APBRmetrics

I've heard a few people (John Hollinger, David Thorpe, etc.) mention a high turnover rate as a rookie as an indicator of a player having a high ceiling. The idea being that it represents a certain amount of creativity and aggressiveness, both positive which can't be taught if not innately present in a player, but which can be improved upon.

I'm wondering if anyone is aware of any statistical analysis tackling this idea. I'm thinking about taking on the project, but don't want to shabbily duplicate something that's already out there.

Certainly seems counterintuitive to me. Maybe I've played with too many players that were turnover machines 10 years ago, and still are now.

I would be surprised if someone found such an effect. If someone finds it, I'd suspect you're finding players that made it into the league despite them being turnover machines, but instead due to their physical attributes. Because of their above average physical attributes they later become above average basketball players

Easy enough to check. Since 1991, just 21 rookies have played 200+ minutes and had as many as 4 TO per 36 min: (Chronologically) Rumeal Robinson, Elliot Perry, Greg Sutton, Tharon Mayes, Jim Jackson, John Crotty, Doug Christie, BJ Tyler, Trevor Ruffin, Randolph Childress, Mark Davis, Cory Alexander, Corliss Williamson, Vince Esposito, Drew Barry, and Jacque Vaughn, in the '90s.
In this Century: Milt Palacio, Smush Parker, DeMarcus Cousins, Armon Johnson, and currently Kyrie Irving.
Irving's TO rate has come down dramatically in the last couple of weeks, so he may well be under 4/36 by season's end.

TO% is another matter, and some of the highest are owned by guys who don't do much with the ball; a couple of illegal picks per game will be a large % of their 'possessions'.

There are also 21 players in the era who played at least 1000 minutes as rookies and had at least 3.5 TO/36 min. : Laettner, Shaq, Oliver Miller, Walt Williams, Shawn Bradley, Glenn Robinson, Bob Sura, Kobe, Iverson, Maggette, Francis, Mateen Cleaves, Tinsley, Smush, Raul Lopez, Marcus Williams, Westbrook, Jonny Flynn, John Wall, Eric Bledsoe, and Cousins.
http://bkref.com/tiny/mI7Xk

In this list, those with greatest success seem to have all come into the league while very young.
Others had flawed games and didn't last so long nor develop so well.

There are 22 rookies with TO/36 > 3 and with TO% > 20, and 1000+ minutes.
These include a number from the previous list, plus: Jalen Rose, Kidd, Antonio Daniels, Rodrick Rhodes, Baron Davis, Anthony Carter, Kaman, Marcus Banks, Steve Blake, TJ Ford, Telfair, Orien Greene, and Jrue Holiday.

Rookies over 2000 minutes with turnovers under 2.5 / 36 minutes

http://bkref.com/tiny/wRPjD

Which list is more impressive?

Perhaps a look at career ws/48 or some other metric would help address the question.

The resulting article

http://www.hickory-high.com/?p=3055

Jeremy Lin has stolen the show since earlier posts in this thread.
Is it to be TO% or TOV% or ToV%? I vote for the former and definitely not the 3rd.
Why is the letter V in there?

TO% is what most use. The later two are clunky but people still know what is meant.

Mike G wrote:Jeremy Lin has stolen the show since earlier posts in this thread.
Is it to be TO% or TOV% or ToV%? I vote for the former and definitely not the 3rd.
Why is the letter V in there?

"TO" is also the natural abbreviation for timeouts. Which would almost never be used in any statistical analysis of an individual player, but could be used when analyzing team statistics (I've actually seen almost no analysis of timeouts except for a few articles about OffRtg off of timeouts, but one could imagine broader deeper analyses of team tactics and game situations which would take timeouts int account).

I'm fine with either TO or TOV. But if we do start better tracking of timeouts, and use "TO" for turnovers, what will the abbreviation for timeouts be? "TOut"? "TOT"?

The only NBA arenas that I've been in in recent years have been Staples Center in LA and the Rose Garden in Portland, IIRC both have arena scoreboards which use "TOL" to indicate timeouts left. That's a good abbreviation (because it quickly tells the fans that the scoreboard is counting timeouts left, not timeouts used), but it's useful only during a game, not useful for analysts after the game is over. Although come to think of it, maybe TOU and TOL (and TOU30 and TOL30 for the 30-second timeouts) could be used by analysts.

To muddy the waters further, I'm pretty sure one of the popular websites uses "TOR" (turnover rate).

The average reduction from rookie season to career TO% among those over 20% TO% was 13%. Some of that was an improvement curve, some of it was perhaps a somewhat random rookie season rate.