given that pbp is readily available, not sure why possessions just aren't counted as any time the ball actually changes possession (be it rebound, steal, block w/ retained possession and score) with giving an additional possession for offensive rebounds (still not sure about that part but on the surface it makes sense to me)

L.A. is getting too much love in regards to their chances to repeat as Western Conference Champs and make it to the finals again IMO. I mean, they are a very good team, but the western conference is still going to be STACKED. It's just so much easier to make it through the playoffs in the east. That's why Detroit has played in 5 straight Eastern Conference Finals whereas you don't see that kind of prolonged dominance in the West save except San Antonio, who probably needs some new role players to be competitive.

The thing is, a year ago Kobe had never taken L.A. out of the first round w/o shaq and Boston was coming off a 26 win season. Now we know that Boston will stay mostly intact, and as of right now they are the best team in the league, and only 1 of the other top 5 teams in the league is in the East. They just have a huge advantage there, the East would have to get shaken up a lot for them to not be big favorites to make it back to the finals. The Lakers were the best team in the west, but there are also probably 5 teams out there that are good enough to compete with them in a seven game series. With a healthy Bynum I def. give the edge to the lakers here, but they have it much harder and they are currently probably the 2nd best team in the league, so them being favorites to win the title next year astounds me. Hell I think Boston is the best team in the NBA, but put them in the West and LA in the east and I like L.A. chances to win a title more.

The general definition for possessions counts an offensive rebound as part of a possession, not as the start of a new one. I'd imagine that play by play data isn't used because that only is available for the last few years while box scores are available for the last few decades, since whenever they began tracking turnovers and offensive rebounds.

yeah I realize that I just think it's dumb. It's an extra possession in my mind, and it's why offensive rebounding studs are really helpful in terms of offensive production (AC Green, Dale Davis, and now David Lee)

There's really no excuse, imo. But that's how they do it.

i agree. i haven't thought too hard about it so maybe it's too simple, but when the shot clock resets it should be a new possession

Yeah, but treating them as a new possession would make them less valuable. Let's say I'm a freak of nature who can grab every rebound on the court, but I'm not very good at shooting. So, I average three shots every time down the court before I get one to go. If I count each one as a new possession, I only average 66.67 points per 100 possession, while if I count it as the extension of a possession, I average 200 points per 100 possessions. So treating it as a new possession would make it so good offensive rebounding teams look less efficient than they really are.

that is the rule isn't it?

when you get an offensive board (and the ball hasn't changed sides) it's extending the possession... why is that dumb exactly to keep it as part of the same possession?

you think every offensive board tip-in, which takes exactly 0.3 seconds or so, should be an extra possession?

anyway the offensive board leading to an extra shot attempt is figured in the metric - FGA... it's fine the way it is imo

also wat dudd said imo

Unfortunately, this information doesn't exist, even in the PBP.

In my little project I'm working on, I'm doing it so that possessions are tallied on every shot/turnover/clock-expiration, and I'm not using stupid best-fit scaling to estimate FTs and rebounds and ****. A 2-shot foul is not worth 0.8 possessions; it is worth 1. I'm also not at all concerned about evening the possessions for both teams.

Ahhhh....it's life-tilting just even thinking about how stupid the possession calculation these guys use is.

hmmm.. that's an interesting point, but really you're only making it look worse because you're changing the denominator. It only looks worse because you're applying an old stat to a new way of thinking. Besides, the current formula already factors in offensive rebounds.

I'd expect it to even out across the league. And you'd have another metric, possessions per game (as well as opponent's possessions per game). It'd be neat to see if certain teams constantly out-possess their opponents by 5 or more possessions.

I have to run, but quickly:

A possession off of an offensive rebound is significantly different from your average possession. It is also significantly different than the bricked shot possession that led to the rebound.

So yes, the 0.3s tip-in is an extra possession, and teams that are exceptionally good on the offensive glass will probably have more of these 0.3s possessions, which happen to be very valuable.

