Weighting Possession stats (TOs/rebounds) against efficiency

[Dr Positivity]

A lot has been made of shot efficiency... how 3 pointers/close to the basket shots are more valuable than midrange 2s, who can give you those shots, TS%. However I believe that number of possessions is greater in value than the quality of a shot.

The idea of this came from the baseball stat-heads. A while ago they figured out the most important part of hitting is not getting out. The 3 outs represent life and extension of the inning, the closer you get to the 3, the worse off you are.

I believe it's similar with possessions. Here's why. Let's say a good shot - say an open 3 or a shot within 5 feet, converts at a .60-.65 TS% rate. Whereas a long contested two pointer converts at .40-45 TS%. These numbers aren't exact, but it's likely somewhere around there... in basketball lingo 40% shooting is bad, 60% is good, but this means even on bad shots you're going to score at 2/3s the rate of a good shot and 2/5s of the time overall. However when you turn the ball over you relinquish your possession completely. Instead of a 'bad' 45%, you get a big fat 0% on that possession. No shot goes up, no chance of points. The difference between .4 and .6 is much smaller than 0 and .4-.6.

Let's take 3 teams. Team A takes 100 shots at a .65 TS% clip. Team B takes 100 shots at a .45 TS% clip. Team C gives up their possession by TO. 100 times. Final scores - Team A: 65 points, Team B 45 points, Team C 0 points.

You're MUCH better off taking a bad shot than giving up your possession by turnover. Giving yourself a 0 conversion rate on a possession is the worst thing you can do. It's a complete whiff. The same applies for offensive rebounding. Giving yourself an extra possession guarantees a .40-.60 chance of scoring points whereas without the rebound, you have a 0. And thus giving the other team another possession is the worst thing you can do defensively.

Now I'm going to look at some team stats to see if it applies in the standings. I used bball-reference's 4 factors, listed here: http://www.basketball-reference.com/lea ... _2010.html

Atlanta - 1st TOs, 6th Orb / 11th eFG, 23rd FT/FGA / 3rd ORTG - Great TO/Orb, average field efficiency, great overall ORTG

Portland - 2nd TOs, 4th Orb / 14th eFG, 4th FT/FGA / 6th ORTG - Great TO/Orb, average field efficiency, great overall ORTG

Dallas - 3rd TOs, 26th Orb / 12th eFG, 18th FT/FGA / 11th ORTG - Average TO/Orb, average field efficiency, average overall ORTG

LA Lakers - 4th TOs, 11th Orb / 15th eFG, 17th FT/FGA / 10th ORTG - Good TO/Orb, average field efficiency, good overall ORTG

Milwaukee Bucks - 5th TOs, 18th Orb / 25th eFG, 30th FT/FGA / 23rd ORTG - Middle ground TO/Orb, horrible field efficiency, below average overall ORTG

New Orleans Hornets - 6th TOs, 24th Orb / 13th eFG, 29th FT/FGA / 17th ORTG - Middle ground To/Orb, below average field efficiency, middle ground overall ORTG

Miami Heat - 7th TOs, 20th Orb / 17th eFG, 16th FT/FGA / 19th ORTG - Middle ground TO/Orb, average field efficiency, average overall ORTG

Toronto Raptors - 8th TOs, 22nd Orb / 6th eFG, 6th FT/FGA / 6th ORTG - Middle ground TO/Orb, great field efficiency, great ORTG

San Antonio Spurs - 9th TOs, 15th Orb / 7th eFG, 22nd FT/FGA / 9th ORTG - Good TO/ORB, middle ground field efficiency, good overall ORTG

Denver Nuggets - 10th TOs, 19th Orb / 9th eFG, 1st FT/FGA / 2nd ORTG - Middle ground TO/ORB, great field efficiency, great overall ORTG

The part that really surprised me here was the impact of FT/FGA. Look at Denver especially. Average TO/ORB, slightly above average eFG, but the best FT/FGA carries them all the way to 2nd. Part of this is because they lead the category by miles. Den is first with .292, OKC is 2nd with .266, Charlotte is 3rd with .258 and it trickles down from there. The inverse are the Bucks who have the 30th FT rate at .179, one of 3 teams under .200.

