFS (12:43:28 PM): Usage rate (available since the 1978 season in NBA); the formula is 40 * (Lg Pace / Tm Pace) * ((FGA + 0.44*FTA + 0.33*AST + TO) / MP). Usage rate estimates the number of possessions a player uses per 40 minutes played. (B-R.com)
BobC (12:44:08 PM): so are you saying this 23/27 is no big jump?
FFS (12:44:26 PM): I'm saying it's a pretty big jump - but the fact is 23 is a pretty good player anyway
FFS (12:44:44 PM): I'm saying that it is easier for a star
FFS (12:44:50 PM): to increase his usage and maintain efficiency
FFS (12:44:52 PM): than a role player
FFS (12:44:58 PM): which is the point of the discussion!
BobC (12:45:08 PM): well then what defines a star? is any player with say touches/min > 1.0 a star?
FFS (12:45:18 PM): this would help be a reason why Iverson is actually more valuable than he looks
FFS (12:45:26 PM): he can maintain his average efficiency with much higher usage
FFS (12:45:36 PM): allowing his teammates to operate at their optimum efficiency
BobC (12:45:53 PM): you are buying into deano's stuff without looking at all the evidence
FFS (12:46:16 PM): I think it makes sense from a logical standpoint
BobC (12:47:12 PM): because in 04-05 iverson had by far his mkost touches/min yet was more efficient than in any of his previous 5-6 seasons
FFS (12:47:27 PM): like I said
FFS (12:47:31 PM): superstars can do that easier
FFS (12:47:41 PM): show me a player like hoiberg
FFS (12:47:45 PM): who can drastically increase his usage
FFS (12:47:48 PM): and see no fall off
FFS (12:47:56 PM): hoiberg was assisted on 96% of his baskets last year
FFS (12:47:59 PM): 96%!
FFS (12:48:06 PM): and you're telling me that he could put a team on his back?
FFS (12:48:10 PM): be the leading scorer
FFS (12:48:18 PM): get open when he's double teamed?
FFS (12:48:44 PM): it seems to me his sole function was to drain shots because he was open off teammates' double teams
FFS (12:48:52 PM): thus if you were to take away high usage players
FFS (12:48:54 PM): such as garnett
FFS (12:48:56 PM): cassell
FFS (12:48:57 PM): etc
BobC (12:49:00 PM): he scored 19 pts/g om 12-13 shots a game
BobC (12:49:04 PM): in the sim
BobC (12:49:19 PM): are you telling me he could not get off 12/13 shots playing 36 min/g??
BobC (12:49:36 PM): the avg SG takes 16 FGA per 36 min?
FFS (12:50:00 PM): his highest rate EVER was 11 FGA per 40 min
BobC (12:50:38 PM): so are you saying is a player got a chance to start he couldn't go from 11 FGA in 40 min to 12/13 in 36min??
FFS (12:51:19 PM): and you're saying he got about 1.5 points per shot?
BobC (12:51:29 PM): yep
BobC (12:51:35 PM): that's what he did in real life
FFS (12:51:54 PM): 1.3 in real life, with 96% of his shots assisted
FFS (12:52:29 PM): do you honestly believe that if fred hoiberg were the best scorer on a team, that he was forced to create open looks for himself, that he'd still shoot 49%?
BobC (12:52:48 PM): again, you said it - not me - that he has taken 11 FGA in 40 min, is it an unrealistic jump to go to 12/13 FGA in 36 min?
FFS (12:52:58 PM): you don't think defenses would say "hmm, he seems to only be able to make a shot off a pass when he's open, lets just not leave him open"?
BobC (12:53:03 PM): do you honestly have any evidence showing it to be otherwise?
BobC (12:53:17 PM): he's only taking 12/13 shots a game
BobC (12:53:27 PM): thats 1 shot every 3 min he plays
FFS (12:54:27 PM): I'm not saying there's an exact correlation, nor am I saying that in every case as usage rises efficiency falls
FFS (12:54:41 PM): I'm saying that in the case of a role player, with low usage, who picks and chooses his shots
FFS (12:55:01 PM): that if you took him and made him the best player on a team, he wouldn't maintain his efficiency
BobC (12:55:05 PM): what is with this role player stuff? what defines hoiberg as a role player?
FFS (12:55:23 PM): In my mind, Hoiberg is a role player because the offense does not run through him
FFS (12:55:33 PM): his role is to wait for a kickout, not be the center of the offense
FFS (12:55:37 PM): he is not keyed on by the defense
BobC (12:56:07 PM): oh right, you're going to sit here and tell me a guy with a 65% ScFG% is always left open??
BobC (12:56:27 PM): that he only ever shoots when he is open and never when guarded?
FFS (12:56:47 PM): not that he's never guarded, just that he's not the first priority of the defense
BobC (12:56:48 PM): that because 96% of his baskets are assisted on that they are EASY shots?
FFS (12:57:10 PM): I would say easier than most shots
FFS (12:57:15 PM): because they're set up by someone else
FFS (12:57:22 PM): hoiberg got open, agreed
FFS (12:57:44 PM): but the reason 96% of his shots are assisted, is that 96% of the shots were a result of someone else giving the kickout
FFS (12:57:52 PM): he didn't create for himself!
BobC (12:58:42 PM): are you under some kind of assumption that an assist is only granted when a player receives a pass and immediately shoots and score without dribbling or moving?
BobC (12:59:11 PM): that all of hoiberg's FGA come from him being left alone and standing still and getting a pass and then hitting an unguarded shot?
FFS (12:59:24 PM): he may have come off a screen
FFS (12:59:32 PM): but the fact is, it's catch then shoot
FFS (12:59:37 PM): he doesn't make a move afterwards
BobC (12:59:38 PM): ah...
BobC (12:59:54 PM): so the t-wolves have SO many offensive weapons that hoiberg is always left alone?
FFS (1:00:09 PM): its not the quantity, its the quality
BobC (1:00:10 PM): tell me, other than garnett, who on that team is so special that
FFS (1:00:17 PM): Garnett is pretty damn good
BobC (1:00:27 PM): hoiberg's man will always leave open a 65% ScFG% shooter?
BobC (1:01:01 PM): well then why doesn't every team with a player like garnett have a 65% ScFG% shooter on their team if its that easy?
FFS (1:01:03 PM): let's look at Steven Hunter
FFS (1:01:30 PM): Hunter has a usage of 12
FFS (1:01:35 PM): shot 61.4% this season
FFS (1:01:43 PM): if you stuck him in Amare's role
FFS (1:01:46 PM): in the phoenix offense
FFS (1:01:50 PM): would he still shoot as well?
FFS (1:01:55 PM): in fact, why didn't they do it?
BobC (1:02:21 PM): ah, beacuse they didn't do it means that hunter can't play? did you by chance just see the contract he got?
FFS (1:02:21 PM): Stoudemire's a tad more efficient
BobC (1:02:47 PM): stoudamire has twice the touches/min of hunter...
FFS (1:02:48 PM): would he have been able to play Amare's role with the same efficiency?
FFS (1:02:51 PM): exactly
FFS (1:03:06 PM): according to you, he could ramp up that usage and still maintain his efficiency
BobC (1:03:23 PM): or really? when did i say this?
FFS (1:03:36 PM): thats what we're arguing about
BobC (1:03:41 PM): when did i say hunter can score like amare?
BobC (1:03:53 PM): no no you just said according to you
FFS (1:04:20 PM): why couldn't hunter ramp up his usage
BobC (1:04:25 PM): when did i say you replace amare with hunter and hunter will score like amare?
BobC (1:05:01 PM): when did i ever say hoiberg can take 25 shots a game?
FFS (1:05:26 PM): see if you can put together a team in the sim where hunter gets plenty of touches
FFS (1:05:32 PM): and see how he does
BobC (1:05:36 PM): define plenty?
BobC (1:05:46 PM): his touches/min now is just 0.50
BobC (1:05:58 PM): who said that could be doubles??
FFS (1:06:39 PM): well here's what I don't understand - why is touch/min a property of a player?
BobC (1:07:00 PM): why is FGA a property of a player
BobC (1:07:05 PM): its just a number
FFS (1:07:11 PM): but I would say FG% is more a property of the player, defined by them
FFS (1:07:24 PM): you act as if hunter couldn't double his touch/min
BobC (1:07:36 PM): it depends on who his teammates are
FFS (1:07:45 PM): why? doesn't it depend more on the coach?
BobC (1:07:56 PM): absolutely not...
FFS (1:07:58 PM): if Amare gets injured for the entire year next year
FFS (1:08:04 PM): and D'Antoni says
FFS (1:08:18 PM): you know what, proceed as normal, just stick Hunter in the offense where amare is
FFS (1:08:25 PM): pretend he's amare
BobC (1:09:37 PM): this idea that a coach can increase a player's touches/min by simply running plays his way is not how nba basketball works
FFS (1:09:38 PM): so just as I couldn't say to Hunter "go shoot 90%", I couldn't also say to the team "go pass him the ball more"?
BobC (1:10:25 PM): i don't know what YOU could do, but the reality of the nba is that if you look at the historical stats, player tuches/min is defeined by the player himself and the touches/min of his teammates
BobC (1:10:59 PM): touches/min are a function of a players ability or desire to get the ball
BobC (1:11:07 PM): not how a coach tells them to play
BobC (1:11:27 PM): a coach can't tell a player what FG% to shoot can he?
