Comments on: A Look at Steals: Does Gambling Pay Off?

By: generic allegra

generic allegra — Mon, 15 Feb 2010 00:13:44 +0000

There’s nothing to worry about.
———————————————————
Signature:buy levitra professional online ecx

By: Jon Nichols

Jon Nichols — Sat, 16 Jan 2010 17:47:48 +0000

schtevie,

It was a question proposed to me by someone, and I don’t think it’s crazy to question if steals really are good. Obviously a steal is a good thing because it causes a turnover, but theoretically more steal conversions could be accompanied by more failed steal attempts. If there were enough failed attempts, they could outweigh the benefits of causing turnovers. The data shows that’s not the case, but I don’t think it’s unreasonable to investigate.

In addition, many fans complain that some of the best defenders as decided by the all-defensive teams are often overrated because of box score stats such as steals, so it’s worth looking at.

-Jon

By: schtevie

schtevie — Sat, 16 Jan 2010 17:41:38 +0000

Jon, has anyone ever alleged that steals are bad? The first order effect is, by definition, positive and large: a steal precludes a scoring attempt on that possession. The only controversy I am aware of is the degree to which the second order gambling costs (suggested to be 0.14 to 0.25 by your numbers) outweigh the offensive fast break gains. I expect that the answer, provided by the symmetric regression of steal rate on OR, will be a clear ‘no’.

By: Jon Nichols

Jon Nichols — Sat, 16 Jan 2010 17:06:28 +0000

schtevie,

I think you’re missing the point of the article. I was not looking for a proportional decrease in defensive efficiency. What I was looking for was any decrease in defensive efficiency at all. The point was to see if steals are good or bad. If they’re good, the slope is negative, and if they’re bad, the slope is positive. How “good” they are was not really the focus of this article, although it certainly could be something to look at going forward. The same goes for the correlation between steals and offensive efficiency.

-Jon

By: schtevie

schtevie — Sat, 16 Jan 2010 16:00:38 +0000

Jon, I am not sure I follow your interpretation of the results. As I understand your numbers, an increase of 1% in steal rate yields an increase (decrease?) in DE of -.93 or -.82. League-wide, DE this year according to B-R.com is 106.8. A 1% decrease in this is -1.07. If I am not mistaken, what your numbers show is that increasing steals has a less than proportionate effect on defensive efficiency. In turn implying that going for steals has adverse defensive consequences.

However, this, of course, is not the complete picture. There should be big gains on the offensive side where open court steals disproportionately yield high return fast breaks. Regressing steal rate on OE will be necessary to complete the story.

By: mattaq

mattaq — Sat, 16 Jan 2010 03:37:30 +0000

i agree, you cannot analyze how much steals help without seeing if possessions beginning with a steal are more likely to generate points.

By: Friday Bolts – 1.15.10 | Daily Thunder.com

Friday Bolts – 1.15.10 | Daily Thunder.com — Fri, 15 Jan 2010 15:03:33 +0000

[...] yourself to get smarter – Jon Nichols of HP writes about if steals are worth the gamble: “In conclusion, based on the evidence Iâ€™ve presented today, I would suggest that lineups [...]

By: Basketball-Statistics.com » Blog Archive » A Look at Steals: Does Gambling Pay Off?

Fri, 15 Jan 2010 04:55:39 +0000

[...] http://www.hardwoodparoxysm.com/2010/01/14/a-look-at-steals-does-gambling-pay-off/ [...]

By: The Mid-Afternoon Milk Mustache, featuring the Kings Dating Game | Stacheketball, an NBA Blog

Thu, 14 Jan 2010 21:17:15 +0000

[...] The Lampshade: Jon Nichols checks (statistically, of course) the correlation between steals and defensive efficiency. [Hardwood Paroxysm] [...]

By: Zeiram

Zeiram — Thu, 14 Jan 2010 16:32:10 +0000

You say that we need to differentiate between certain kinds of steals to have a conclusive say on wether steals are bad or good for a defensive. But that is not the point, steals are never bad, because a steal means you actually got the ball. That never hurts a team, gambling WITHOUT getting the steal is what hurts teams. So if we had a stat that would count steal attempts and we would build a percentage out of the steal attempts and the successfull steals (kind of like fg%), we would see that people with low steal% but high numbers in steal attempts hurt the defense just like high volume shooters with low fg% hurt their offense. So applying high pressure and creating a turnover and a steal is the equivalent to an inside shot while straying from your man to gamble the passing lanes is a three launched with a hand in your face. One thing is efficient and will help you win, the other will not.