APBRmetrics

Posted: **Tue Apr 12, 2016 1:17 pm**

This point was made in another thread:

Mudiay's usage being almost as high as Paul means he is probably exposing himself to about the same level of TO risk / danger as Paul,...

Turnovers are associated with Assists, certainly. And with scoring attempts, among other things. Around the league, a player's expected turnovers can be estimated by weighting his other raw stats and summing them:

Code: Select all

Weight   Stat
.258    Assists
.103    Blocks
.088    Fouls
.069    2FG 
.055    FTA
.051    3FGA
.044    2FGA
.040    DReb
-.012   Minutes

No weight on/correlation with OReb or Stl; none with 3FG and FT, apart from the attempts.
A negative correlation with minutes says a low minute guy can be expected to get more TO, per x min.
Biggest total season deviations from this formula :

Code: Select all

.  low turnovers          high turnovers
-82   Chris Paul        57   James Harden
-59   Mike Conley       53   Robert Covington
-56   Al Horford        51   Nerlens Noel
-47   Dirk Nowitzki     48   Emmanuel Mudiay
-45   Jimmy Butler      47   D'Angelo Russell
-43  LaMarcus Aldridge  45   Paul George
-41   Kawhi Leonard     43   Brandon Knight
-39   Anthony Davis     38   Markieff Morris
-37   Kemba Walker      37   Ben McLemore
-36   Ricky Rubio       36   Alex Len
-35  Langston Galloway  36   Bojan Bogdanovic
-34   DeMar DeRozan     35   Dwight Howard
-33   Al Jefferson      34   Mario Hezonja
-33   Ish Smith         33   Nicolas Batum
-32   Isaiah Thomas     32   Raul Neto
-32   DeAndre Jordan    32   Stanley Johnson
-30   Pau Gasol         32   DeMarcus Cousins
-29   Chris Bosh        28   Tyson Chandler
-27   Marvin Williams   28   Dion Waiters
-27   Aaron Gordon      27   Markieff Morris

Among players with >1000 minutes, the best/worst TO/36 min. rates, relative to norms:

Code: Select all

.  low turnovers /36       high turnovers /36
-1.29  Chris Paul        1.14   Robert Covington
-1.20  Mike Conley       1.05   Ben McLemore
-.88   Blake Griffin     1.03   Mario Hezonja
-.83   Al Horford        1.01   Nerlens Noel
-.79   Dirk Nowitzki      .94   Markieff Morris
-.74   Jimmy Butler       .92   Emmanuel Mudiay
-.74  LaMarcus Aldridge   .84   D'Angelo Russell
-.68   Kawhi Leonard      .82   JaKarr Sampson
-.68  Langston Galloway   .82   Brandon Knight
-.65   Anthony Davis      .82   Raul Neto
-.61   Ricky Rubio        .81   Alex Len
-.60   Ish Smith          .72   Stanley Johnson
-.59   Chris Bosh         .71   James Harden
-.56   T.J. Warren        .70   Tyson Chandler
-.56   Aaron Gordon       .70   Joe Ingles
-.56  Shaun Livingston    .66   Bojan Bogdanovic
-.52   Jeremy Lamb        .62   Paul George
-.52   Jared Sullinger    .60   Dwight Howard
-.52   Ish Smith          .58   Manu Ginobili
-.51   Pau Gasol          .57   Jahlil Okafor

Rather a lot of rookies in the right hand lists -- and some with bad hands?

EDIT -- My data are a few days old. Al Jefferson is now >1000 min., and per36 he's right up there with Chris Paul in turnovers saved.

Posted: **Tue Apr 12, 2016 6:24 pm**

I don't recall seeing such a correlation of box score stats and turnovers before. Good stuff.

New data increases attention to passes, potential assists and total usage but I don't recall any major work on turnovers. If there was been something good and deep recently, please give a link. Have teams focused on it more than the press / outside analysts? Perhaps given the results of getting to lowest ever TO%. Are there types and times of passes that have seen significantly greater reductions of has the change been non-descript / just general?

Most of the correlations have direction explanations. The correlations with blocks and fouls imo indirectly lead predominantly to interior defensive workers who might not be major offensive forces on average, have bad hands or dribble, or see passes very sporadically and under time duress and likely operate inside with more threats of reach-ins. Or to in general risk-takers regardless of size and offensive skill?

Posted: **Wed Apr 13, 2016 12:41 am**

I thought of risk takers, too -- so why no correlation with steals? That's a gamble.
Shot blockers are often too tall to be coordinated?
Some guys with very high TO% and low Usg% are getting lots of fouls by illegal screens. This is both a foul and a turnover, so there's a correlation.

There have been about 11,600 blocked shots this year. At .103 TO/Blk, that's about 1200 TO in the league "due to" blocks. This is just 3.6% of the TO this year.
Assists account for about 42% of TO.

Posted: **Wed Apr 13, 2016 6:17 am**

Steal thiefs know how to steal and perhaps are best aware how to prevent it? Have the hand and foot coordination and quickness and mental acumen necessary for both?

APBRmetrics

High and low turnover guys.

High and low turnover guys.

Re: High and low turnover guys.

Re: High and low turnover guys.

Re: High and low turnover guys.