The difference between a players usage rate in Hollingers page and yours is always around 2
I'm calculating them differently now. Please see my definitions for Usg% and TOV%.
I see, is there a reason why you stopped counting assists as possessions?
Not really digging these new percentages, I prefer knowing the percentage of a players personal possessions, than what his team does while hes on the floor.
Im not sure what assist % means either, percentage of teammate field goals assisted while on the court. What exactly does this show and why the change in philosophy? I mean the only one you havent changed is Reb Rate, I was half expecting it to be % of team rebounds corralled when a player is on the floor.
They seem useful but I guess I'll have to do my own work for the rest of the possession stats
Joined: 05 Jan 2005 Posts: 603 Location: Columbus, OH
Posted: Tue Jan 22, 2008 7:21 am Post subject:
Chronz1 wrote:
Not really digging these new percentages, I prefer knowing the percentage of a players personal possessions, than what his team does while hes on the floor.
Im not sure what assist % means either, percentage of teammate field goals assisted while on the court. What exactly does this show and why the change in philosophy?
I find it extremely useful. As an example, Steve Nash's assist percentage in 2006-07 was 50.1%. In other words, when Nash was on the court there was about a 50% chance that he got the assist when a teammate scored. To me that's more interesting than knowing he had 39.3 assists per 100 (FGA + 0.44*FTA + AST + TOV) (the old formula).
Chronz1 wrote:
I mean the only one you havent changed is Reb Rate, I was half expecting it to be % of team rebounds corralled when a player is on the floor.
Yeah, but I added offensive and defensive rebound percentages, so that had to be disappointing. :-) _________________ Regards,
Justin Kubatko
Basketball-Reference.com
Joined: 21 Jan 2005 Posts: 46 Location: Montevideo, Uruguay
Posted: Tue Jan 22, 2008 9:49 am Post subject:
Yes, but the new assist rate is basically the old assist rate*usage, right? (Not mathematically, I know, but that are the relevant factors) I do not really see the improvement (there seems to be less information there to me), and it doesn't really measures passing ability (or propensity to pass). It's more of a team context thing (how much a team depends on X), right?
Joined: 13 Oct 2005 Posts: 507 Location: Atlanta, GA
Posted: Tue Jan 22, 2008 11:18 am Post subject:
I still don't understand why people are so hung up on the old definition of "assist rate"... It merely measured assists as a fraction of a player's individual possessions consumed (and it didn't even do that right -- the denominator was not technically individual possessions); I fail to see how this measures passing ability as much as passing tendency.
If a large % of Player A's possessions are devoted to assists, it could very well mean that he was a good passer... but it could just as easily mean that he simply never shot, never dribbled, never drew fouls, etc. He might not be a good passer, but if all you do is pass, eventually you'll get assists, a high fraction of your poss. will be assists, and you'll look like a good passer by this metric even if you're not creating much of anything for your teammates.
IMHO, the new formula is much better to determine passing ability. Yes, it's technically measuring a player's passing role in the offense -- but doesn't that say something about passing ability in and of itself? And in the new formula, you've got to be actively getting assists and making plays while on the floor to look like a good passer, as opposed to just concerning yourself with devoting a large % of your own possessions to assists.
The only issue is that it doesn't really take into account the quality of one's passes; by simply using made FG by teammates, you don't really capture: A) passes that didn't turn into baskets, which could at least partially be the passer's fault, and B) "incomplete" passes that result in turnovers. You might say the solution to this brings us back to the old definition, since it at least takes into account turnovers... but I say why not take the Pomeroy definition one step further and calculate individual assists as a fraction of team possessions while on the floor? This would indirectly capture the quality of the passing, and still show how active the player is in the offense as a passer.
Joined: 13 Oct 2005 Posts: 507 Location: Atlanta, GA
Posted: Tue Jan 22, 2008 5:25 pm Post subject:
jkubatko wrote:
And here's where the train jumps the track...
Ha, yeah, we've been in this same exact place before (without a resolution), so sorry for hijacking the thread to go back there again. But just the same, it is a (somewhat) relevant part of the B-R.com discussion, because the plain truth is that your site is the most visible NBA stat site on the web, and the odds are good that whichever version you choose to put on your pages automatically becomes the commonly-accepted definition.
