 |
APBRmetrics
The statistical revolution will not be televised.
|
| View previous topic :: View next topic |
| Author |
Message |
asimpkins
Joined: 30 Apr 2006
Posts: 245
Location: Pleasanton, CA
|
 Posted: Fri May 04, 2007 11:18 am Post subject: |
 |
|
| Flint wrote: | | I appreciate the value of creating an open shot, but it still doesn't make any sense to me to say Iverson increases his teams efg despite shooting well below it. What you seem to be implying is that if you replaced Iverson wiith a player who took as many shots while shooting at the team average, the teams Efg will still decline. That's just strange. Honestly, that just seems illogical. |
I hate to be the third person to jump in on this point, but you may find this fact to be far-fetched, but it certainly isn't illogical. Iverson's lower eFG% does lower his team's average. Obviously. But his ability to warp the defense, to draw the best defenders, to create open looks for his teammates, and to take tough end-of-shot-clock shots in place of his teammates increases the eFG% of his teammates.
So we have two factors. One positive and one negative. If the positive one out weighs the negative one, then we have a net positive increase. It's not illogical. It might not be true -- maybe players do mostly play in isolation from each other like you suggest -- but it is logically sound.
My stance on Iverson. He certainly wouldn't be the anywhere near my first choice to start a team with, but he's far above an average player. His ultimate value depends on, probably more than most players, how well his teammates complement him. That said, I think he has the raw talent to be an even better player if his mental approach to the game improved. I wonder if a lot of his critics get caught up on that -- noticing his mistakes and how much better he could be, but taking for granted all the stuff he is doing right. |
|
| Back to top |
|
 |
Flint
Joined: 25 Mar 2007
Posts: 112
|
| Posted: Fri May 04, 2007 11:29 am Post subject: |
|
|
I did think it was a great post. And I have added Basketball on Paper to my cart. I wish I could respond at similar length, because there are a lot of things to say, but I am just too busy at the moment. Suffice to say that I dont agree with the argument, that my logic would follow Berri's in that first link, but that I have total respect for Oliver's work from what I know of it, and I shouldn't really comment until I have read it.
Bottom line though is that I think it's correct to say that Mutombo, Mckie, Snow, Lynch and Hill had more to do with making Iverson look good, then the other way around. I think scorers and role players are complements. They have a symbiotic relationship, and are equally valuable. A role player like David Lee is more important than a "star" scorer like Marbury or Iverson. He does that whole rebounding and put back thing a lot better than anyone in the league, which though you basically dismiss it it is in an incredibly valuable skill. Shawn Marion barely touches the ball during a game. Does this make him less valuable? Is it possible that low usage players "create space" for high usage players to operate in, just as much as the other way around?
You could equally say that Iverson needs the ball a lot to be productive, and unless you find a bunch of players who dont need the ball to be successful, you wont win many games. Does that make sense?
Bottom line, elite players do what scorers do better than most scorers, and what role players do better than most role players. Iverson doesn't match that description. |
|
| Back to top |
|
|
jeffpotts77
Joined: 18 Feb 2005
Posts: 150
Location: Cambridge, MA
|
| Posted: Fri May 04, 2007 11:48 am Post subject: |
|
|
Removed my post because I didn't want to distract from the discussion.
Last edited by jeffpotts77 on Fri May 04, 2007 12:05 pm; edited 1 time in total |
|
| Back to top |
|
|
Ben F.
Joined: 07 Mar 2005
Posts: 391
|
| Posted: Fri May 04, 2007 11:53 am Post subject: |
|
|
| Flint wrote: | | I did think it was a great post. And I have added Basketball on Paper to my cart. I wish I could respond at similar length, because there are a lot of things to say, but I am just too busy at the moment. Suffice to say that I dont agree with the argument, that my logic would follow Berri's in that first link, but that I have total respect for Oliver's work from what I know of it, and I shouldn't really comment until I have read it. |
Which part of the argument do you not agree with? The final conclusion, or the part where I say: "If the teammates were to give him the ball every time down the floor and refuse to take any shots, he would NOT shoot 90%. If we can agree on this we're essentially on the same page."?
| Flint wrote: | | Bottom line though is that I think it's correct to say that Mutombo, McKie, Snow, Lynch and Hill had more to do with making Iverson look good, then the other way around. I think scorers and role players are complements. They have a symbiotic relationship, and are equally valuable. |
If you stopped here I think most would agree with you as well. That's what I was getting at in my post - that there is a relationship going on that goes beyond "Player X shoots 60% therefore his value is Y." Those numbers are dictated by context, so to understand them you have to understand the context.
| Flint wrote: | | A role player like David Lee is more important than a "star" scorer like Marbury or Iverson. He does that whole rebounding and put back thing a lot better than anyone in the league, which though you basically dismiss it it is in an incredibly valuable skill. |
See, now we get into a tougher place - I was not arguing about the importance of any player to their team, but rather about the simple relationship of usage and efficiency. As I said before, if we can agree on the first point, about the offensive rebounder shooting worse when asked to move outside of his role, then we agree about that relationship. You can still think Iverson doesn't have as much value as people think, because you might say that optimizing the offense around him isn't as good as most believe. But the relationship still exists. (And for the record, I wasn't trying to dismiss offensive rebounding - I had actually included a sentence in there about how clearly the player had to have great position and be able to put it back in quickly, and that other players might not be able to do it as well, but I felt it took away from the main argument. That being said, there's no denying that a put back look off of an offensive rebound is close to the highest percentage shot there is.)
