APBRmetrics

Hi everyone. Recently I was searching for better rating system than PER and then decided to make my own. Then I saw this site. In a couple of days I generated my formula by regression (I first learned what it means here). Used box-scores.

In a couple of days I learned a lot of formulas. Especially specific team possession formula was useful. But I can't seem to find which "player total possession" formula is the most accurate.

I used this one at first:

Player's Total Possessions Formula=(Player's Scoring Possessions)+(Player's Non-Scoring Possessions)=(Player Field Goal Attempts)-(Player Missed Shots)*(Team Offensive Rebounding Percentage)+0.37*(Player Assists)-0.37*(Field Goals Made)*Q/R+(Player Turnovers)+0.4*(Player Free Throw Attempts)

Where;

Q=5*(Player Minutes)*(Team Assist Total)/(Team Total Minutes)-(Player Assists)
R=5*(Player Minutes)*(Team Field Goals Made)/(Team Minutes)-(Player Assists)

I'm not sure about R. It could be

R=5*(Player Minutes)*(Team Field Goals Made)/(Team Minutes)-(Player Field Goals Made)

instead of Player Assists. Still, both ways I can't seem to produce the correct poss. when I compare it with poss.(by using usage and ORTG stats) on B-R.

Is there a more accurate formula or there's something wrong here?

Thought someone else might respond.

Bump.

Player possession formulas are problematic. The fact of the matter is that in defining possessions on a player level you are introducing a bias in your metric that will limit it. Possessions are great for team level metrics, but they don't make sense on the player level. Basketball is a team game, after all.

Player possession formulas are problematic. The fact of the matter is that in defining possessions on a player level you are introducing a bias in your metric that will limit it. Possessions are great for team level metrics, but they don't make sense on the player level.

naysayers - no vision...

bchaikin wrote:naysayers - no vision...

Expand upon this. Have you actually done this analysis on both a team and player level? Do you understand the issue? Do you actually believe that any one player ever 'uses' a possession?

Expand upon this. Have you actually done this analysis on both a team and player level?

yep...

Do you understand the issue?

what i don't understand is why anyone would say that in defining possessions on a player level you are introducing a bias in your metric that will limit it... they don't make sense on a player level...

why would you make such a statement? have you ever charted player possessions in a game? over a season?...

http://www.82games.com/dribbles.htm

Yes I've seen that, it's a great piece. I say it because teams use possessions, players take opportunities.

... teams use possessions, players take opportunities.

This is true most of the time. Sometimes a player just loses the ball or throws up a bad shot.

Still, there is probably some use for the allotment of "possessions used" to individuals.
There's also a great potential for misuse. With "possessions" in the denominator, a player can have a crazy high ORtg, if all he does is park in the corner and shoot when someone passes to him.

If that guy is your teammate, you have an inordinate responsibility to facilitate the offense; and you will invariably engage in riskier plays in order to incorporate his one offensive skill.

Having more than one such teammate on the floor, there's exponentially more burden on the skilled facilitator. The irony is that the most skilled offensive player can display the worst "offensive rating", and vise-versa.

Mike G wrote:

... teams use possessions, players take opportunities.

This is true most of the time. Sometimes a player just loses the ball or throws up a bad shot.

Still, there is probably some use for the allotment of "possessions used" to individuals.
There's also a great potential for misuse. With "possessions" in the denominator, a player can have a crazy high ORtg, if all he does is park in the corner and shoot when someone passes to him.

If that guy is your teammate, you have an inordinate responsibility to facilitate the offense; and you will invariably engage in riskier plays in order to incorporate his one offensive skill.

Having more than one such teammate on the floor, there's exponentially more burden on the skilled facilitator. The irony is that the most skilled offensive player can display the worst "offensive rating", and vise-versa.

This is exactly what I'm getting at. My point was simply that the possessions framework is great on a team level, but far less obvious, and (as can be seen by looking at Wins Produced) indeed often misleading at the player level. Bad shots and turnovers to me represent opportunity usage, not possession usage, because it is usually the fault of many players on a team that the ball gets turned over or a bad shot gets attempted.