BobC (1:11:39 PM): its a function of the player
FFS (1:11:45 PM): thats what I was trying to clarify
FFS (1:11:54 PM): why you thought of touch/min as a property of the player
FFS (1:11:57 PM): just like FG%
BobC (1:12:23 PM): this idea of some in the discussion group thinking you can increase a players touches by simply running plays for that person unmasks that persons misunderstanding of the nba game
FFS (1:14:54 PM): well, that changes things
BobC (1:15:05 PM): what changes?
FFS (1:15:06 PM): my main argument I suppose was based on the fact that you could change touch/min
Back to top
View user's profile Send private message
Dan Rosenbaum
Joined: 03 Jan 2005
Posts: 541
Location: Greensboro, North Carolina
PostPosted: Mon Aug 29, 2005 2:20 pm Post subject: Reply with quote
What's interesting is that in the Bob's sim, we have actual evidence on how the top two leading scorers might do with increased touches.
Fred Hoiberg had a possession usage of 12.1% last season and scored 13.7 points per 40 minutes. In his highest possession usage season (16.5% in 1999-2000), he scored just 13.2 points per 40 minutes. His true shooting percentage was down, and his turnover rate was much higher with the greater possession usage.
Bruce Bowen had a possession usage of 11.6% last season and scored 10.3 points per 40 minutes. In his highest possession usage season (12.8% in 1997-1998), he scored just 10.4 points per 40 minutes. His true shooting percentage was down, and his turnover rate was higher with the greater possession usage.
So when we look back at their careers, the top two leading scorers on this simulated team did not increase their scoring when their usage increased. If historical patterns persisted, then that would take away more than 10 points per game away from the team that Bob simulated. Instead of a .500 team, it would be one of the worst in the NBA.
And actually, we probably picked the wrong example to maximize Hoiberg's exploits. If we wanted him to win the MVP award, this is probably the team we would put him with.
PG: Snow (36), McKie (12)
SG: McKie (24), Buckner (24)
SF: Hoiberg (36), Ryan Bowen (12)
PF: Ruffin (36), Madsen (12)
C: Ratliff (36), Ervin Johnson (12)
My guess is that this lineup in Bob's sim will win greater than 50 games and Hoiberg will score more than 25 points per game. Atlanta's bench was wasting too many touches in the previous simulation. Stocked with low usage players in the starting lineup and bench, Hoiberg would be set free to be a dominant NBA player.
Usage vs. Efficiency (Ben F., 2005)
Usage vs. Efficiency (Ben F., 2005)
Last edited by Crow on Wed May 11, 2011 5:35 pm, edited 1 time in total.
Re: Usage vs. Efficiency
NickS
Joined: 30 Dec 2004
Posts: 384
PostPosted: Mon Aug 29, 2005 3:10 pm Post subject: Reply with quote
ffs,
thanks for posting that exchange, I appreciate seeing the elaboration by bob on what he is and isn't claiming.
I haven't finished reading the whole exchange but so far it's very interesting.
Back to top
View user's profile Send private message
marc200
Joined: 23 Aug 2005
Posts: 14
Location: 40 miles north of Lebron James' birthplace
PostPosted: Mon Aug 29, 2005 4:03 pm Post subject: Reply with quote
Bob seemed a tiny bit obtuse to me in that exchange with FFS. The point isn't that the COACH changes touches per minute by calling plays. It's that the team is constructed with everyone playing roles. That happens in a million ways from the moment the general manager chooses the players, and is probably reinforced informally all the time through everyone's assumptions about who plays what role. The whole Minnesota team knows to look for Garnett as a first offensive option, and Garnett himself knows he bears that responsibility. The team is not charging down the court on every possession thinking "let's set up the offense...OK, where's Hoiberg...gotta get the ball in his hands". The defense keys on Garnett as well. Do we have any idea how Hoiberg will perform with an entire defense keying on him?
And it's pretty funny to see Kevin Ollie referred to hear as a high efficiency player (is he statistically? Couldn't quite get that straight?). I had the misfortune of watching Mr. Ollie during what must have been one of his highest usage seasons. This was when he played for the Cavs during Lebron's rookie year. The offense went through him often as the PG and believe me it was not pleasant to watch. He was helpless out there. Not because he made stupid decisions, but because he couldn't create at all.
_________________
Marcus Stanley
Back to top
View user's profile Send private message
NickS
Joined: 30 Dec 2004
Posts: 384
PostPosted: Mon Aug 29, 2005 10:35 pm Post subject: Reply with quote
An anecdotal piece of evidence in favor of the null hypothesis (players do not lose offensive efficiency when they have to take on more responsibility to create offense).
Look at the 2003 Denver nuggets, a team that was desperately lacking in players with the ability to create offense. The consensus view about that team was that they played hard on deffense but their offense was terrible because they lacked scorers.
Their offense was terrible, but the fact that they had a number of low volume scorers may not have made it any worse.
Look at all of the players from that nuggets team and see if any of them improved their offensive rating when they had better scorers around them.
Most of them don't. Nene improved slightly but he was a rookie in 2003 so his improvement could be explained for other reasons. James Posey had a temendous improvement in Ortg in 2004 which is so large as to be difficult to explain by any measure. Rodney White had a slight jump but there's no sign that, as a group, they played bettwer when they were playing with better scorers.
(note, however, that Juwan Howard's was much better on dallas the year before he stayed at his 2003 level after he went to Orlando but Orlando was also a terrible team.)
There are enough that improve that it isn't a strong piece of evidence but I find it interesting.
Back to top
View user's profile Send private message
Kevin Pelton
Site Admin
Joined: 30 Dec 2004
Posts: 979
Location: Seattle
PostPosted: Mon Aug 29, 2005 10:53 pm Post subject: Reply with quote
That discussion only further convinces me that a big reason this debate is as intractable as it has been is that there are two very different prisms through which the game of basketball is being viewed.
Bob is speaking about touches. Everyone else is speaking about possessions. These are similar concepts, but certainly not the same.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
bchaikin
Joined: 27 Jan 2005
Posts: 689
Location: cleveland, ohio
PostPosted: Tue Aug 30, 2005 2:46 am Post subject: Reply with quote
What's interesting is that in the Bob's sim, we have actual evidence on how the top two leading scorers might do with increased touches.
what's interesting with the simulation is that you have actual statistical evidence for any simulation results - by simply looking at players' real life numbers for comparison...
Fred Hoiberg had a possession usage of 12.1% last season and scored 13.7 points per 40 minutes. In his highest possession usage season (16.5% in 1999-2000), he scored just 13.2 points per 40 minutes. His true shooting percentage was down, and his turnover rate was much higher with the greater possession usage.
so, his "...true shooting percentage was down, and his turnover rate was much higher, with the greater possession usage...". factual statements, but quite misleading as they don't tell the true story about fred hoiberg...
1st off in 04-05 hoiberg's turnover rate was just 2%, meaning just 2 turnovers per 100 touches, which is exceptionally low, one of the lowest marks ever for a player. in fact it is so low that it is the 2nd lowest turnovers/touch mark for a player since 1977-78 for all players playing at least 1000 minutes in a season (that's out of 6023 player seasons with 1000+ minutes played since 1977-78 ). so pretty much any amount of "normal" turnovers will be significantly higher than his 04-05 rate...
2nd hoiberg only played 845 minutes in 99-00, and yes his turnover rate was much higher than in 04-05. but it certainly wasn't high by any standard - i have it at 5% or 5 turnovers per 100 touches, which is about average for an SG. but those 845 minutes was just the 5th highest amount of minutes he ever played in a single season. why not look at the seasons he played his most minutes? wouldn't that be of more significance statistically speaking? why skew this discussion by looking at just a small sample size of data?...
if you look at the 4 seasons where he played his most minutes, his rate of turnovers was just 3%, 3%, 4%, and 2% - these are very low. as a matter of fact that rate of 5% turnovers in 99-00 was the highest rate of turnovers for hoiberg in all 8 seasons where he played at least 500 minutes. so you just happened to choose the season where he had his worst rate of turnovers, and it came in a season where he played all of 845 minutes...
and yes his shooting was down that season, but again you chose a season out of the 8 that he played at least 500 minutes that was his 2nd worst shooting season. in the 4 seasons where he played his most minutes his Scoring FG%s were 59%, 51%, 60%, and 65%, for an average of 59% (weighted to total minutes). so when he has gotten the playing time he has been an excellent shooter, one of the best in the league, and in the two seasons where he played his most minutes, he shot a ScFG% of 59% and 60%, and one of those two seasons he had his highest touches/min and the other his lowest touches/min. so his touches/min, whether his career highest or lowest, certainly hasn't affect his excellent shooting...
so to imply that his shooting percentage was "down" by choosing a season where he played less than 1000 minutes, when he played 4 other seasons of 1000+ minutes, doesn't truly reflect hoiberg's shooting ability. he played 10 seasons in the league, and played 2/3 of his career minutes in just 4 of those seasons, and shot a Scoring FG% of 59%. the other 1/3 of his career minutes (3200 total) were stretched out over 6 seasons, and he shot a Scoring FG% of just 52% over that range. so if anything this is a player who shoots much better when given the playing time, regardless of touches/min. his touches/min were at 0.95 in 99-00, but at 0.98 in 00-01...
hoiberg played his most minutes in a season (2247) in 00-01 with the bulls. that season he also had his highest touches/min mark of 0.98, which is about average for an SG. that season his Scoring FG% was 59%, the 3rd highest mark in the league that season for an SG (behind brent barry's 65% and ray allen's 60%). in his season with his 2nd most minutes (03-04) he had a Scoring FG% of 60% with just 0.65 touches/min...
hoiberg could have had his touches/min doubled over his 2nd highest mark that 99-00 season you mentioned, but - tell me, from the point of view of a professional statistician - how statistically significant is it to imply hoiberg shot worse with high touches/min if it came in just 845 minutes, when he played significantly more minutes in 4 other seasons with much better shooting and lower turnovers and just as high touches/min in one of those seasons playing over twice as many minutes?...
also in 00-01 hoiberg and brent barry had virtually identical numbers. both shot extremely well, both had close to 1.0 touches/min (barry 1.05, hoiberg 0.98 ), both shot the ball with 1/5 of all their touches, and both passed the ball with 2/3 of their touches. they even got fouled as often as one another per touch, and got simlar amounts of steals and blocks. yet barry went on to have a season just 3 years later with 30% more touches than in 00-01 with no loss in shooting efficiency, and hoiberg 3 years later got about 30% less touches than in 00-01 and no loss in shooting efficiency. why should one increase and the other decrease?...
