The Turnover Question. (stats geeks, would love your thoughts)

Are you just making excuses for Lin?

Remember Harden is leading the league in turnovers for 2 years in a row and he's a true superstar... Lololololo

 

I always think 24 seconds violation(a turnover) is generally better than a low percentage shot, which more often than not ends up with a fastbreak and easy points for opponents.

Same holds true for defending shots: Good close out often lead to missed shots and ball possession, so technically it's better than blocking it outbounds.

However, 24 secs viloation make you look bad, and highlight block looks way more impressive...

 

Are there stats anywhere that distinguish live-ball turnovers from dead-ball turnovers?

Live ball turnovers are killers, because they hurt you on both ends.

Edit:

nbawowy.com helps answer the question:

http://nbawowy.com/query/qpjq25f59nfm9529

According to this, opponents shoot 57.9 eFG% off of steals against us. They shoot 47.6 eFG% on shots that aren't off of steals. They shoot 40.2 eFG% off of non-steal turnovers (so, I think a dead ball situation, where our defense is set).

 

Turnovers (of decent NBA players) only hurt the team when the player stop playing aggressively due to the concern of turnovers.

 

I know the Rockets track live-ball vs. dead-ball turnovers, because early last season Morey said that they weren't too concerned about Lin's turnovers because a lot of them were dead-ball turnovers (charges and passes thrown out of bounds.) But it appears to me (from memory) that Lin's live-ball turnovers are up this season. Would be curious to see these stats for the whole team.

 

Click the ball-handling tab in that link, and that expand the TOV Details towards the bottom of the page.

 

Not sure if this is what you are talking about, but last season, stats show that games where Lin had high TO's correlated more with wins, whereas games where Lin had low TOs correlated more with losses. For Harden, the TO stats made more sense. More TOs for Harden correlated more with team losses. As a guess, its possible that in games where Lin had high TOs, he played more aggressively, and thus we were more likely to win. Its only a guess.

 

correct if you have a beard they are ok but if you just signed with adidas they are not

 

Thanks, durvasa

 

So durvasa's link clears that up: Lin is indeed committing lots of live-ball turnovers this year.

And no Tim, that's a separate issue. But it could be true: In other words Lin's aggressiveness might, say, lead to higher turnovers but it also might lead to higher team TS%, which could outweigh the turnovers. That would be an interesting study: to see the link generally between turnovers and TS%.

 

This is one of those things that requires a lot more analysis to fully understand. Low turnovers, as defined by the box score, have been shown to correlate pretty strongly with wins (Dean Oliver assigns it a weight of .25 in his book "Four Factors of Basketball Success", with the other factors being scoring, rebounds, and free throws). However, one can perform other analyses that can show that in some situations high turnovers correlate with higher scoring (including higher scoring efficiency), which seems to intuitively run counter to the notion that low turnovers correlate with winning.

What I would say is that the idea that turnovers are, in and of themselves, classified incorrectly is probably correct. In some sense, each possession can result in either positive points (scoring the basket) or giving the ball to the other team. Within those two ultimate outcomes, however, are a whole slew of potential actions, each with their own probabilities of turning into either positive points or giving the ball to the other team. One example of this is simply shooting the ball, which can result in making the basket or missing, and then the results of missing are then determined by who gets the rebound, which resets the entire flow chart all over again (an offensive rebound having a high percentage to turn into points, while a defensive rebound means a loss of possession).

Taking into account every single possibility & probability, one can see how simply classifying something as a "turnover" or not might magnify what is in actuality a pretty small problem in the context of an NBA game. If roughly half of your possessions will ultimately result in a change of possession anyway, then the ~10-15% chance that there is a turnover could be understood as a comparatively small proportion of that pie.

What you really want to see is a flow chart or a stochastic process where each state has with it a corresponding probability. For example, shooting the ball might give you a 50% chance of points, with a 25% chance of an offensive rebound and a 25% chance of a defensive rebound. Having a better scorer changes those probabilities to 40/20/20. Having a better rebounder changes them to 50/35/15.