This leads me to look at the difference between high and low TO rates, efficiency, and FT/FGA. Could it be FT/FGA's large impact is because there's such a gap between teams compared to the other categories? While I've concluded keeping your possession is more important than shot selection, if every team in the league fits within a small TO range but a larger efficiency one, then reasonably the latter would become more important

For TOs .125 ranks 5th, Utah's .140 ranks 26th, however Atlanta is an outlier in 1st with a .113 to 2nd's .121 and there's a huge drop-off in the bottom 4 with 27. Minny - .145, 28. Boston - .147, 29. Clippers - .148, and 30. Charlotte - .151. The difference between Charlotte's at 30 and Utah's at 26 is as big as Utah at 26 and Denver's at 8.

Looking at eFG. Boston is 5th with with .522 and Charlotte is 24th is .488. But once again there are outliers. Phx's 1st is .545, a large gap over Cle in 2nd at .534 and then the bottom teams are 30. Nets - .458, 29. Detroit - .469, 28. Chicago .474, 27. Washington - .477 26. Minny .477. Once again the gap between 30 New Jersey and 24 Charlotte is as big as between 24 Charlotte and 6 Toronto. Even the gap between Charlotte and Minnesota, a simple 24 to 26 drop, is the same as 24 to 14.

In FT/FGA the difference between 30 Milwaukee and 25 Sacramento is the same as between 25 Sacramento and 9 New Jersey. The difference between 1 Denver and 4 Portland is the same as between 4 Portland and 22 San Antonio (wow).

For increments going down the list, teams are closest in TOs, the gap in Orb and eFG is a little bigger, and the gap in FT/FGA is by far the biggest (cue Donaghy sign!)

But nevertheless in TOs, Orb, eFG, FT/FGA, the majority of teams in the 5-25 range will be within striking distance of each other and then the main outliers are the top and bottom teams. If you're an outlier in one of these for good or bad, it's going to have as large an impact as any variation within the middle ground. However getting back to my original point, can we then take the outliers for TOs and Orbs and weigh them against the efficiency ones? Let's try:

TOs:
High end outlier: Atlanta - 1st with .113. ORTG: 3rd
Low end outliers: Minnesota - 27th with .145, 28. Boston with .147, 29. Clippers with .148, 30. Charlotte with .151. ORTG: Minnesota - 29th, Boston - 15th, Clippers - 27th, Charlotte - 24th

Orb:
High end outliers - 1st Memphis .317, 2nd Detroit .304. ORTG: Memphis - 12th, Detroit - 22nd
Low end outliers - 27 New York .227, 28 Boston .228, 29 Indiana .214, 30 GS .212. ORTG: New York - 18th, Boston - 15th, Indiana - 27th, Golden State - 13th

eFG:
High end outliers - 1st Phoenix .545. ORTG: 1st
Low end outliers - 26 Minnesota .477 27 Washington .477 28 Chicago .474 29 Detroit .469 30 New Jersey .458. ORTG: Minnesota - 29th, Washington - 25th, Chicago - 26th, Detroit - 22nd, New Jersey - 30th

FT/FGA
High end outliers - 1st Denver .292. ORTG: 2nd
Low end outliers - 30th Milwaukee .179, 29th Hornets .188, 28th New York .197. ORTG: Milwaukee - 23rd, New Orleans - 17th, New York - 18th