Bruce Bowen had a possession usage of 11.6% last season and scored 10.3 points per 40 minutes. In his highest possession usage season (12.8% in 1997-1998), he scored just 10.4 points per 40 minutes. His true shooting percentage was down, and his turnover rate was higher with the greater possession usage.
bowen's touches/min have ranged from 0.50 to 0.70 over his career, looking at the seasons where he's played at least 1000 minutes, and yes in his case he shot best with fewer touches. but there are players who like hoiberg shoot very well with their highest and lowest touches/min, and some who shoot better with their higher touches/min...
So when we look back at their careers, the top two leading scorers on this simulated team did not increase their scoring when their usage increased. If historical patterns persisted, then that would take away more than 10 points per game away from the team that Bob simulated. Instead of a .500 team, it would be one of the worst in the NBA.
again your statement is in part true. but it fails to take into account what happens when you place certain players into situations different from what they have experienced, but that others have, and dismisses all those situations where the evidence does show players touches/min changing - sometimes significantly - based on who their teammates are (such as when they are traded, signed as free agents, etc)...
for example, how would the 04-05 performances of the t-wolves players have changed had reggie evans of the sonics (probably the lowest touches/min of any starting PF in the league last season) played 40 min/g at PF instead of kevin garnett for minnesota? garnett got 1.6 touches/min in 04-05, over 3 times what reggie evans got in 04-05. forget an entire team of low touches/min players, how about the change of just one player - a low touches/min starter playing 40 min/g replacing a very high touches/min player, especially a frontcourt player - and the other t-wolves players playing the same minutes that they did in 04-05?...
how would you approach this situation? is it your opinion that players like fred hoiberg are simply incapable of handling the ball more often than they have in real life, simply because they have never been in a situation in the nba that has not called upon for them to do so? to me this sounds vaguely familiar to the one time arguement that players who have only played say 15-20 min/g over their career are not capable of playing 36-40 min/g when called upon to do so. but there are many examples of just this occuring (like expansion teams), and without a decrease in their abilities, just as there are many examples of players who have experienced significant increases or decreases in their touches/min based on who their teammates are without significant changes in their efficiency...
when i run a simulation of the 04-05 t-wolves such that their players end up playing close to their total minutes actually played in 04-05, the sub pattern looks like this:
min/g - player - (touches/min)
16 olowokandi (0.50)
40 garnett (1.51)
32 szczerbiak (0.90)
32 sprewell (0.97)
20 hudson (1.41)
8 madsen (0.43)
20 griffin (0.68 )
24 hassell (0.67)
16 hoiberg (0.69)
20 cassell (1.78 )
4 e.johnson
8 a.carter
a simulation of 8200 games against the wolves 04-05 schedule several times, and averaging the results, shows the team with an average W-L record of 44.5-37.5. they went 44-38 in real life, and had the per game point differential of a team that should have gone 45-37...
if i switch the minutes of hoiberg and sprewell, and give hoiberg 36 min/g and sprewell just 12 min/g:
min/g - player - (touches/min)
16 olowokandi (0.53)
40 garnett (1.57)
32 szczerbiak (0.92)
12 sprewell (0.94)
20 hudson (1.55)
8 madsen (0.46)
20 griffin (0.71)
24 hassell (0.73)
36 hoiberg (0.75)
20 cassell (1.82)
4 e.johnson
8 a.carter
and run the simulation, they average a W-L record of 49-33. hoiberg averages 8.5 FGA/g, and 2.7 FTA/g. his touches/min increase to 0.75...
then replacing garnett with reggie evans:
min/g - player - (touches/min)
16 olowokandi (0.65)
40 r.evans (0.59)
32 szczerbiak (1.14)
12 sprewell (1.07)
20 hudson (1.74)
8 madsen (0.51)
20 griffin (0.79)
24 hassell (0.85)
36 hoiberg (0.89)
20 cassell (1.92)
4 e.johnson
8 a.carter
the team drops to an average a 41.5-40.5 W-L record playing against the wolves 04-05 schedule. but as you can see everyones touches/min increased with evans replacing garnett. szczerbiak increased 19%, olowokandi 18%, hoiberg 16%, hudson 11%, etc. hoiberg took 10.1 FGA/g and 3.2 FTA/g. since hoiberg has had touches/min as high as 0.98 in 00-01 with the bulls, i don't see this as being unrealistic...
but what if you then replace szczerbiak with some other player who's touches/min are lower than his, and someone like hoiberg then has his touches/min increase another 10%-15% to the point that he is taking 12/13 FGA/g and maybe 3.8 FTA/g and scoring 18-19 pts/g. so now instead of 1.0 touches/min he's up to 1.1 touches/min. he's still only taking 12/13 shots playing 36 min/g, and the average SG in the NBA over the past few seasons averaging just 32 min/g or more takes on average 15-16 FGA/g, so hoiberg is still taking less FGA/g than just the average SG...
And actually, we probably picked the wrong example to maximize Hoiberg's exploits. If we wanted him to win the MVP award, this is probably the team we would put him with.
PG: Snow (36), McKie (12)
SG: McKie (24), Buckner (24)
SF: Hoiberg (36), Ryan Bowen (12)
PF: Ruffin (36), Madsen (12)
C: Ratliff (36), Ervin Johnson (12)
My guess is that this lineup in Bob's sim will win greater than 50 games and Hoiberg will score more than 25 points per game.
my guess is that you are indeed an expert on the simulation. i'll take your word for it on this one...
- implying that all these high percentage low usage shooters can ramp up their usage without penalty...
ok, i'll bite..... just what kind of penalty? do you have any inkling as to how to "penalize" players because they are put into a situation where they are called upon to handle the ball more often than they have previously in their careers? again i can show you players with increased touches/min that play as well as or better than they have previously, and some that are worse, but no consistent pattern one way or the other...
would this be a standard across the board penalty? an individual penalty? would it entail something like a decrease in FG%? an increase in turnovers? a decrease in FGA/min? a 15 yard penalty?...
bottom line - do you have any evidence whatsoever showing that players as a whole or even in general experience a decrease in some aspect of player performance with an increase in touches/min or your usage possessions?...
in BoP you discuss derek fisher, but i find in 04-05 fisher on the warriors with his highest touches/min in 6 seasons, and a ScFG% of .512 that is above his prior to that year career .496 ScFG%, and his career high scoring average despite playing less minutes than in 2 other seasons with the lakers...
you also mention robert horry. i have his touches/min ranging from 0.70 to 0.96 during his career, and i can't see a discernible difference in offensive performance based on touches/min, other than that associated with more touches (total points for example). looking at his top 4 seasons in touches/min (0.88-0.96) he has a ScFG% of .528, and in his bottom 4 seasons in touches/min (0.70-0.77) a ScFG% of .529. what i do see with horry is a slight increase in steals and blocked shots with lower touches/min, but that's on the defensive end...
"...implies that the people running the NBA are not just a little wrong. It implies also that the fundamental nature of basketball is poorly understood."
just what fundamental nature of basketball might you be speaking of? i've always thought of the fundamental nature of basketball as being put the ball in the hole and the team with the most points wins....
It implies that any sort of linear weights rating is wrong -- you should normalize a player's shots to 20% (or something) and turnovers (to something else) then put it together with rebounds, etc. It implies that pretty much every rating method is wrong because the context in which players are being used is incorrect.
wow!... thats pretty powerful.... every rating method?...
how is the "...context in which players are being used is incorrect..."? because some players increase efficiency with an increase in touches/min and some others show a decrease with an increase in touches/min, but that there is not definitive pattern saying efficiency always decreases with an increases in touches/min (or possessions)?...
would slapping some kind of penalty on player efficiency for every player that increases their touches/min or possessions during their career solve this conundrum?...
Dan's method which is totally different from others here has to be wrong because it is flawed by the decision to not let Fred Hoiberg shoot 25x per game.
i've never said in any way that dan's methodology is wrong. if dan does not want to let fred hoiberg shoot 25 times per game, i say go for it...
This is not just a matter of a tiny little assumption that has to be proven. This is a principle that really underlies the game of basketball.
sounds like not being able to show statistically that efficiency definitely always decreases with an increase in touches/min or possessions has shaken the foundations of basketball....
- Ultimately this is a principle I have used to make projections. It is just one of many tools to do that. How do we explain Hoiberg's extreme variation in TS%?
again if you look at the seasons where he's played his most minutes you'll find he's an excellent shooter. as you include seasons of less and less minutes played you are bound to find more variation in TS% or ScFG%, but looking at his seasons of most minutes played, i would not call the changes extreme. on the contrary, he's simply a very good to excellent shooter...
Bob seemed a tiny bit obtuse to me in that exchange with FFS.
just a tiny bit? not trying hard enough i guess...