It still inserts empty possessions in the denominator, so it only serves to hurt offensive rebounding teams, and it wouldn't even out because different teams have different styles and differing offensive rebounding percentage. If you want to see who gets more offensive rebounds, just compare offensive rebounding, you don't need to mess with a metric like efficiency.

Yes, but those show up in the general definition of efficiency. Even if a team scores a putback on every single offensive rebound, if you treat them as a new possession, it hurts a team's overall efficiency because the league average efficiency is over 100 points per possession, at least by the current definition.

it's not an empty possession if you score on it. I think that's what we're driving at. Though I do see some of your concerns.

what about the times where you don't get a putback and run another offensive set? if you get an offensive rebound you are getting another chance to run another offensive set, no? sounds like a new possession to me, otherwise the shot clock shouldn't reset. so my rationale is that a new clock signifies a new possession and i don't think the ball has to change hands for you to get a new possession.

lol

No, by resetting it is an empty possession. Currently offensive rebounds don't count as an empty possession because they don't reset. But by resetting, it does, because the possession that resulted in the offensive rebound scored zero points, otherwise there wouldn't be a rebound.

tarheel / kc / kbfc -- ok so basically you are looking at the points per 100 possession metric, which is a measuer of offensive / defensive efficiency (e.g. 103 pts per 100 possessions, your established baseline, is more offensively efficient than 98 pts per 100 possessions). if this is correct i am with you.

now, based on this, how does counting an offensive board as a possession (which in general i would think (intuitvely) should icnrese your efficiency)... hellp accurately measure offensive efficiency by incerasing the denominator and creating a less efficient number?

if the current POS formula has a numerator offset somehow, can someeone explain it to me? is it weighted? algorithmic?

uh yeh, which is the (sort of uh obvious?) reasoning as to why the possession denominator increases and (i would think) you'd need a numerator offset to fully measure the efficiency of the team in those "2 possessions" in your mind

It doesn't, the only thing in the numerator is points. I have no idea why they would think treating offensive rebounds as a new possession is a good thing, since possessions primarily are used to measure efficiency.

but for the points per 100 possession metric, the offensive rebound gives you anotehr chance to improve your efficiency by not increasing teh possession total, yeh?

and if efficiency is what you're measuring... i think the metric should try to reflect + / - ?

kbfc - very curious:

i haven't read dean oliver's book in a while... so can you or someone else familiar with this stat (like jofa) explain why it is "******ed"?

erm -- ok looking throu oliver's formula... it looks like he does something liek this:

Possessions = FGA-OR+TO+.4*FTA

meaning he takes total FGA, *subtracts* offensive rebounds (which makes sense *decreasing* the denominator to make the team more efficient), adds in turnovers, and adds some value for a FTA as a possession

what is wrong with this formula (theoretically, not sure how gets to 0.4 though it makes sense it should be <0.5 possession) exaclty?

Sometimes you shoot and 1's, which are already counted by fga, so .4 is an estimate for the ratio ft's to possessions used, plus I'd imagine some correction for the amount of possessions extended by an offensive rebound off of a free throw.

wat? of course it's looking worse cause you're changing the denominator, it's points / possessions -- it's not some mystical formula it's simple math

is this "new way of thinking" about possessions somehow obviating the basic laws of arithmetic?

if the "new way of thinking" is about completely changing the offensive efficiency metric altogether, then talk about that and show us a new formula. one that isn't about having "total poitns" in the numerator and "total possessions" in the denominator.

to me points per 100 possessino seems fine, maybe the possession formula needs tweaking, but the current baseline methodology seems pretty straightforward and makes sense

dudd - gotcha, thanks

in a vacuum, what does "out possessing" a team tell you on any level about the team's perofrmance exactly?

outpossessing a team with no other stat applied to it (to give you an actual performance metric) is 100% meaningless