Author Message
huevonkiller
Joined: 25 May 2010
Posts: 15
Location: Miami, Fl
PostPosted: Fri Jun 18, 2010 8:54 pm Post subject: Offensive Win Shares vs other barometers. Reply with quote
I was reading a very interesting blog post at basketball-reference recently. It lead me to wonder about the accuracy of Paine's o-rating/%pos method versus Offensive Win Shares. Neil Paine seems to rate offensive abilities (in some instances at least), using this conversion:
Quote:
For this post, I want to create a simple lineup efficiency model that combines Dean and Eli's findings -- specifically Eli's tradeoff for average players (+/-1.25 in a 107.5 ORtg league), but Dean's distinction between high-usage, mid-usage, and low-usage effects on personal efficiency (the effect on low-usage players is twice the effect on high-usage ones). What we in essence have, then, is a simple algebra problem: let x be the tradeoff for low-usage (<=18%) players and y the tradeoff for high-usage ones (>=23%). (x + y) / 2 = 1.25, and x = 2*y. What are x and y?
x = 1.6667, y = 0.8333
This means that in a 107.5-ORtg environment, the efficiency trade-off for increasing or decreasing usage by 1% is as follows:
Player Type Tradeoff
High Usage (>=23%) 0.833
Mid Usage (18-23%) 1.250
Low Usage (<=18%) 1.667
Knowing this, he wrote the following piece on the Greatest Laker of all time:
http://www.basketball-reference.com/blog/?p=6269
However by applying Offensive Win Shares (which is Justin Kubatko's formula I found out) one would get various different conclusions if they wanted to determine the Greatest Laker.
A point was raised in the Greatest Laker discussion that Eli's study has flaws, and in Kobe's case he has not shown the ability to trade an extremely high usage rate for a Pau Gasol-like Offensive Rating. Historically this trend is rather consistent and continues in the playoffs. He simply plateaus at about a 115 offensive rating at best. He has a 111 o-rating on 31.7 usage rate from ages 22-31 in the post-season. 114 O-rating on 32.7 usage rate from ages 27-31. Curiously Wade also seems to have this problem.
I then read the Calculating Win Shares page again, and found this:
Quote:
2. Calculate offensive possessions for each player. James had an estimated 1928.1 offensive possessions in 2008-09.
3. Calculate marginal offense for each player. Marginal offense is equal to (points produced) - 0.92 * (league points per possession) * (offensive possessions). For James this is 2345.9 - 0.92 * 1.083 * 1928.1 = 424.8. Note that this formula may produce a negative result for some players.
4. Calculate marginal points per win. Marginal points per win reduces to 0.32 * (league points per game) * ((team pace) / (league pace)). For the 2008-09 Cavaliers this is 0.32 * 100.0 * (88.7 / 91.7) = 30.95.
5. Credit Offensive Win Shares to the players. Offensive Win Shares are credited using the following formula: (marginal offense) / (marginal points per win). James gets credit for 424.8 / 30.95 = 13.73 Offensive Win Shares.
http://www.basketball-reference.com/about/ws.html
It seems that Justin has indeed accounted for %pos as well. What problems are there with Offensive Win Shares then, because I thought it was the bread and butter at basketball-reference for a reason?
In the 2009 post-season, Pau Gasol didn't have the most OWS on the Lakers even with the higher O-rating and comparable minutes played. I thought the usage-efficiency tradeoff was always taken into consideration by Offensive Win Shares.
I suspect Eli's study has problems with extremely high usage rate players. Please shed more light on which metric is superior, preferable, however you want to describe it.
Last edited by huevonkiller on Sat Jun 19, 2010 2:43 pm; edited 1 time in total
Back to top
View user's profile Send private message
Kevin Pelton
Site Admin
Joined: 30 Dec 2004
Posts: 979
Location: Seattle
PostPosted: Fri Jun 18, 2010 9:13 pm Post subject: Re: Offensive Win Shares vs other barometers. Reply with quote
huevonkiller wrote:
A point was raised in the Greatest Laker discussion that Eli's study has flaws ...
I'm not exactly seeing that.
Quote:
... in Kobe's case he has not shown the ability to trade an extremely high usage rate for a Pau Gasol-like Offensive Rating. Historically this trend is rather consistent and continues in the playoffs. He simply plateaus at about a 115 offensive rating at best. He has a 111 o-rating on 31.7 usage from ages 22-31 in the post-season. 114 O-rating on 32.7 usage from ages 27-31. Curiously Wade also seems to have this problem.