The whole Minnesota team knows to look for Garnett as a first offensive option, and Garnett himself knows he bears that responsibility.
agreed... but what if he went down with an injury? or was traded away (for ben and rasheed wallace say)? how would the wolves players roles change? or do they change at all? or is it simply that who then gets the ball and when is simply dictated by how often the players generally handle the ball, but now that's increased slightly for each player because a player who had very high touches/min is now gone and replaced by a player who gets lower touches/min (or possessions)?...
The team is not charging down the court on every possession thinking "let's set up the offense...OK, where's Hoiberg...gotta get the ball in his hands".
well if reggie evans replaced kevin garnett in the t-wolves lineup, would it then be "...get szczerbiak the ball?..." or griffin? or hassell? or sprewell?...
The defense keys on Garnett as well. Do we have any idea how Hoiberg will perform with an entire defense keying on him?
do we have any idea how szczerbiak would perform if keyed on? griffin? hassell? how are any of these different than hoiberg? or are they all the same?...
And it's pretty funny to see Kevin Ollie referred to hear as a high efficiency player
has anybody in this thread referred to kevin ollie as a high efficiency player?...
I had the misfortune of watching Mr. Ollie during what must have been one of his highest usage seasons. This was when he played for the Cavs during Lebron's rookie year. The offense went through him often as the PG and believe me it was not pleasant to watch. He was helpless out there. Not because he made stupid decisions, but because he couldn't create at all.
yes 03-04 was ollie's highest touches/min, along with his 01-02 season with the bulls...
but i see very little difference between kevin ollie in 03-04 and eric snow in 04-05. both got 1.3 touches/min, both passed the ball with about 3/4 of all their touches, both shot the ball with about 1 out of every 7 touches. the only real difference i see between the two is ollie got fouled over twice as often per touch, and shot better than snow, and snow was the better defender than ollie. they even got steals and blocked shots at about the same rates...
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Dan Rosenbaum
Joined: 03 Jan 2005
Posts: 541
Location: Greensboro, North Carolina
PostPosted: Tue Aug 30, 2005 3:54 am Post subject: Reply with quote
BobC, you were so focused on debating that you completely missed the point of why I chose the 1999-2000 season for Hoiberg. Of the three seasons where Hoiberg has a possession rate of 16.0 percent or higher, it was the one with the most minutes played. The whole point of the exercise was to compare to a season where Hoiberg's possession usage approached what it would be in the simulation that you ran.
There is no magical number of minutes where all of the results become statistically significant. The standard errors for 2004-2005 (when Hoiberg played 1272 minutes) likely would only be about 20 percent smaller than in 1999-2000 (when he played 845 minutes).
And since you spent so many words debating every other point I made, I will assume that the lineup that I presented did indeed results in more than 50 wins and more than 25 points per game for Hoiberg.
And ignoring your simulator for a moment, do you really truly believe that Minnesota would only win 3 less games with Reggie Evans playing Garnett's minutes if 24 minutes of Sprewell were replaced by Hoiberg? Does the simulator ever produce results that you don't believe?
Back to top
View user's profile Send private message Send e-mail Visit poster's website Yahoo Messenger
kjb
Joined: 03 Jan 2005
Posts: 865
Location: Washington, DC
PostPosted: Tue Aug 30, 2005 6:49 am Post subject: Reply with quote
As is often the case in a discussion such as this one, I see truths on both sides. The long-held belief in hoops is that there is some relationship between efficiency and usage -- that as usage increases, efficiency will usually decline. This belief is fundamental enough that Dean can generate those skills curves, and his approach seemed to help the Sonics last season.
What I'd like to know is whether the examples cited of players increasing efficiency with increased usage are evidence of something systemic, or whether they're anomalies. I don't think anyone would assert that efficiency ALWAYS declines as usage increases. The question is -- what does the historical evidence actually show on this subject?
I think the study discussed here is a good start. As a stat analysis community, we ought to be going back to the stats to see what they show. And, (in my opinion) we ought not be getting stuck on digging for examples that prove our point, but instead should be looking for broader evidence. Is this efficieny/usage theory accurate, or is it another theory destined for the conventional wisdom bone yard?
Author Message
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
PostPosted: Fri Sep 02, 2005 4:28 pm Post subject: Reply with quote
bchaikin wrote:
If this inverse relationship holds, if it is significant at 95%, will anyone care?
it would be significant information even if it wasn't significant at 95%...
OK. I can run this.
Following the procedure I laid out above and getting the usage percentage average based on weighted average by minutes played (though it doesn't matter really as long as the individual averages are approximate), I find the following:
- An inverse relationship shows up between individual offensive rating and percentage of possessions used.
- For every increase of 1% in usage, offensive rating drops by about 0.6 over all the players.
- This coefficient is significant at well over 99%. There is roughly a 1 in 10 to the power of 38 that it's mere luck. According to my old quantum phys professor, that would be roughly the odds of banging your head against the wall and none of the molecules in your head actually hitting any of the molecules in the wall, thus avoiding pain.
- The regression has an r2 of 0.01, a case of low fit to the data being actually a meaningful result. If I were to re-add in the player averages, the r2 goes up to about 0.1.
- If you run the regression on different segments of the data -- high use games or low use games, you get essentially the same results.
- If you run the regression on players averaging high or low use, you see that the low use players (I used 18% as a cut off) are more sensitive to increases than high use players (23%). But both are still very significant. This implies that increasing Hoiberg's possessions 5% causes a bigger decline (about twice the size) than a similar increase in, say, Kevin Garnett. Or, from an optimization perspective, taking Garnett's possessions (who increases in efficiency only a little) and giving them to Hoiberg (who declines in efficiency a lot) has a pretty big cost in even this crude analysis.
This is much cruder than several of the studies I've done, but it generally reproduces the observations I've had in other proprietary studies as well as what I showed in Basketball on Paper. Clearly there is a lot of noise in this kind of analysis, which is why Bob can find examples where things just look to fluctuate around some average. But fluctuating around some average is always a zeroth order assumption. Using this elementary principle not only improves predictions but fits with a very reasonable model of the game of basketball.
What shooting percent do you use, Bob, to make projections for the coming year for, say, Allen Iverson? His career average? Last year's value? Some running average over a few years? Something more sophisticated? Because we know that the value you choose matters in your simulation and guys definitely fluctuate.
_________________
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
NickS
Joined: 30 Dec 2004
Posts: 384
PostPosted: Fri Sep 02, 2005 4:35 pm Post subject: Reply with quote
HoopStudies wrote:
- An inverse relationship shows up between individual offensive rating and percentage of possessions used.
- For every increase of 1% in usage, offensive rating drops by about 0.6 over all the players.
[...]
- If you run the regression on players averaging high or low use, you see that the low use players (I used 18% as a cut off) are more sensitive to increases than high use players (23%). But both are still very significant. This implies that increasing Hoiberg's possessions 5% causes a bigger decline (about twice the size) than a similar increase in, say, Kevin Garnett.
Thanks for running this. Very interesting. I was expecting that there would be a correlation, but I'm slightly surprised it's that large.
Does this mean that you're back from your vacation?
Back to top
View user's profile Send private message
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
PostPosted: Fri Sep 02, 2005 4:50 pm Post subject: Reply with quote
NickS wrote:
HoopStudies wrote:
- An inverse relationship shows up between individual offensive rating and percentage of possessions used.
- For every increase of 1% in usage, offensive rating drops by about 0.6 over all the players.
[...]
- If you run the regression on players averaging high or low use, you see that the low use players (I used 18% as a cut off) are more sensitive to increases than high use players (23%). But both are still very significant. This implies that increasing Hoiberg's possessions 5% causes a bigger decline (about twice the size) than a similar increase in, say, Kevin Garnett.
Thanks for running this. Very interesting. I was expecting that there would be a correlation, but I'm slightly surprised it's that large.
Does this mean that you're back from your vacation?
In studies where I actually include more stuff, the slope actually increases.
There is so much noise in this crude way of looking at it that the slope gets closer to 0.
Including what you suggested -- opponent D rating -- would increase that slope.
And, yeah, I'm back from vacation. Made a million phone calls but it looks like everyone is gone for Labor Day.
So I'm stuck putting together pictures, running errands, and doing regressions... Let's see if this Img tag actually works (see FFSBasketball post)...
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers.
Last edited by HoopStudies on Fri Sep 02, 2005 5:16 pm; edited 1 time in total
Back to top
View user's profile Send private message Visit poster's website
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Fri Sep 02, 2005 5:13 pm Post subject: Reply with quote
Here, I transferred it over to imageshack.us for you, which allows remote linking.
It's very nice.
Back to top
View user's profile Send private message
bchaikin
Joined: 27 Jan 2005
Posts: 688
Location: cleveland, ohio
PostPosted: Fri Sep 02, 2005 11:07 pm Post subject: Reply with quote
What shooting percent do you use, Bob, to make projections for the coming year for, say, Allen Iverson? His career average? Last year's value? Some running average over a few years? Something more sophisticated?
for the simulation i always use the most recent regular season data. in the off season that's the final regular season stats of the previous season. during the season its the regular season stats up and through the previous weekend (once about 18-20 games have been played).....
but in addition i also run (simulate) the player's stats from two seasons earlier and three seasons earlier if the data exists, to check for consistency. take joe johnson for example, other than his individual defense (outside of steals, blocks, and def reb), the only major difference between his 04-05 and 03-04 stats is that he shot much much better from 3pt range (with more 3pt FGA) in 04-05, 48% to 31%. that difference between his 04-05 and 03-04 3pt shooting is huge, and results (along with other minor changes in his game) in a difference of close to 4-5 wins per average 82 games (playing him 40 min/g) in favor of his 04-05 stats....
so using his 03-04 stats in any 04-05 simulation if i remember i think he produced wins at the rate of just 4-5 more wins per average 82 games than some of the worst starting SGs in 04-05 (playing 40 min/g). but using his 04-05 stats that gets bumped up to 9-10 more wins per average 82 games than some of the worst starting SGs in 04-05. that's quite a difference (5-6 wins, maximized by him playing 40 min/g)...
so any analyses i run for joe johnson using his 04-05 stats is tempered by also quoting his simulation results using his 03-04 stats (and 02-03 for comparison), saying to the effect of should his 3pt shooting go south like it did in 03-04, expect those wins to drop...
for players who played few minutes in a season, say < 500, i will often add their 04-05 and 03-04 stats together and use that for the simulation...
actually the difference in using allen iverson for 40 min/g using his 04-05 versus his 03-04 stats is also about 5-6 wins per average 82 games in favor of his 04-05 stats...
