APBRmetrics

Posted: **Fri Feb 28, 2014 8:26 pm**

At this NBA.com page --
http://stats.nba.com/leaguePlayerGenera ... PerPage=25

... I notice players have a Pace column. That's presumably from pbp -- and I haven't seen this before.
NBA.com apparently uses a different Pace formula than that used at b-r.com -- most teams are about 2-2.5 possessions per game higher at NBA.com

In any case, knowing the player's Pace and minutes; and the team's Pace and minutes; we can deduce the pace of the team when the player is off the floor, and the difference between them.
Something is wack with Minnesota, perhaps. This player list is limited to those who average >20 mpg, but it seems all Wolves players have a Pace above that of their team.

Code: Select all

.player           tm    %Min    Team     On     Off    diff
Kevin Love        MIN    .71    99.7   101.5    95.2    6.3
Corey Brewer      MIN    .65    99.7   101.8    95.9    5.9
Ricky Rubio       MIN    .65    99.7   101.8    95.9    5.8
John Wall         WAS    .76    95.7    97.0    91.3    5.8
Kevin Durant      OKC    .78    97.7    99.0    93.3    5.7

Kyle Lowry        TOR    .75    94.5    95.9    90.4    5.5
Russell Westbrook OKC    .32    97.7   101.3    96.0    5.3
Trevor Ariza      WAS    .68    95.7    97.2    92.5    4.7
Derrick Rose      CHI    .11    93.1    97.1    92.6    4.5
Kevin Martin      MIN    .56    99.7   101.7    97.2    4.4

.player           tm    %Min    Team     On     Off    diff
Carlos Boozer     CHI    .55    93.1    95.1    90.8    4.2
Ty Lawson         DEN    .58    99.7   101.3    97.4    3.9
Mario Chalmers    MIA    .54    94.4    96.1    92.5    3.6
Victor Oladipo    ORL    .66    96.0    97.2    93.7    3.5
Ramon Sessions    MIL    .50    94.4    96.1    92.6    3.5

Isaiah Thomas     SAC    .71    96.9    97.8    94.6    3.2
Jrue Holiday      NOP    .41    94.3    96.2    93.0    3.2
Eric Gordon       NOP    .63    94.3    95.4    92.4    3.0
Deron Williams    BKN    .46    94.1    95.7    92.7    3.0
Tim Hardaway Jr.  NYK    .45    92.7    94.3    91.4    2.9

Ball-pushers, mostly -- and Carlos Boozer. Maybe he just shoots early?

Players who slow the pace of the game:

Code: Select all

.player                 tm    %Min   Team     On     Off     diff
Josh McRoberts         CHA    .61    95.3    93.9    97.5    -3.7
Shaun Livingston       BKN    .50    94.1    92.5    95.7    -3.3
Michael Kidd-Gilchrist CHA    .34    95.3    93.3    96.3    -3.1
Ray Allen              MIA    .50    94.4    92.9    95.9    -3.0
Kevin Garnett          BKN    .38    94.1    92.3    95.2    -2.9

Pablo Prigioni         NYK    .31    92.7    90.8    93.6    -2.9
Channing Frye          PHX    .60    98.3    97.2    99.9    -2.8
Tayshaun Prince        MEM    .51    92.2    91.0    93.5    -2.6
Terrence Jones         HOU    .52    98.3    97.2    99.6    -2.5
Gerald Henderson       CHA    .68    95.3    94.5    96.9    -2.5

Hornets are the anti-Wolves, perhaps.
We have to assume these Diff numbers are often relative to a player's backup (or the player he's backing up).
I dunno -- is this anything?

Posted: **Fri Feb 28, 2014 10:46 pm**

I'm surprised CP3 isn't on the list. he is the first name I thought when it came to pace killers.

Posted: **Fri Feb 28, 2014 11:41 pm**

No kidding. If I'm reading this right, he just misses the list of pace boosters and leads the Clipps.

Code: Select all

.player          tm    %Min   Team     On     Off     diff
Chris Paul      LAC    .49    98.0    99.4    96.7    2.6
Jamal Crawford  LAC    .63    98.0    98.7    96.8    1.9
Matt Barnes     LAC    .35    98.0    98.6    97.7     .9
JJ Redick       LAC    .30    98.0    98.4    97.9     .5
Blake Griffin   LAC    .75    98.0    98.1    97.9     .2
Jared Dudley    LAC    .52    98.0    97.5    98.6   -1.1
Darren Collison LAC    .48    98.0    97.4    98.6   -1.2
DeAndre Jordan  LAC    .75    98.0    97.7    99.1   -1.4

Maybe he burns seconds in the half court, but he also takes advantage of early ops?
Don't ask me if subbing out Paul and putting in Collison makes the team 3.8 poss/G slower. But it would seem to be so.

Posted: **Sat Mar 01, 2014 10:59 am**

That seems like something the RAPM framework could work very well for. Instead of having 'points' as dependent variable one can just use 'length of possession'. Should be a little more accurate than the raw numbers

Posted: **Sat Mar 01, 2014 11:49 am**

I have built a box score model for this - the box score is actually rather good at capturing the required information in this instance. I'll dig out the basic model coeffs when I have a moment. As J.E. suggests the RAPM framework can also be used for this.

Posted: **Sat Mar 01, 2014 12:17 pm**

... limited to those who average >20 mpg, ... it seems all Wolves players have a Pace above that of their team.

Actually, Turiaf is the one pace-killer on the list.
Now that I look, I see Love, Pekovic, Rubio, Brewer, and Martin have played 900 minutes as a unit. Next most-used replaces Peko with Turiaf (134 min.) Other lineups are all <80 minutes.

So it's that old multi-colinearity thing. The starters get out on the break and take the first good shot in the halfcourt. Any substitution is a downgrade in efficiency-at-high-speed.
Maybe over-use of the starters is also their problem late in close games.

Posted: **Sat Mar 01, 2014 4:59 pm**

If one wants to isolate individual pace and individual impact on pace, it would be good to take it down to offensive and defensive pace instead of just overall pace. Hypothetically one might even be able to take impact on pace down to Factor level to see more about how the player is impacting pace thru these actions (and maybe others).

APBRmetrics

Players who affect their teams' Pace

Players who affect their teams' Pace

Re: Players who affect their teams' Pace

Re: Players who affect their teams' Pace

Re: Players who affect their teams' Pace

Re: Players who affect their teams' Pace

Re: Players who affect their teams' Pace

Re: Players who affect their teams' Pace