Well, that's a rather debatable "problem." I would love to have a problem like that on my team.
I think the important thing to remember is there are two different issues at play here. One is the individual player's skill curve (usage-efficiency tradeoff), and one is the general skill curve.
For players like Bryant and Wade, and DeanO shows this in Basketball on Paper, their efficiency is relatively flat compared to their usage. For role players, the skill curve is very steep. Ask them to usage a few more possessions and it changes their efficiency substantially.
Now, to the extent this is a problem, I think it's like this--guys like Bryant and Wade are at their most valuable when they're alongside players like Smush Parker and Joel Anthony, not surrounded by a lot more offensive talent. But obviously that's not going to maximize team performance, unless it's a situation like the Iverson Sixers where you get four defensive players and just let one guy do all the scoring.
When we're looking to value the importance of their high usage rates, however, I would argue that this is irrelevant. The extra possessions they're taking on are valuable because of how they impact their teammates' efficiency, not so much their own.
As for which measure is the best, you're looking for certainty that doesn't exist with our current statistics. Neil's method probably gets you 95% of the way there, and so do Offensive Win Shares. It just so happens they're not always exactly the same 95%.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
huevonkiller
Joined: 25 May 2010
Posts: 15
Location: Miami, Fl
PostPosted: Fri Jun 18, 2010 9:28 pm Post subject: Re: Offensive Win Shares vs other barometers. Reply with quote
Kevin Pelton wrote:
I'm not exactly seeing that.
Well then, I was trying to discuss the nuances. No disrespect meant to the study in general. Surely there could be instances where it is off ("95%" accuracy I suppose?).
Quote:
Well, that's a rather debatable "problem." I would love to have a problem like that on my team.
I think the important thing to remember is there are two different issues at play here. One is the individual player's skill curve (usage-efficiency tradeoff), and one is the general skill curve.
For players like Bryant and Wade, and DeanO shows this in Basketball on Paper, their efficiency is relatively flat compared to their usage. For role players, the skill curve is very steep. Ask them to usage a few more possessions and it changes their efficiency substantially.
Now, to the extent this is a problem, I think it's like this--guys like Bryant and Wade are at their most valuable when they're alongside players like Smush Parker and Joel Anthony, not surrounded by a lot more offensive talent. But obviously that's not going to maximize team performance, unless it's a situation like the Iverson Sixers where you get four defensive players and just let one guy do all the scoring.
When we're looking to value the importance of their high usage rates, however, I would argue that this is irrelevant. The extra possessions they're taking on are valuable because of how they impact their teammates' efficiency, not so much their own.
As for which measure is the best, you're looking for certainty that doesn't exist with our current statistics. Neil's method probably gets you 95% of the way there, and so do Offensive Win Shares. It just so happens they're not always exactly the same 95%.
Thanks that is a great response.
Back to top
View user's profile Send private message
eyriq
Joined: 04 Jun 2008
Posts: 54
Location: Orlando
PostPosted: Sat Jun 19, 2010 8:12 pm Post subject: Reply with quote
That was a fascinating read, and you bring up good points. This leads me to a question: I saw him use Z-Scores for determining best defenses and offenses of all time, so why not do this here as well? I've started to do this on my own, just as a way to interpret ORtg and especially DRtg as something more transferable, and I guess I'm also wondering if this is a flawed analysis.
Back to top
View user's profile Send private message
DSMok1
Joined: 05 Aug 2009
Posts: 608
Location: Where the wind comes sweeping down the plains
PostPosted: Mon Jun 21, 2010 8:52 am Post subject: Reply with quote
eyriq wrote:
That was a fascinating read, and you bring up good points. This leads me to a question: I saw him use Z-Scores for determining best defenses and offenses of all time, so why not do this here as well? I've started to do this on my own, just as a way to interpret ORtg and especially DRtg as something more transferable, and I guess I'm also wondering if this is a flawed analysis.
I don't like the Z-Scores, as I don't think the spread of the rest of the league is that important (unless you are looking at "dominance"). I'd like at offensive/defensive rating percentage points from league average as a pretty good proxy for how good a team was on each end. If league average was similar, then just points above or below average. The spread of the rest of the league, I think, is not very relevant.