Because we know that the value you choose matters in your simulation and guys definitely fluctuate.
i don't "choose" values, i only use actual data. so it doesn't matter how much anyone's stats fluctuate from one season to the next because i can run simulations using a player's data from any previous seasons (or the current season if the analysis is done in season). if for example a player was hurt in his most recent season but not the previous two, and he's now healthy, you can use all 3 sets of single season data and quantify the result of the most recent data saying he was hurt...
but in all honesty its been my experience that most players do not fluctuate that much from one season to the next., not upwards of the 5-6 wins difference you see in joe johnson or iverson above. yes you do see the few that do like how shawn marion and amare stoudemire were both much better in 04-05 than in 03-04, but for the most part players do not tend to fluctuate by more than 2-3 wins from one season to the next, and many are less than that when you normalize their minutes played (i.e. in the sim run them for the same amount of minutes using their stats from different seasons regardless of how many min/g they actually played in those different seasons)....
the only time i will actually change a player's stats from their real life numbers to something else is when running what-if scenarios, i.e. how much better or worse would a team be if for example player A shot 50% from 2pt range rather than his current 45%, or player B got 25% more rebounds, or player C had a much better or worse individual defensive FG% rating...
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: Thu Sep 08, 2005 9:05 am Post subject: Reply with quote
Dan Rosenbaum wrote:
I think the instrumental variables approach that I am suggesting could work at the game-level or the season-level. It would be interesting to do both.
I had an idea that might be fun. There is the Journal of Quantitative Analysis in Sports that DeanO and Roland are helping start. Done right, I think this would be a nice article for that journal.
http://www.bepress.com/jqas/
I am very busy this Fall and probably only have time to serve as the organizer of this, but I was wondering who might be interested in being part of this project. We probably can't have 10 co-authors, but I suspect that even four or five would not be completely unworkable. Here is what I think needs to be done.
(1) putting together the season-level data
(2) putting together the game-level data
(3) devising the empirical strategy
(4) running the regressions
(5) putting together the tables
(6) writing up the results/writing the paper
(7) revising the paper
(Cool organizing the process
(9) handling the submission
With the limited time I have this Fall, I think that I am best employed doing (3), (4), part of (5), some guidance on (6), (7), (Cool, and (9). If there were folks who were interested in helping with (1), (2), (5), (6), and (7), that would be great.
Let me know either through e-mail, PM, or in this thread if you might be interested. If we dozens of people interested in helping out, we may need to make some choices about whose name goes on the paper, but it would be good to get many people involved in this. (But it would look bad to have a list of co-authors longer than the article itself - although I have seen that done with medical journals.)
Having worked on papers for medical journals (day job), I could probably help some with #6-#9, and I would LOVE to be involved in it.
Back to top
View user's profile Send private message Send e-mail AIM Address
kjb
Joined: 03 Jan 2005
Posts: 865
Location: Washington, DC
PostPosted: Thu Sep 08, 2005 3:08 pm Post subject: Reply with quote
I can help with the writing/revision process. I won't have time to help with the data side of the project, however -- plus, I'm a better writer than mathemetician.
Back to top
View user's profile Send private message AIM Address Yahoo Messenger
Dan Rosenbaum
Joined: 03 Jan 2005
Posts: 541
Location: Greensboro, North Carolina
PostPosted: Thu Sep 08, 2005 3:41 pm Post subject: Reply with quote
I have been swamped with schoolwork over the past few weeks, as my posts here and at my blog has dwindled to almost nothing. I owe e-mails to lots of people this weekend.
This weekend I plan to send out an e-mail to those of you who have expressed interest. We probably have about 10 people that have expressed interest in some fashion or another and 10 co-authors probably is too many. But what I was hoping to do is have maybe half those folks be co-authors and the other half be involved in the process and get an acknowledgment in the credits. On the other hand, we could work it like some medical journal articles where we have 10 co-authors for a 10 sentence paper. Very Happy
Back to top
View user's profile Send private message Send e-mail Visit poster's website Yahoo Messenger
gabefarkas
Joined: 31 Dec 2004
Posts: 1313
Location: Durham, NC
PostPosted: Fri Sep 09, 2005 8:21 am Post subject: Reply with quote
happy to help in whatever way -- gfarkas@gmail.com
Back to top
View user's profile Send private message Send e-mail AIM Address
Mountain
Joined: 13 Mar 2007
Posts: 1527
PostPosted: Mon Dec 17, 2007 9:32 pm Post subject: Reply with quote
Is the research project / paper suggested in this thread still a possibility sometime in future? A good next step after past APBR member papers groundwork? Would running it as suggested then still be the approach favored today or modify?
Back to top
View user's profile Send private message
Dan Rosenbaum
Joined: 03 Jan 2005
Posts: 541
Location: Greensboro, North Carolina
PostPosted: Tue Dec 18, 2007 12:06 am Post subject: Reply with quote
Mountain wrote:
Is the research project / paper suggested in this thread still a possibility sometime in future? A good next step after past APBR member papers groundwork? Would running it as suggested then still be the approach favored today or modify?
Probably not by me. I just don't have the time right now.
Joined: 30 Dec 2004
Posts: 384
PostPosted: Mon Aug 29, 2005 3:10 pm Post subject: Reply with quote
ffs,
thanks for posting that exchange, I appreciate seeing the elaboration by bob on what he is and isn't claiming.
I haven't finished reading the whole exchange but so far it's very interesting.
Back to top
View user's profile Send private message
marc200
Joined: 23 Aug 2005
Posts: 14
Location: 40 miles north of Lebron James' birthplace
PostPosted: Mon Aug 29, 2005 4:03 pm Post subject: Reply with quote
Bob seemed a tiny bit obtuse to me in that exchange with FFS. The point isn't that the COACH changes touches per minute by calling plays. It's that the team is constructed with everyone playing roles. That happens in a million ways from the moment the general manager chooses the players, and is probably reinforced informally all the time through everyone's assumptions about who plays what role. The whole Minnesota team knows to look for Garnett as a first offensive option, and Garnett himself knows he bears that responsibility. The team is not charging down the court on every possession thinking "let's set up the offense...OK, where's Hoiberg...gotta get the ball in his hands". The defense keys on Garnett as well. Do we have any idea how Hoiberg will perform with an entire defense keying on him?
And it's pretty funny to see Kevin Ollie referred to hear as a high efficiency player (is he statistically? Couldn't quite get that straight?). I had the misfortune of watching Mr. Ollie during what must have been one of his highest usage seasons. This was when he played for the Cavs during Lebron's rookie year. The offense went through him often as the PG and believe me it was not pleasant to watch. He was helpless out there. Not because he made stupid decisions, but because he couldn't create at all.
_________________
Marcus Stanley
Back to top
View user's profile Send private message
NickS
Joined: 30 Dec 2004
Posts: 384
PostPosted: Mon Aug 29, 2005 10:35 pm Post subject: Reply with quote
An anecdotal piece of evidence in favor of the null hypothesis (players do not lose offensive efficiency when they have to take on more responsibility to create offense).
Look at the 2003 Denver nuggets, a team that was desperately lacking in players with the ability to create offense. The consensus view about that team was that they played hard on deffense but their offense was terrible because they lacked scorers.
Their offense was terrible, but the fact that they had a number of low volume scorers may not have made it any worse.
Look at all of the players from that nuggets team and see if any of them improved their offensive rating when they had better scorers around them.
Most of them don't. Nene improved slightly but he was a rookie in 2003 so his improvement could be explained for other reasons. James Posey had a temendous improvement in Ortg in 2004 which is so large as to be difficult to explain by any measure. Rodney White had a slight jump but there's no sign that, as a group, they played bettwer when they were playing with better scorers.
(note, however, that Juwan Howard's was much better on dallas the year before he stayed at his 2003 level after he went to Orlando but Orlando was also a terrible team.)
There are enough that improve that it isn't a strong piece of evidence but I find it interesting.
Back to top
View user's profile Send private message
Kevin Pelton
Site Admin
Joined: 30 Dec 2004
Posts: 979
Location: Seattle
PostPosted: Mon Aug 29, 2005 10:53 pm Post subject: Reply with quote
That discussion only further convinces me that a big reason this debate is as intractable as it has been is that there are two very different prisms through which the game of basketball is being viewed.
Bob is speaking about touches. Everyone else is speaking about possessions. These are similar concepts, but certainly not the same.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
bchaikin
Joined: 27 Jan 2005
Posts: 689
Location: cleveland, ohio
PostPosted: Tue Aug 30, 2005 2:46 am Post subject: Reply with quote
What's interesting is that in the Bob's sim, we have actual evidence on how the top two leading scorers might do with increased touches.
what's interesting with the simulation is that you have actual statistical evidence for any simulation results - by simply looking at players' real life numbers for comparison...
