APBRmetrics

Yep that's my major issue now. I don't know if I should adjust for touch time or not. And if I did, what speed would I use? Ideally you'd have the distance right when the shooter catches it.

I'm still confused about what 'spacing the floor' is. I have some ideas, but I'm not sure there's consensus. Is there some seminal article about it out there?

Steve Shea's approach / write up mentioned here earlier sounds like it is an valuable contribution. Perhaps going beyond convex hull area, one could characterize the contours, descriptively and mathematically. What contours do best / worst in general and against specific play types? And ideally should the analysis be made dynamic analysis over time. I haven't read the article yet to know what is covered vs. not yet.

(subtweet: Time... There is a place for / big value in doing something big / long drives and perhaps a little too for small returns of tennis service * many returns, I think).

So would floor stretching be more accurate term in this case? Whereas "spacing" can include an eg. like as mentioned, Dwight Howard in the post drawing defenders

My bigger plan was to create a WAR type of stats that gave players credit 1/4 for volume, 1/4 for the efficiency of their possessions vs average efficiency, 1/4 floor spacing and 1/4 defense. Having the volume category in there would be what attempts to give Howard more credit in comparison to Tyson Chandler for the post ability that draws more defenders.

The thing about a player like Howard is that in the old days you would probably want to double team him, and actually draw a wing defender away from the 3pt line towards the basket. I think teams have realized it's better to defend Howard with single coverage in the post, so as not to leave any shooters open. Now, if you're an offense with a weak shooting SF, then you might pull off him and double more in the post.

This is why offenses need to have as many shooters as possible.

AcrossTheCourt wrote:You have a problem with a linear assumption. A shot from 10 feet does not have half the spacing effect as a shot from 20 feet. And, of course, shots from behind the line have a much greater spacing effect than short jumpers.

And yeah, you can't ignore position. It's really powerful. Height*3PA is too conservative in my opinion.

I agree with both points, though one of the reasons why adding assisted % into the equation brought out satisfactory results for me of identifying the spot up floor stretch guys, may be related to your concerns.

According to hoop data's still available info for the 2012-2013 season, assisted % splits:

53.9% at rim
43.1% 3-9 ft
44.2% 10-15 ft
61.7% 16-23 ft
83% 3pt

So part of the way the difference between 3pt spacing and midrange spacing is accounted for more than just in the distance, is that the 3pt shot is almost 2x as likely to be assisted and nobody should be trying to get spot up 10 footers anymore

The other theory one might have is that because big men dribble less, they are more likely to be assisted on their FGs, which would help give big men credit for spacing more than wings. However the splits for assisted % don't support this idea the best:

44.0% PG
60.2% SG
65.8% SF
66.0% PF
63.8% C

I was surprised that PFs and Cs are close to SG and SFs in assisted % on average. I'm guessing it's because of put back points not being assisted.

In terms of "adjusting for position" for bigs vs wings, in the same way Jason Kidd's rebounding or James Harden's assists are more special for the position they play, just having context of where the other bigs rank to see how impressive that player's floor spacing is, might be enough to account for it