| Flint wrote: | | Shawn Marion barely touches the ball during a game. Does this make him less valuable? Is it possible that low usage players "create space" for high usage players to operate in, just as much as the other way around? You could equally say that Iverson needs the ball a lot to be productive, and unless you find a bunch of players who dont need the ball to be successful, you wont win many games. Does that make sense? |
Again, getting into an area of value that I'm trying to stay away from. Right now I'm just trying to present the logic of the argument. But I would say that you're right on with this point, and I myself am not convinced that the relationship works backwards as well. In other words, it makes logical sense to me to think that if Chuck Hayes were asked to do more in the offense that he would become much less efficient. It doesn't necessarily make sense to me to say that if Iverson were asked to do less in the offense he would become more efficient, for exactly the reason you specify.
I'd love it if Dean Oliver or anyone else could chime in on that subject, but that's something I thought about with regards to the Denver trade.
I think, however, that most would agree with you that Iverson needs the right teammates around him just as much as role players need Iverson. That, in my mind, is what the whole process of team building is about, why it fascinates me. A lot of players can be horrible in one context and great in another, the trick is fitting all the parts together. |
|
| Back to top |
|
|
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
|
| Posted: Fri May 04, 2007 12:13 pm Post subject: |
|
|
| Ben F. wrote: | ...
Again, getting into an area of value that I'm trying to stay away from. Right now I'm just trying to present the logic of the argument. But I would say that you're right on with this point, and I myself am not convinced that the relationship works backwards as well. In other words, it makes logical sense to me to think that if Chuck Hayes were asked to do more in the offense that he would become much less efficient. It doesn't necessarily make sense to me to say that if Iverson were asked to do less in the offense he would become more efficient, for exactly the reason you specify.
I'd love it if Dean Oliver or anyone else could chime in on that subject, but that's something I thought about with regards to the Denver trade.
I think, however, that most would agree with you that Iverson needs the right teammates around him just as much as role players need Iverson. That, in my mind, is what the whole process of team building is about, why it fascinates me. A lot of players can be horrible in one context and great in another, the trick is fitting all the parts together. |
The skill curves I presented in BoP have different shapes (slopes) and those slopes illustrate how much guys can ramp up efficiency by reducing their usage. Those slopes are definitely different for different players, meaning that cutting player Xs usage by 2% may not increase his efficiency as much as cutting player Ys usage by 2%. Understanding why they're different -- what are the underlying context or tactical reasons -- is part of that fit process.
It's like any strategy game, chess being a convenient example -- you can lose with better pieces, you can win with worse. Part of it is fitting the pieces together, getting a good game plan, and hoping you match up right.
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers. |
|
| Back to top |
|
|
Ben F.
Joined: 07 Mar 2005
Posts: 391
|
| Posted: Fri May 04, 2007 1:03 pm Post subject: |
|
|
| HoopStudies wrote: | | The skill curves I presented in BoP have different shapes (slopes) and those slopes illustrate how much guys can ramp up efficiency by reducing their usage. Those slopes are definitely different for different players, meaning that cutting player Xs usage by 2% may not increase his efficiency as much as cutting player Ys usage by 2%. Understanding why they're different -- what are the underlying context or tactical reasons -- is part of that fit process. |
Right, the curves make sense to me but I was under the impression that the extreme low-usage end of players like Iverson's curves were largely extrapolated from the general trend, since finding data on games where Iverson's usage was actually that low is near impossible (and even if found the sample size is too small to draw an actual conclusion from it).
It's just that while I can understand the logic of: "When these players can pick their spots more they take higher percentage shots," from observation I think it's only true to a certain extent. If Iverson didn't have to take shots at the end of the shot clock his efficiency would increase, but I don't think it would help him at all to become a role player jump shooter - my feeling is that forcing him out of his comfort level of high usage would actually decrease his efficiency. This would make his skill curve almost bell curve shaped. I was wondering if you have data to back up either side of it, or simply a logical counter-argument. |
|
| Back to top |
|
|
Neil Paine
Joined: 13 Oct 2005
Posts: 774
Location: Atlanta, GA
|
| Posted: Fri May 04, 2007 1:25 pm Post subject: |
|
|
| Dean - In this thread, I said I hoped that there might be indicators in a player's skill set (his size, athleticism, ballhandling ability, etc.) that could predict how much usage a player could add without a loss in efficiency. In your work with skill curves, have you found any specific indicators (in a player's stats or physical profile) that directly influence the shape of a player's skill curve? |
|
| Back to top |
|
|
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
|
| Posted: Fri May 04, 2007 1:39 pm Post subject: |
|
|
| Ben F. wrote: | | HoopStudies wrote: | | The skill curves I presented in BoP have different shapes (slopes) and those slopes illustrate how much guys can ramp up efficiency by reducing their usage. Those slopes are definitely different for different players, meaning that cutting player Xs usage by 2% may not increase his efficiency as much as cutting player Ys usage by 2%. Understanding why they're different -- what are the underlying context or tactical reasons -- is part of that fit process. |
Right, the curves make sense to me but I was under the impression that the extreme low-usage end of players like Iverson's curves were largely extrapolated from the general trend, since finding data on games where Iverson's usage was actually that low is near impossible (and even if found the sample size is too small to draw an actual conclusion from it).