Fred Hoiberg had a possession usage of 12.1% last season and scored 13.7 points per 40 minutes. In his highest possession usage season (16.5% in 1999-2000), he scored just 13.2 points per 40 minutes. His true shooting percentage was down, and his turnover rate was much higher with the greater possession usage.
so, his "...true shooting percentage was down, and his turnover rate was much higher, with the greater possession usage...". factual statements, but quite misleading as they don't tell the true story about fred hoiberg...
1st off in 04-05 hoiberg's turnover rate was just 2%, meaning just 2 turnovers per 100 touches, which is exceptionally low, one of the lowest marks ever for a player. in fact it is so low that it is the 2nd lowest turnovers/touch mark for a player since 1977-78 for all players playing at least 1000 minutes in a season (that's out of 6023 player seasons with 1000+ minutes played since 1977-78 ). so pretty much any amount of "normal" turnovers will be significantly higher than his 04-05 rate...
2nd hoiberg only played 845 minutes in 99-00, and yes his turnover rate was much higher than in 04-05. but it certainly wasn't high by any standard - i have it at 5% or 5 turnovers per 100 touches, which is about average for an SG. but those 845 minutes was just the 5th highest amount of minutes he ever played in a single season. why not look at the seasons he played his most minutes? wouldn't that be of more significance statistically speaking? why skew this discussion by looking at just a small sample size of data?...
if you look at the 4 seasons where he played his most minutes, his rate of turnovers was just 3%, 3%, 4%, and 2% - these are very low. as a matter of fact that rate of 5% turnovers in 99-00 was the highest rate of turnovers for hoiberg in all 8 seasons where he played at least 500 minutes. so you just happened to choose the season where he had his worst rate of turnovers, and it came in a season where he played all of 845 minutes...
and yes his shooting was down that season, but again you chose a season out of the 8 that he played at least 500 minutes that was his 2nd worst shooting season. in the 4 seasons where he played his most minutes his Scoring FG%s were 59%, 51%, 60%, and 65%, for an average of 59% (weighted to total minutes). so when he has gotten the playing time he has been an excellent shooter, one of the best in the league, and in the two seasons where he played his most minutes, he shot a ScFG% of 59% and 60%, and one of those two seasons he had his highest touches/min and the other his lowest touches/min. so his touches/min, whether his career highest or lowest, certainly hasn't affect his excellent shooting...
so to imply that his shooting percentage was "down" by choosing a season where he played less than 1000 minutes, when he played 4 other seasons of 1000+ minutes, doesn't truly reflect hoiberg's shooting ability. he played 10 seasons in the league, and played 2/3 of his career minutes in just 4 of those seasons, and shot a Scoring FG% of 59%. the other 1/3 of his career minutes (3200 total) were stretched out over 6 seasons, and he shot a Scoring FG% of just 52% over that range. so if anything this is a player who shoots much better when given the playing time, regardless of touches/min. his touches/min were at 0.95 in 99-00, but at 0.98 in 00-01...
hoiberg played his most minutes in a season (2247) in 00-01 with the bulls. that season he also had his highest touches/min mark of 0.98, which is about average for an SG. that season his Scoring FG% was 59%, the 3rd highest mark in the league that season for an SG (behind brent barry's 65% and ray allen's 60%). in his season with his 2nd most minutes (03-04) he had a Scoring FG% of 60% with just 0.65 touches/min...
hoiberg could have had his touches/min doubled over his 2nd highest mark that 99-00 season you mentioned, but - tell me, from the point of view of a professional statistician - how statistically significant is it to imply hoiberg shot worse with high touches/min if it came in just 845 minutes, when he played significantly more minutes in 4 other seasons with much better shooting and lower turnovers and just as high touches/min in one of those seasons playing over twice as many minutes?...
also in 00-01 hoiberg and brent barry had virtually identical numbers. both shot extremely well, both had close to 1.0 touches/min (barry 1.05, hoiberg 0.98 ), both shot the ball with 1/5 of all their touches, and both passed the ball with 2/3 of their touches. they even got fouled as often as one another per touch, and got simlar amounts of steals and blocks. yet barry went on to have a season just 3 years later with 30% more touches than in 00-01 with no loss in shooting efficiency, and hoiberg 3 years later got about 30% less touches than in 00-01 and no loss in shooting efficiency. why should one increase and the other decrease?...
Bruce Bowen had a possession usage of 11.6% last season and scored 10.3 points per 40 minutes. In his highest possession usage season (12.8% in 1997-1998), he scored just 10.4 points per 40 minutes. His true shooting percentage was down, and his turnover rate was higher with the greater possession usage.
bowen's touches/min have ranged from 0.50 to 0.70 over his career, looking at the seasons where he's played at least 1000 minutes, and yes in his case he shot best with fewer touches. but there are players who like hoiberg shoot very well with their highest and lowest touches/min, and some who shoot better with their higher touches/min...
So when we look back at their careers, the top two leading scorers on this simulated team did not increase their scoring when their usage increased. If historical patterns persisted, then that would take away more than 10 points per game away from the team that Bob simulated. Instead of a .500 team, it would be one of the worst in the NBA.
again your statement is in part true. but it fails to take into account what happens when you place certain players into situations different from what they have experienced, but that others have, and dismisses all those situations where the evidence does show players touches/min changing - sometimes significantly - based on who their teammates are (such as when they are traded, signed as free agents, etc)...
for example, how would the 04-05 performances of the t-wolves players have changed had reggie evans of the sonics (probably the lowest touches/min of any starting PF in the league last season) played 40 min/g at PF instead of kevin garnett for minnesota? garnett got 1.6 touches/min in 04-05, over 3 times what reggie evans got in 04-05. forget an entire team of low touches/min players, how about the change of just one player - a low touches/min starter playing 40 min/g replacing a very high touches/min player, especially a frontcourt player - and the other t-wolves players playing the same minutes that they did in 04-05?...
how would you approach this situation? is it your opinion that players like fred hoiberg are simply incapable of handling the ball more often than they have in real life, simply because they have never been in a situation in the nba that has not called upon for them to do so? to me this sounds vaguely familiar to the one time arguement that players who have only played say 15-20 min/g over their career are not capable of playing 36-40 min/g when called upon to do so. but there are many examples of just this occuring (like expansion teams), and without a decrease in their abilities, just as there are many examples of players who have experienced significant increases or decreases in their touches/min based on who their teammates are without significant changes in their efficiency...
when i run a simulation of the 04-05 t-wolves such that their players end up playing close to their total minutes actually played in 04-05, the sub pattern looks like this:
min/g - player - (touches/min)
16 olowokandi (0.50)
40 garnett (1.51)
32 szczerbiak (0.90)
32 sprewell (0.97)
20 hudson (1.41)
8 madsen (0.43)
20 griffin (0.68 )
24 hassell (0.67)
16 hoiberg (0.69)
20 cassell (1.78 )
4 e.johnson
8 a.carter
a simulation of 8200 games against the wolves 04-05 schedule several times, and averaging the results, shows the team with an average W-L record of 44.5-37.5. they went 44-38 in real life, and had the per game point differential of a team that should have gone 45-37...
if i switch the minutes of hoiberg and sprewell, and give hoiberg 36 min/g and sprewell just 12 min/g:
min/g - player - (touches/min)
16 olowokandi (0.53)
40 garnett (1.57)
32 szczerbiak (0.92)
12 sprewell (0.94)
20 hudson (1.55)
8 madsen (0.46)
20 griffin (0.71)
24 hassell (0.73)
36 hoiberg (0.75)
20 cassell (1.82)
4 e.johnson
8 a.carter
and run the simulation, they average a W-L record of 49-33. hoiberg averages 8.5 FGA/g, and 2.7 FTA/g. his touches/min increase to 0.75...
then replacing garnett with reggie evans:
min/g - player - (touches/min)
16 olowokandi (0.65)
40 r.evans (0.59)
32 szczerbiak (1.14)
12 sprewell (1.07)
20 hudson (1.74)
8 madsen (0.51)
20 griffin (0.79)
24 hassell (0.85)
36 hoiberg (0.89)
20 cassell (1.92)
4 e.johnson
8 a.carter
the team drops to an average a 41.5-40.5 W-L record playing against the wolves 04-05 schedule. but as you can see everyones touches/min increased with evans replacing garnett. szczerbiak increased 19%, olowokandi 18%, hoiberg 16%, hudson 11%, etc. hoiberg took 10.1 FGA/g and 3.2 FTA/g. since hoiberg has had touches/min as high as 0.98 in 00-01 with the bulls, i don't see this as being unrealistic...
but what if you then replace szczerbiak with some other player who's touches/min are lower than his, and someone like hoiberg then has his touches/min increase another 10%-15% to the point that he is taking 12/13 FGA/g and maybe 3.8 FTA/g and scoring 18-19 pts/g. so now instead of 1.0 touches/min he's up to 1.1 touches/min. he's still only taking 12/13 shots playing 36 min/g, and the average SG in the NBA over the past few seasons averaging just 32 min/g or more takes on average 15-16 FGA/g, so hoiberg is still taking less FGA/g than just the average SG...
And actually, we probably picked the wrong example to maximize Hoiberg's exploits. If we wanted him to win the MVP award, this is probably the team we would put him with.
PG: Snow (36), McKie (12)
SG: McKie (24), Buckner (24)
SF: Hoiberg (36), Ryan Bowen (12)
PF: Ruffin (36), Madsen (12)
C: Ratliff (36), Ervin Johnson (12)
My guess is that this lineup in Bob's sim will win greater than 50 games and Hoiberg will score more than 25 points per game.
my guess is that you are indeed an expert on the simulation. i'll take your word for it on this one...
