Is shooting a lot of three-pointers a bad thing?

[Minstrel]

This is an article from Basketball Prospectus, explaining the unexpected success of the Atlanta Hawks (unexpected considering they lost Josh Childress, drafted no one and added no major free agents). It's about the Hawks, but it largely discusses the surprising correlation between shooting a lot of three-pointers and offensive success.

An excerpt:

Let's take a look at the league's most improved offenses along with a mystery stat that has seemed to mirror their overall improvement.

                OFFENSIVE RATING         MYSTERY STAT
Team          0708    0809   Diff     0708    0809   Diff

Cleveland    107.5   115.1   +7.6     .190    .256   +.066
Portland     108.5   115.6   +7.1     .218    .258   +.040
Miami        102.2   107.7   +5.5     .217    .244   +.027
New Jersey   105.6   111.0   +5.4     .222    .253   +.031
Atlanta      108.3   111.2   +2.9     .165    .279   +.114
New York     105.6   107.6   +2.0     .215    .344   +.129

Any guesses as to what the mystery stat might be? Here's a hint--and what makes this so interesting. The mystery stat is not an "outcome" stat like a shooting percentage or offensive rebounding. Instead, it's a tendency stat, and one that in theory should not necessarily have anything to do with the performance of an offense. Got it? Maybe the enormous leap by the Knicks tipped you off that the mystery stat has to do with three-point shooting. It is, in fact, the percentage of the team's field-goal attempts that have come from beyond the arc.

On a league-wide basis, teams have attempted the same percentage of their shots as threes this year as last (22.2 percent), so that doesn't explain the change for the improved offensive teams. All of them were at league average or below a year ago; now they're all above it, some (including the Hawks) dramatically so.

If this seems like something more than a coincidence, that's because it is. Looking merely at three attempts, without any regard for success, is a surprisingly decent indicator of offensive performance. The correlation between 3A/FGA and Offensive Rating so far this season is .562 (a correlation of 1 or -1 indicates two variables move in lockstep, while a correlation of 0 means no relationship), almost as good as the correlation between three-point percentage and Offensive Rating (.590).

My friend David Locke of The Fan Sports in Salt Lake City took a slightly different look at the numbers last week, finding that the top ten teams in the league in three attempts per possession were winning at a combined .620 clip. Considered either way, the evidence seems to point to one conclusion: The old adage "live by the three, die by the three" is about half correct.

There's more, including a contrast with the Philadelphia Sixers, who have been disappointing and are not taking many three-pointers as a percentage of their field goal attempts.

http://www.basketballprospectus.com/article.php?articleid=487

 

[BatumKaboom]

Not if you make them!

 

[GoBlazersGo]

Not if you make them!

[/thread]

 

[BalancedMan]

A big part is "if you shoot them well". The Hawks added several good outside shooters this last offseason, which is why they are more likely to shoot and made the three. I read the article a few days ago and I don't think threes are good if they are made at a good clip. 40% 3PT is 60% 2PT.

 

[bobf]

Two teams

Team A shoots all 3 pointers and makes 1/3 of them.
Team B shoots all 2 pointers and makes 1/2 of them.

Each team gets 70% of the defensive rebounds and 30% of the offensive rebounds.

There are no turnovers, fouls, etc. just shots and rebounds.

Who is expected to win?

 

[Iwatas]

Each team scores 1 point each possession, with an equal number of possessions (same rebound rate for both teams).

Seems they should both score the same. What is the flaw?

 

[BrianFromWA]

Team A scores the same as Team B in the first iteration (1st shot), but then gets a larger number of offensive rebounds, and converting at the same rate would start to pull away.

Just to use 100 possessions, A scores 100 on 33.3 made 3ptFG. Team B scores 100 on 50 made FGs. But A gets 30% of the 67 offensive rebounds (20) and converts 6.67 of those into made 3ptFG (still shooting 33.3%, remember?) for 20 points, while Team B scores 50% of it's 15 offensive rebounds for 7.5 made FGs and 15 points. Now it's 120-115 Team A. But A got 30% of it's "second shot" rebounds, as did B...continue out to 3 iterations and it's 124-118 Team A.

 

[bobf]

Team A scores the same as Team B in the first iteration (1st shot), but then gets a larger number of offensive rebounds, and converting at the same rate would start to pull away.

Just to use 100 possessions, A scores 100 on 33.3 made 3ptFG. Team B scores 100 on 50 made FGs. But A gets 30% of the 67 offensive rebounds (20) and converts 6.67 of those into made 3ptFG (still shooting 33.3%, remember?) for 20 points, while Team B scores 50% of it's 15 offensive rebounds for 7.5 made FGs and 15 points. Now it's 120-115 Team A. But A got 30% of it's "second shot" rebounds, as did B...continue out to 3 iterations and it's 124-118 Team A.

Yup... another way to solve it.

A = points per possession team A
B = points per possession team B

A = prob(make) x 3 + prob(miss) x ORB% x A
A = (1/3) x 3 + (2/3) x 0.3 x A
A = 1 + 0.2A
A = 1.25

B = prob(make) x 2 + prob(miss) x ORB% x B
B = (1/2) x 2 + (1/2) x 0.3 x B
B = 1 + 0.15B
B = 1.1765

So the final score would be like 125 to 118

So 3 pointers have a built-in advantage to offset the long rebound.

 

[bobf]

Each team scores 1 point each possession, with an equal number of possessions (same rebound rate for both teams).

Seems they should both score the same. What is the flaw?

They will have an equal number of major possessions but not an equal number of minor possessions and shots.
Minor possessions = Major possessions + Offensive Rebounds
Shots = Minor possessions
Points = Shots (on average)

The team that misses the most shots will get the most offensive rebounds.

 

[Iwatas]

Thanks. Well explained!

 

[LameR]

Also, it's more likely for the two-point shooting team to end up with free throws than the three-point shooting team.

 

[Denny Crane]

Don't offensive rebounds count as another possession? If so, it really screws up the math presented here.

I think it's true that the 2pt shooting team will draw way more fouls.

I think it's also true that shooting 33% from 3pt is the same as shooting 50% from 2pt. Teams shooting 40% from 3pt may well overcome any advantage at the FT line for the 2pt shooting team.

 

[Ed O]

I would be interested to see the correlation between volume three point shooting teams and offensive rebounding rate. With longer shots, there SEEM to be more long rebounds, with defenders more spaced and therefore more offensive board opportunities. I'm not sure that's actually the way that it turns out, though...

Ed O.