Never bothered looking into Pomeroy's rates but if thats where all this change stems from then it must be worth checking out.
Quote:
I find it extremely useful. As an example, Steve Nash's assist percentage in 2006-07 was 50.1%. In other words, when Nash was on the court there was about a 50% chance that he got the assist when a teammate scored. To me that's more interesting than knowing he had 39.3 assists per 100 (FGA + 0.44*FTA + AST + TOV) (the old formula).
You may be right, I need to familiarize myself with these rates before I can really comment on them, but I like what asimpkins had to say in the thread: In an ideal world, I'd like both of them to be offered, but if I can only have one then I think Pomeroy's answers a more important question.
Im not sold on assists not having to count in usage rates but thats for another thread and Ill keep an open mind, you've given me a great reason to look through b-r's databasa
Quote:
Yeah, but I added offensive and defensive rebound percentages, so that had to be disappointing.
Ha, yeah, we've been in this same exact place before (without a resolution), so sorry for hijacking the thread to go back there again. But just the same, it is a (somewhat) relevant part of the B-R.com discussion, because the plain truth is that your site is the most visible NBA stat site on the web, and the odds are good that whichever version you choose to put on your pages automatically becomes the commonly-accepted definition.
Bingo, I wouldnt go as far as automatically but its because of this change that Ill even bother to look into them. But to tell you the truth, before this thread I havent really heard much about them
Joined: 05 Jan 2005 Posts: 603 Location: Columbus, OH
Posted: Wed Jan 23, 2008 11:20 am Post subject:
davis21wylie2121 wrote:
But just the same, it is a (somewhat) relevant part of the B-R.com discussion, because the plain truth is that your site is the most visible NBA stat site on the web, and the odds are good that whichever version you choose to put on your pages automatically becomes the commonly-accepted definition.
Joined: 03 Jan 2005 Posts: 728 Location: Washington, DC
Posted: Wed Jan 23, 2008 12:05 pm Post subject:
One thought on usage percentage -- Why not just use Dean O's possession formulation? You're calculating it already to get ortg. Or is this formulation pretty close to Dean's anyway? _________________ My blog
IMHO, the new formula is much better to determine passing ability. Yes, it's technically measuring a player's passing role in the offense -- but doesn't that say something about passing ability in and of itself? And in the new formula, you've got to be actively getting assists and making plays while on the floor to look like a good passer, as opposed to just concerning yourself with devoting a large % of your own possessions to assists.
What about players whos primary role isnt to distribute, has multiple good passers on the team or isnt required score in general, a guy like Ben Wallace. Yea hes not going to dominate offensively but I could atleast understand that the one thing he could do is find the open man/cutter.
This is my stance on the subject based on my limited understanding, both have their advantages, one shows how much a team relies on the player, the other shows what a player did with his personal possessions in his own role. I just dont understand why one must be eliminated in favor of the other.
As a rookie Wade didnt play much but I could understand that he made a concentrated effort to involve his teammates when he did play, his personal possessions reflect that.
Did he really improve as a passer, or did he just start getting more playing time? I guess Ill just miss being able to know what a player did, and while knowing how much a team relies on a player is important, so is being able to break down a players individual game. These new percentages seem more like the old rates+MPG.
And throwing out assists from the turnover rate in general doesnt make much sense to me, wouldnt it punish players whos primary role is to pass, they get punished for committing the turnover and losing the possession but absolutely no credit for creating the assist in another possession.
Joined: 05 Jan 2005 Posts: 603 Location: Columbus, OH
Posted: Fri Jan 25, 2008 9:20 am Post subject:
The new rates are not affected by playing time. Wade's assist percentage changed because his role in the offense changed. _________________ Regards,
Justin Kubatko
Basketball-Reference.com
Sorry, I have to complain about the move to "per 36". All it does is ensure that every pundit and fan alike is going to keep debating players based on per game statistics. Per 40 was bad enough. Why not just go with per 48? Per 48 makes since. It's the length of a basketball game. Per 40 or per 36, no matter how confident of a reason you have for using it, is inherently arbitrary. And arbitrariness does not belong in analytical thinking.
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