Looking at that (admittedly small sample size), the ones hardest to come back from appear to be TOs and field efficiency, and the weak Orb teams do OK. For Golden State and New York that's because they have a decent efficiency 8th and 10th... they play small, sacrificing size (offensive rebounds) for this efficiency. Using the prior logic about possessions vs shot selection we can say this is not a wise trade-off. The Warriors and Knicks place the standings show this is true. Their overall ORTG is not bad, only 13th and 18th, but keep in mind they also sacrificed defense to run out these offense only lineups... so middle ground offense does not cut it and sacrificing offensive rebounds for higher efficiency is part of the reason why. Boston survives with weak TOs and Orbs because their eFG is 5th overall and they are 8th in FT/FGA... so despite being top 5 from the field in efficiency, they rank 15th overall in ORTG

For FT/FGA the teams that buck the trend are Milwaukee, New Orleans, and New York. Milwaukee and New Orleans are 5th and 6th in TOs which really helps them, whereas New York has their smallball thing.

After that I wasn't satisfied with the data, so I went farther through the list and attempted to find teams who performed identically in 3 of the 4 categories to see how large the impact was. In scientific terms this isolates 3 of the factors as constants and lets the 4th become the variable:

Orlando vs Utah- (difference in Orb)
In eFG Orlando has .530 and Utah has .524. In TOs Orlando is .137 and Utah is .140. In FT/FGA Utah is .253 and Utah is .248, making up for some of the eFG difference. In Orb Utah is .267 and Orl is .245. Overall ORTG Utah is 110.8 and Orlando is 110.4. So that .22 only made up .4ppg, translating to .84 for 1ppg

Portland vs OKC - (difference in TOs)
Portland has a .499 eFG to Okc's .493 (14th to 18th), TOs is .258 to .257 in OKC's favor, FT/FGA is .266 to .253 in OKC's favor which makes up for the eFG drop. TOs is the big drop. Portland is .121 and 2nd, OKC is .139 and 24th. Overall ORTG has Portland at 6th and OKC at 13th, or 111.0 to 108.0. So that roughly .018 difference made up 3 points overall in offense, translating to .06 to 1ppg

Dallas vs New Orleans - (difference in FT/FGA)
In eFG Dallas is .504 and New Orleans is .503. In TOs Dallas is .123 and New Orleans is .126. In Orb New Orleans is .249 and Dallas is .245. Almost as close as you can get in those 3 categories. Now FT/FGA. Dallas is .222, New Orleans is .188, a .34 difference. Overall ORTG Dallas is 11th with 108.8 and New Orleans is 17th with 107.0. So that .34 FT/FGA gap made up 1.8ppg, translating to .19 for 1ppg

Milwaukee vs Chicago 09 - (difference in eFG). Couldn't find anyone this year so I had to go back

In TOs Milwaukee is .131, Chicago .133. In Orb Milwaukee is .278, Chicago .280. In FT/FGA both are .329. In eFG Chicago is .293 and Milwaukee .283. Overall ORTG is 108.4 for Chicago and 106.7 for Milwaukee, difference between 14th and 23rd. So .01 eFG difference changes 1.7ppg, translating to .59 for 1 ppg.

This is of course a small sample size and the 3 constants in each comparison are not exactly equal. So over the next little while I'm going to keep digging through examples with 3 of the same, and trying to isolate each factor and see if I can get a consistent number and pattern out of it

[Jimmy76]

wow impressive analysis

so your basic conclusion is that maximizing the number of possessions is more important than what you make of those possessions if im reading this right

I think you would be better off just crunching the numbers than trying to look through comparable team tendencies

for instance a missed shot+defensive rebound is equivalent to a TO so you can weigh the odds of a shot being made or missed (or drawing fts) and if missed the odds it results in a defensive rebound vs an offensive one and then (somehow) use that to weigh efficiency vs possession management

correct me if i interpreted anything wrong I just did a pretty quick read through your post

[rrravenred]

Very nice, Dr. M.

I suppose the FTA are important because they are a high percentage scoring opportunity which is NOT affected by any opposition defensive pressure, and that the ability to obtain these opportunities at a minimal risk of giving the opposition a chance to obtain possession (i.e. taking a field goal, irrespective of whether it is made or missed) has a high correlation to general offensive efficiency.