- implying that all these high percentage low usage shooters can ramp up their usage without penalty...
ok, i'll bite..... just what kind of penalty? do you have any inkling as to how to "penalize" players because they are put into a situation where they are called upon to handle the ball more often than they have previously in their careers? again i can show you players with increased touches/min that play as well as or better than they have previously, and some that are worse, but no consistent pattern one way or the other...
would this be a standard across the board penalty? an individual penalty? would it entail something like a decrease in FG%? an increase in turnovers? a decrease in FGA/min? a 15 yard penalty?...
bottom line - do you have any evidence whatsoever showing that players as a whole or even in general experience a decrease in some aspect of player performance with an increase in touches/min or your usage possessions?...
in BoP you discuss derek fisher, but i find in 04-05 fisher on the warriors with his highest touches/min in 6 seasons, and a ScFG% of .512 that is above his prior to that year career .496 ScFG%, and his career high scoring average despite playing less minutes than in 2 other seasons with the lakers...
you also mention robert horry. i have his touches/min ranging from 0.70 to 0.96 during his career, and i can't see a discernible difference in offensive performance based on touches/min, other than that associated with more touches (total points for example). looking at his top 4 seasons in touches/min (0.88-0.96) he has a ScFG% of .528, and in his bottom 4 seasons in touches/min (0.70-0.77) a ScFG% of .529. what i do see with horry is a slight increase in steals and blocked shots with lower touches/min, but that's on the defensive end...
"...implies that the people running the NBA are not just a little wrong. It implies also that the fundamental nature of basketball is poorly understood."
just what fundamental nature of basketball might you be speaking of? i've always thought of the fundamental nature of basketball as being put the ball in the hole and the team with the most points wins....
It implies that any sort of linear weights rating is wrong -- you should normalize a player's shots to 20% (or something) and turnovers (to something else) then put it together with rebounds, etc. It implies that pretty much every rating method is wrong because the context in which players are being used is incorrect.
wow!... thats pretty powerful.... every rating method?...
how is the "...context in which players are being used is incorrect..."? because some players increase efficiency with an increase in touches/min and some others show a decrease with an increase in touches/min, but that there is not definitive pattern saying efficiency always decreases with an increases in touches/min (or possessions)?...
would slapping some kind of penalty on player efficiency for every player that increases their touches/min or possessions during their career solve this conundrum?...
Dan's method which is totally different from others here has to be wrong because it is flawed by the decision to not let Fred Hoiberg shoot 25x per game.
i've never said in any way that dan's methodology is wrong. if dan does not want to let fred hoiberg shoot 25 times per game, i say go for it...
This is not just a matter of a tiny little assumption that has to be proven. This is a principle that really underlies the game of basketball.
sounds like not being able to show statistically that efficiency definitely always decreases with an increase in touches/min or possessions has shaken the foundations of basketball....
- Ultimately this is a principle I have used to make projections. It is just one of many tools to do that. How do we explain Hoiberg's extreme variation in TS%?
again if you look at the seasons where he's played his most minutes you'll find he's an excellent shooter. as you include seasons of less and less minutes played you are bound to find more variation in TS% or ScFG%, but looking at his seasons of most minutes played, i would not call the changes extreme. on the contrary, he's simply a very good to excellent shooter...
Bob seemed a tiny bit obtuse to me in that exchange with FFS.
just a tiny bit? not trying hard enough i guess...
The whole Minnesota team knows to look for Garnett as a first offensive option, and Garnett himself knows he bears that responsibility.
agreed... but what if he went down with an injury? or was traded away (for ben and rasheed wallace say)? how would the wolves players roles change? or do they change at all? or is it simply that who then gets the ball and when is simply dictated by how often the players generally handle the ball, but now that's increased slightly for each player because a player who had very high touches/min is now gone and replaced by a player who gets lower touches/min (or possessions)?...
The team is not charging down the court on every possession thinking "let's set up the offense...OK, where's Hoiberg...gotta get the ball in his hands".
well if reggie evans replaced kevin garnett in the t-wolves lineup, would it then be "...get szczerbiak the ball?..." or griffin? or hassell? or sprewell?...
The defense keys on Garnett as well. Do we have any idea how Hoiberg will perform with an entire defense keying on him?
do we have any idea how szczerbiak would perform if keyed on? griffin? hassell? how are any of these different than hoiberg? or are they all the same?...
And it's pretty funny to see Kevin Ollie referred to hear as a high efficiency player
has anybody in this thread referred to kevin ollie as a high efficiency player?...
I had the misfortune of watching Mr. Ollie during what must have been one of his highest usage seasons. This was when he played for the Cavs during Lebron's rookie year. The offense went through him often as the PG and believe me it was not pleasant to watch. He was helpless out there. Not because he made stupid decisions, but because he couldn't create at all.
yes 03-04 was ollie's highest touches/min, along with his 01-02 season with the bulls...
but i see very little difference between kevin ollie in 03-04 and eric snow in 04-05. both got 1.3 touches/min, both passed the ball with about 3/4 of all their touches, both shot the ball with about 1 out of every 7 touches. the only real difference i see between the two is ollie got fouled over twice as often per touch, and shot better than snow, and snow was the better defender than ollie. they even got steals and blocked shots at about the same rates...
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Dan Rosenbaum
Joined: 03 Jan 2005
Posts: 541
Location: Greensboro, North Carolina
PostPosted: Tue Aug 30, 2005 3:54 am Post subject: Reply with quote
BobC, you were so focused on debating that you completely missed the point of why I chose the 1999-2000 season for Hoiberg. Of the three seasons where Hoiberg has a possession rate of 16.0 percent or higher, it was the one with the most minutes played. The whole point of the exercise was to compare to a season where Hoiberg's possession usage approached what it would be in the simulation that you ran.
There is no magical number of minutes where all of the results become statistically significant. The standard errors for 2004-2005 (when Hoiberg played 1272 minutes) likely would only be about 20 percent smaller than in 1999-2000 (when he played 845 minutes).
And since you spent so many words debating every other point I made, I will assume that the lineup that I presented did indeed results in more than 50 wins and more than 25 points per game for Hoiberg.
And ignoring your simulator for a moment, do you really truly believe that Minnesota would only win 3 less games with Reggie Evans playing Garnett's minutes if 24 minutes of Sprewell were replaced by Hoiberg? Does the simulator ever produce results that you don't believe?
Back to top
View user's profile Send private message Send e-mail Visit poster's website Yahoo Messenger
kjb
Joined: 03 Jan 2005
Posts: 865
Location: Washington, DC
PostPosted: Tue Aug 30, 2005 6:49 am Post subject: Reply with quote
As is often the case in a discussion such as this one, I see truths on both sides. The long-held belief in hoops is that there is some relationship between efficiency and usage -- that as usage increases, efficiency will usually decline. This belief is fundamental enough that Dean can generate those skills curves, and his approach seemed to help the Sonics last season.
What I'd like to know is whether the examples cited of players increasing efficiency with increased usage are evidence of something systemic, or whether they're anomalies. I don't think anyone would assert that efficiency ALWAYS declines as usage increases. The question is -- what does the historical evidence actually show on this subject?
I think the study discussed here is a good start. As a stat analysis community, we ought to be going back to the stats to see what they show. And, (in my opinion) we ought not be getting stuck on digging for examples that prove our point, but instead should be looking for broader evidence. Is this efficieny/usage theory accurate, or is it another theory destined for the conventional wisdom bone yard?
Author Message
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
PostPosted: Fri Sep 02, 2005 4:28 pm Post subject: Reply with quote
bchaikin wrote:
If this inverse relationship holds, if it is significant at 95%, will anyone care?
it would be significant information even if it wasn't significant at 95%...
OK. I can run this.
Following the procedure I laid out above and getting the usage percentage average based on weighted average by minutes played (though it doesn't matter really as long as the individual averages are approximate), I find the following:
- An inverse relationship shows up between individual offensive rating and percentage of possessions used.
- For every increase of 1% in usage, offensive rating drops by about 0.6 over all the players.
- This coefficient is significant at well over 99%. There is roughly a 1 in 10 to the power of 38 that it's mere luck. According to my old quantum phys professor, that would be roughly the odds of banging your head against the wall and none of the molecules in your head actually hitting any of the molecules in the wall, thus avoiding pain.
- The regression has an r2 of 0.01, a case of low fit to the data being actually a meaningful result. If I were to re-add in the player averages, the r2 goes up to about 0.1.
- If you run the regression on different segments of the data -- high use games or low use games, you get essentially the same results.
- If you run the regression on players averaging high or low use, you see that the low use players (I used 18% as a cut off) are more sensitive to increases than high use players (23%). But both are still very significant. This implies that increasing Hoiberg's possessions 5% causes a bigger decline (about twice the size) than a similar increase in, say, Kevin Garnett. Or, from an optimization perspective, taking Garnett's possessions (who increases in efficiency only a little) and giving them to Hoiberg (who declines in efficiency a lot) has a pretty big cost in even this crude analysis.
This is much cruder than several of the studies I've done, but it generally reproduces the observations I've had in other proprietary studies as well as what I showed in Basketball on Paper. Clearly there is a lot of noise in this kind of analysis, which is why Bob can find examples where things just look to fluctuate around some average. But fluctuating around some average is always a zeroth order assumption. Using this elementary principle not only improves predictions but fits with a very reasonable model of the game of basketball.
