RealGM

I haven't read Basketball On Paper, so I'm not totally familiar with the formula for what goes into Ortg. Basketball-reference defines Ortg as "Offensive Rating (available since the 1977-78 season in the NBA); for players it is points produced per 100 posessions"

True Shooting % is defined in the same place "True Shooting Percentage; the formula is PTS / (2 * (FGA + 0.44 * FTA)). True shooting percentage is a measure of shooting efficiency that takes into account field goals, 3-point field goals, and free throws."

Isn't this just points per possession divided by 2, though? To wit: player scores 2 points on 1-2 shooting, that's 2 points in 2 possessions, or 1 point per possession. Yeah? TS% put that at 2/(2*2), or .50. It makes sense as a shooting % because you receive two points for one basket, so .50 is the same as 1 point per possesion. Which would be an Ortg of 100, right?

It seems like different ways of expressing the same relationship, at least the way Ortg is defined. Is this accurate? Is there some straight line formula you can draw to calculate Ortg from TS%, and vice versa? Or do they take different approaches to measuring the same thing, leading to slightly different results?

Ortg also includes assists and offensive rebounds.

Doctor MJ wrote:Ortg also includes assists and offensive rebounds.

ahhh...I was wondering why those #s seemed higher than was TS% would predict

Does that (or any stat for that matter) try to account for a what I guess I'd call a failed assist attempt?

That is, players get credited positively if they kick out to a 3-pt shooter who sticks the jumper with an assist (and presumably this ups their Ortg as well), but there seems to be no statistical measure to take into account when a player "uses" the ball to dish to a shooter who then misses. Is that any less of a possession used than an assist?

I've always wondered about that. Would also be interested to see fg% and TS% and points per possession numbers on things like that...how efficiently players score when being dished to by say, Steve Nash, vs how they score when either unassisted or being dished to by another player.

Does that make sense?

The definition of a possession used on basketball-reference.com is slightly weird (see this Wikipedia article: http://en.wikipedia.org/wiki/APBRmetrics)

Among the growing list of APBRmetric basketball statistics here are some of the most important ones gaining increased usage:
Offensive Rating/Offensive Efficiency and Defensive Rating/Defensive Efficiency, on a team level, are calculated as points scored and points allowed per 100 possessions. Possessions are usually estimated by the following formula:
Possessions = .96 * (FGA − ORb + TO + (.44 * FTA))

So if you get blocked a lot, but grab your own rebound, then your PPP will be fairly high. I dunno, this has never sat well with me..

Ripp wrote:The definition of a possession used on basketball-reference.com is slightly weird (see this Wikipedia article: http://en.wikipedia.org/wiki/APBRmetrics)

Possessions = .96 * (FGA − ORb + TO + (.44 * FTA))

So if you get blocked a lot, but grab your own rebound, then your PPP will be fairly high. I dunno, this has never sat well with me..

How so? The amount of possessions doesn't change with an ORB.

mysticbb wrote:

Ripp wrote:The definition of a possession used on basketball-reference.com is slightly weird (see this Wikipedia article: http://en.wikipedia.org/wiki/APBRmetrics)

Possessions = .96 * (FGA − ORb + TO + (.44 * FTA))

So if you get blocked a lot, but grab your own rebound, then your PPP will be fairly high. I dunno, this has never sat well with me..

How so? The amount of possessions doesn't change with an ORB.

You are right, but this stat (which has the goal of estimating the number of possessions used by a team or player) doesn't seem to account for that, at least at the player level. If I'm blocked 10 times in one play, but keep grabbing the OReb and eventually score, the estimate for the number of possessions I've used that day is pretty badly off..

^--- Actually, I guess I'm wrong....when you get blocked, that counts as a FGA, right? So if you get blocked and grab your own rebound, they negate each other...

Ripp wrote:You are right, but this stat (which has the goal of estimating the number of possessions used by a team or player) doesn't seem to account for that, at least at the player level. If I'm blocked 10 times in one play, but keep grabbing the OReb and eventually score, the estimate for the number of possessions I've used that day is pretty badly off..

Ripp wrote:^--- Actually, I guess I'm wrong....when you get blocked, that counts as a FGA, right? So if you get blocked and grab your own rebound, they negate each other...

Yes, when someone gets blocked, the shot attempt still counts as FGA. Thus the scoring efficiency would be incredible low. In terms of possessions a FGA and a ORB negate each other. Which is logical, because the team is still in possession of the ball after a shot attempt and the offensive rebound.

mopper8 wrote:Does that (or any stat for that matter) try to account for a what I guess I'd call a failed assist attempt?

That is, players get credited positively if they kick out to a 3-pt shooter who sticks the jumper with an assist (and presumably this ups their Ortg as well), but there seems to be no statistical measure to take into account when a player "uses" the ball to dish to a shooter who then misses. Is that any less of a possession used than an assist?

I've always wondered about that. Would also be interested to see fg% and TS% and points per possession numbers on things like that...how efficiently players score when being dished to by say, Steve Nash, vs how they score when either unassisted or being dished to by another player.

Does that make sense?

Oliver's ORtg and DRtg are based on reading box scores not play by play, so it's not doing anything as sophisticated as that. it's for this reason, particularly with the DRtg, that he advises people to take them with a grain of salt.

Individual ORating/DRating doesn't mean much. But when you're talking about a team or a lineup's O/D Rating, it's helpful.

TS% is the best measure for shooting efficiency.

Now I remember the example that made me dislike the above possessions formula:

Suppose I only live off of Orbs, that is the only way I score. Say I score at 60%, never get fouled, never turn it over, etc.

TS% will say I produce 1.2 point per possesion, PPP (where individual possessions will be estimated by the formula linked above) will say I use zero offensive possessions, and thus have infinite efficiency.

So yeah, the possessions formula at an individual level is screwy.

Actually the formula you listed isn't for calculating individual possessions, but for getting an estimation of the teams total possessions.

If a player is scoring off a ORB he still is using an individual possession and his ORtg wouldn't be infinite.

The formula for individual possessions is:

Possessions = Scoring Possessions + Missed FG Part + Missed FT Part + TOV

A description of that in more detail can be found in Appendix 1 of Dean Oliver's "Basketball On Paper".

Ah, I see. Hmm, I've been meaning to order that book...maybe I should do so.

Do it, BOP is a great read.

tsherkin wrote:Do it, BOP is a great read.

Yup, Oliver has the best combination of statistical smarts, basketball knowledge, and common sense I see in the statistical community.

Yeah, wish I'd bought the book woulda saved me some time over the past coule of days. We have a pretty interesting discussion going on now in the RealGM Raptors forum (viewtopic.php?f=32&t=1047850), where we are discussing some of the merits of Oliver's PDSS and Drtg methodology.

Thread is crazy long, but might be fun to read parts of it when you have the chance. Page 1 and then starting again on page 24 and 25 is where it starts to get interesting. I also have a post on page 39 that might be interesting to read.

Ripp wrote:Yeah, wish I'd bought the book woulda saved me some time over the past coule of days. We have a pretty interesting discussion going on now in the RealGM Raptors forum (viewtopic.php?f=32&t=1047850), where we are discussing some of the merits of Oliver's PDSS and Drtg methodology.

Thread is crazy long, but might be fun to read parts of it when you have the chance. Page 1 and then starting again on page 24 and 25 is where it starts to get interesting. I also have a post on page 39 that might be interesting to read.

I should post in the Raptors thread from time to time, lol.