It's just that while I can understand the logic of: "When these players can pick their spots more they take higher percentage shots," from observation I think it's only true to a certain extent. If Iverson didn't have to take shots at the end of the shot clock his efficiency would increase, but I don't think it would help him at all to become a role player jump shooter - my feeling is that forcing him out of his comfort level of high usage would actually decrease his efficiency. This would make his skill curve almost bell curve shaped. I was wondering if you have data to back up either side of it, or simply a logical counter-argument. |
You're hinting at how you'd actually figure things out...
But, no, skill curves are built only to be monotonic. In reality, contextual changes can make the data appear to be non-monotonic (if I recall, Iverson's raw data from 2002 suggested that he was more efficient in high usage games, but you can then break those down and see why that was suggested but not a fair comparison). The work you have to do to produce skill curves is not trivial, but it also builds a big understanding of why things happen...
So I have lots of experience with data working through the reasons why the raw data goes up when a skill curve goes down. The reasons range from different competition (some guys just kill bad Ds then drop off a cliff against good ones, where a skill curve is against an average D) to the tactics of opponents (a guy may get a lot of good shots and be efficient but that's because opponents know that it's still the best way to slow down the team as a whole). It has happened where I can't explain it. It's rare, but not ridiculously so. Those are often the guys, too, where the algorithm generates a skill curve that doesn't make sense. Then the next year, they'll be fine and normal. But not bell curve shaped 
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers. |
|
| Back to top |
|
|
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
|
| Posted: Fri May 04, 2007 1:41 pm Post subject: |
|
|
| davis21wylie2121 wrote: | | Dean - In this thread, I said I hoped that there might be indicators in a player's skill set (his size, athleticism, ballhandling ability, etc.) that could predict how much usage a player could add without a loss in efficiency. In your work with skill curves, have you found any specific indicators (in a player's stats or physical profile) that directly influence the shape of a player's skill curve? |
I forgot about that. And, while an interesting idea with relevance to scouting, I can't say I've done it. It's hard. How do you quantify "shape" of a curve? Quantifying athleticism and ballhandling ability is hard, too, though there are attempts. There are enough questions in how I'd do it that I haven't done it.
Ahem, I am looking for interns this summer. If people are interested, feel free to email me.
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers. |
|
| Back to top |
|
|
Statman
Joined: 20 Feb 2005
Posts: 242
Location: Arlington, Texas
|
| Posted: Fri May 04, 2007 2:05 pm Post subject: |
|
|
| Harold Almonte wrote: |
4.- A.I. tend to concentrate all the scoring entrophy on himself, then his teammates tend to get a better FG% playing with him.
|
I like how you phrased this - and I think it's pretty much true. I had been trying to say this in a more round about way. |
|
| Back to top |
|
|
Statman
Joined: 20 Feb 2005
Posts: 242
Location: Arlington, Texas
|
| Posted: Fri May 04, 2007 2:43 pm Post subject: |
|
|
| Flint wrote: |
Bottom line though is that I think it's correct to say that Mutombo, Mckie, Snow, Lynch and Hill had more to do with making Iverson look good, then the other way around. |
I don't agree. All those low usage players pretty much ensure that AI takes more shots - and score more points - so THAT would be true. However - playing with nothing but below average usage players will hurt any and every player's shooting efficiency overall (not to mention lower his assist totals and increase his turnovers). A lower shooting efficiency doesn'y make AI look "good" in your eyes - it makes him look worse.
As for the "role" players - all of them played better statistically (especially shooting efficency) WITH Iverson than they did pre Iverson or post Iverson. Heck - the fact that you even stated that you thought a number of these guys (all of whom outside of Mutambo would be looked at as below average NBA players when looking at their careers without Iverson) had a better year than Iverson that year kinda shows me how he made them "look better" than they made him.
Once again looking at that year - Iverson had to take pretty much all the tough shots - being completely surrounded by low usage guys - ALL season. This would be even worse against good defenses. He still somehow had the same ts% than the rest of the players on the team - who got to pick and choose when they shot (often dunks, put backs, & wide open jumpers) because non of them were skilled enough offensively to burden more of the load. Iverson also turned the ball over MUCH less than the rest of the "role" players you mention in relation to usage.
He also played many more minutes than anyone else on the team per game.
So - how did ANY of these players have a better season than him? Do you seriously think that team would have been BETTER without Iverson than, say, Lynch or McKie?? |
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
Powered by phpBB © 2001, 2005 phpBB Group
|
FAQ
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