What shooting percent do you use, Bob, to make projections for the coming year for, say, Allen Iverson? His career average? Last year's value? Some running average over a few years? Something more sophisticated? Because we know that the value you choose matters in your simulation and guys definitely fluctuate.
_________________
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
NickS
Joined: 30 Dec 2004
Posts: 384
PostPosted: Fri Sep 02, 2005 4:35 pm Post subject: Reply with quote
HoopStudies wrote:
- An inverse relationship shows up between individual offensive rating and percentage of possessions used.
- For every increase of 1% in usage, offensive rating drops by about 0.6 over all the players.
[...]
- If you run the regression on players averaging high or low use, you see that the low use players (I used 18% as a cut off) are more sensitive to increases than high use players (23%). But both are still very significant. This implies that increasing Hoiberg's possessions 5% causes a bigger decline (about twice the size) than a similar increase in, say, Kevin Garnett.
Thanks for running this. Very interesting. I was expecting that there would be a correlation, but I'm slightly surprised it's that large.
Does this mean that you're back from your vacation?
Back to top
View user's profile Send private message
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
PostPosted: Fri Sep 02, 2005 4:50 pm Post subject: Reply with quote
NickS wrote:
HoopStudies wrote:
- An inverse relationship shows up between individual offensive rating and percentage of possessions used.
- For every increase of 1% in usage, offensive rating drops by about 0.6 over all the players.
[...]
- If you run the regression on players averaging high or low use, you see that the low use players (I used 18% as a cut off) are more sensitive to increases than high use players (23%). But both are still very significant. This implies that increasing Hoiberg's possessions 5% causes a bigger decline (about twice the size) than a similar increase in, say, Kevin Garnett.
Thanks for running this. Very interesting. I was expecting that there would be a correlation, but I'm slightly surprised it's that large.
Does this mean that you're back from your vacation?
In studies where I actually include more stuff, the slope actually increases.
There is so much noise in this crude way of looking at it that the slope gets closer to 0.
Including what you suggested -- opponent D rating -- would increase that slope.
And, yeah, I'm back from vacation. Made a million phone calls but it looks like everyone is gone for Labor Day.
So I'm stuck putting together pictures, running errands, and doing regressions... Let's see if this Img tag actually works (see FFSBasketball post)...
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers.
Last edited by HoopStudies on Fri Sep 02, 2005 5:16 pm; edited 1 time in total
Back to top
View user's profile Send private message Visit poster's website
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Fri Sep 02, 2005 5:13 pm Post subject: Reply with quote
Here, I transferred it over to imageshack.us for you, which allows remote linking.
It's very nice.
Back to top
View user's profile Send private message
bchaikin
Joined: 27 Jan 2005
Posts: 688
Location: cleveland, ohio
PostPosted: Fri Sep 02, 2005 11:07 pm Post subject: Reply with quote
What shooting percent do you use, Bob, to make projections for the coming year for, say, Allen Iverson? His career average? Last year's value? Some running average over a few years? Something more sophisticated?
for the simulation i always use the most recent regular season data. in the off season that's the final regular season stats of the previous season. during the season its the regular season stats up and through the previous weekend (once about 18-20 games have been played).....
but in addition i also run (simulate) the player's stats from two seasons earlier and three seasons earlier if the data exists, to check for consistency. take joe johnson for example, other than his individual defense (outside of steals, blocks, and def reb), the only major difference between his 04-05 and 03-04 stats is that he shot much much better from 3pt range (with more 3pt FGA) in 04-05, 48% to 31%. that difference between his 04-05 and 03-04 3pt shooting is huge, and results (along with other minor changes in his game) in a difference of close to 4-5 wins per average 82 games (playing him 40 min/g) in favor of his 04-05 stats....
so using his 03-04 stats in any 04-05 simulation if i remember i think he produced wins at the rate of just 4-5 more wins per average 82 games than some of the worst starting SGs in 04-05 (playing 40 min/g). but using his 04-05 stats that gets bumped up to 9-10 more wins per average 82 games than some of the worst starting SGs in 04-05. that's quite a difference (5-6 wins, maximized by him playing 40 min/g)...
so any analyses i run for joe johnson using his 04-05 stats is tempered by also quoting his simulation results using his 03-04 stats (and 02-03 for comparison), saying to the effect of should his 3pt shooting go south like it did in 03-04, expect those wins to drop...
for players who played few minutes in a season, say < 500, i will often add their 04-05 and 03-04 stats together and use that for the simulation...
actually the difference in using allen iverson for 40 min/g using his 04-05 versus his 03-04 stats is also about 5-6 wins per average 82 games in favor of his 04-05 stats...
Because we know that the value you choose matters in your simulation and guys definitely fluctuate.
i don't "choose" values, i only use actual data. so it doesn't matter how much anyone's stats fluctuate from one season to the next because i can run simulations using a player's data from any previous seasons (or the current season if the analysis is done in season). if for example a player was hurt in his most recent season but not the previous two, and he's now healthy, you can use all 3 sets of single season data and quantify the result of the most recent data saying he was hurt...
but in all honesty its been my experience that most players do not fluctuate that much from one season to the next., not upwards of the 5-6 wins difference you see in joe johnson or iverson above. yes you do see the few that do like how shawn marion and amare stoudemire were both much better in 04-05 than in 03-04, but for the most part players do not tend to fluctuate by more than 2-3 wins from one season to the next, and many are less than that when you normalize their minutes played (i.e. in the sim run them for the same amount of minutes using their stats from different seasons regardless of how many min/g they actually played in those different seasons)....
the only time i will actually change a player's stats from their real life numbers to something else is when running what-if scenarios, i.e. how much better or worse would a team be if for example player A shot 50% from 2pt range rather than his current 45%, or player B got 25% more rebounds, or player C had a much better or worse individual defensive FG% rating...
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: Thu Sep 08, 2005 9:05 am Post subject: Reply with quote
Dan Rosenbaum wrote:
I think the instrumental variables approach that I am suggesting could work at the game-level or the season-level. It would be interesting to do both.
I had an idea that might be fun. There is the Journal of Quantitative Analysis in Sports that DeanO and Roland are helping start. Done right, I think this would be a nice article for that journal.
http://www.bepress.com/jqas/
I am very busy this Fall and probably only have time to serve as the organizer of this, but I was wondering who might be interested in being part of this project. We probably can't have 10 co-authors, but I suspect that even four or five would not be completely unworkable. Here is what I think needs to be done.
(1) putting together the season-level data
(2) putting together the game-level data
(3) devising the empirical strategy
(4) running the regressions
(5) putting together the tables
(6) writing up the results/writing the paper
(7) revising the paper
(Cool organizing the process
(9) handling the submission
With the limited time I have this Fall, I think that I am best employed doing (3), (4), part of (5), some guidance on (6), (7), (Cool, and (9). If there were folks who were interested in helping with (1), (2), (5), (6), and (7), that would be great.
Let me know either through e-mail, PM, or in this thread if you might be interested. If we dozens of people interested in helping out, we may need to make some choices about whose name goes on the paper, but it would be good to get many people involved in this. (But it would look bad to have a list of co-authors longer than the article itself - although I have seen that done with medical journals.)
Having worked on papers for medical journals (day job), I could probably help some with #6-#9, and I would LOVE to be involved in it.
Back to top
View user's profile Send private message Send e-mail AIM Address
kjb
Joined: 03 Jan 2005
Posts: 865
Location: Washington, DC
PostPosted: Thu Sep 08, 2005 3:08 pm Post subject: Reply with quote
I can help with the writing/revision process. I won't have time to help with the data side of the project, however -- plus, I'm a better writer than mathemetician.
Back to top
View user's profile Send private message AIM Address Yahoo Messenger
Dan Rosenbaum
Joined: 03 Jan 2005
Posts: 541
Location: Greensboro, North Carolina
PostPosted: Thu Sep 08, 2005 3:41 pm Post subject: Reply with quote
I have been swamped with schoolwork over the past few weeks, as my posts here and at my blog has dwindled to almost nothing. I owe e-mails to lots of people this weekend.
This weekend I plan to send out an e-mail to those of you who have expressed interest. We probably have about 10 people that have expressed interest in some fashion or another and 10 co-authors probably is too many. But what I was hoping to do is have maybe half those folks be co-authors and the other half be involved in the process and get an acknowledgment in the credits. On the other hand, we could work it like some medical journal articles where we have 10 co-authors for a 10 sentence paper. Very Happy
Back to top
View user's profile Send private message Send e-mail Visit poster's website Yahoo Messenger
gabefarkas
Joined: 31 Dec 2004
Posts: 1313
Location: Durham, NC
PostPosted: Fri Sep 09, 2005 8:21 am Post subject: Reply with quote
happy to help in whatever way -- gfarkas@gmail.com
Back to top
View user's profile Send private message Send e-mail AIM Address
Mountain
Joined: 13 Mar 2007
Posts: 1527
PostPosted: Mon Dec 17, 2007 9:32 pm Post subject: Reply with quote
Is the research project / paper suggested in this thread still a possibility sometime in future? A good next step after past APBR member papers groundwork? Would running it as suggested then still be the approach favored today or modify?
Back to top
View user's profile Send private message
Dan Rosenbaum
Joined: 03 Jan 2005
Posts: 541
Location: Greensboro, North Carolina
PostPosted: Tue Dec 18, 2007 12:06 am Post subject: Reply with quote
Mountain wrote:
Is the research project / paper suggested in this thread still a possibility sometime in future? A good next step after past APBR member papers groundwork? Would running it as suggested then still be the approach favored today or modify?
Probably not by me. I just don't have the time right now.