The three-point shot?

[NetsForce]

I had these graphs laying around and figured I'd post them, maybe they'll generate a discussion, maybe they won't, either way they look pretty:



Thinking back what prompted me to generate the graphs was some debate over whether or not the three-point shot was "ruining" the game of basketball... I remember one of the arguments being made was that players nowadays were much more efficient shooting the three-point shot than they were in the past, the guy didn't cite any numbers so I figured I'd see for myself if that was the case...

Anyway here are some notes:

1994-1995: Shorter three-point line
1995-1996: Shorter three-point line
1996-1997: Shorter three-point line
1998-1999: Lockout (50 game season)

[tsherkin]

Well, given the increasing number of guys who can shoot 40% or better from downtown, this is just playing the numbers, right? The three-point shot, especially the corner three (which San Antonio has been exploiting for years) has a really big pay-out if you can make it at a high percentage.

And if you've got a guy like Duncan or Dwight, then you're shooting 3s all day long because you've still got high-efficiency interior play.

[CellarDoor]

Well, given the increasing number of guys who can shoot 40% or better from downtown, this is just playing the numbers, right? The three-point shot, especially the corner three (which San Antonio has been exploiting for years) has a really big pay-out if you can make it at a high percentage.

And if you've got a guy like Duncan or Dwight, then you're shooting 3s all day long because you've still got high-efficiency interior play.

To add to that, the three point shot wasn't valuable to a player growing up in the 70s. Even in the 80s it wasn't that prolific, so there's been an exponential growth of guys who can hit that shot and while there are illegal defenses now, the zone rules allow for a lot of room for shooters on the perimeter. As the game is evolving it's becoming a legit weapon and three point specialists are more and more finding their places.

[Miller4ever]

The spikes from 94 to 97 is attributed to the shorter three-point line, so it's just basically been climbing.

Here's what I predict will happen:

The 3-point shot is getting more and more attempts as the average range of a player increases. We have 7-feet tall guys who shoot the three. However, there is a limit to how effective the strategy of using the 3 to spread the floor can be. At a certain point, a combination of defenses adjusting, the faster defenders that will close out to meet that range, and the reaction catches up to the trend. It's the same thing with the forward pass in football. It's become such a specialist skill that there's one guy who gets paid millions of dollars to throw the pass forward and get more yardage, despite the fact that running it almost guarantees more control over the result. The first team to perfect it manages to dominate, but defenses begin to catch up. I think that it's possible we'll see the attempts go up even more, but the 3PTA% will settle, much like the balance between running and passing has been struck in football.

[JoNoNRose]

The spikes from 94 to 97 is attributed to the shorter three-point line, so it's just basically been climbing.

Here's what I predict will happen:

The 3-point shot is getting more and more attempts as the average range of a player increases. We have 7-feet tall guys who shoot the three. However, there is a limit to how effective the strategy of using the 3 to spread the floor can be. At a certain point, a combination of defenses adjusting, the faster defenders that will close out to meet that range, and the reaction catches up to the trend. It's the same thing with the forward pass in football. It's become such a specialist skill that there's one guy who gets paid millions of dollars to throw the pass forward and get more yardage, despite the fact that running it almost guarantees more control over the result. The first team to perfect it manages to dominate, but defenses begin to catch up. I think that it's possible we'll see the attempts go up even more, but the 3PTA% will settle, much like the balance between running and passing has been struck in football.

Actually, passing is becoming more and more popular in football. I think perhaps a more accurate comparison is that in both sports the less conservative method is becoming popular. Yes running is safe, but passing gives much more options and offenses are realizing it is the smarter route. Similar to 3 pt shooting, while it may not be as safe as a shot in the post, it is one of the most efficient shots if you have good shooters and teams are moving more towards whats most efficient and less whats safer.

[Miller4ever]

Not to keep getting off-topic, but the NFL is having a great year in terms of QB's. I honestly think this year is a unique one because Brees and Favre got magic surgery, and Manning is doing super-well, and McNabb is healthy, and other QB's are having extraordinary years, so I think things will calm back down next season.

[Bleeding Blue]

The reason the numbers are up is because of the way the games are officiated. Defenders cant touch WR's and QB's without a penalty, thus higher passing %'s and yards. The NFL wants more passing because it scores more points and it more exciting to watch. Most fans dont want to see a 7-3 game, they want to see a 41-35 shootout. (Sorry to keep this off-topic)

[tsherkin]

FWIW as far as three-point shooting is concerned, Hollinger sees things evening out now as teams adjust to defending the 3. It's in one of his recent pieces and it talks about how the increase in usage is slowing down and percentages are slowing down. He also talked about the decline in offensive rebounding as a result of greater floor-spacing.

Meantime, as a corollary, I think we'll see a new interest in the mid-range and post games. The NBA wheel continues to turn. It's all in the cycles.

[Doctor MJ]

The reason the numbers are up is because of the way the games are officiated. Defenders cant touch WR's and QB's without a penalty, thus higher passing %'s and yards. The NFL wants more passing because it scores more points and it more exciting to watch. Most fans dont want to see a 7-3 game, they want to see a 41-35 shootout. (Sorry to keep this off-topic)

This. People like watching passes, so rules have been changed to favor quarterbacks. A change in rules could easily render the passing game dead. Take out pass interference (you can basically hit any player at any time except a receiver so taking out that rule isn't as arbitrary as it sounds) and the teams will focus a lot less on passing.

[Doctor MJ]

As to the general thread, yes, players today are a lot better at the 3 ball. Even when % levels off, the level of difficulty with greater volume means skill is going up. Course none of this correlates very strongly with talent, players from earlier times would've been better if they'd have been encouraged to shoot the 3 more.

[rrravenred]

Well... the incentive of the extra point justifies the specialisation in skill, keeping in mind the "greater good" of floor spacing, even if a shot is (ostensibly) inefficient, is something that's going to be determined by how kindly they treat other forms of scoring. You de-emphasise slashing (through, for example, clamping down on "crab-dribbles" and the like) then you make slashing a rarer and less-effective feature of the game, incentivising both mid-and-long range shooting. Then it comes down to how effective jumpshooting is cmopared to a post-game or a neutered slashing game.

[Mamba Venom]

The NBA has more 3 point snipers than ever.

The Magic made it to the Finals w/ all those shooters. The NBA is CHANGING.

[TrueLAfan]

There are more threes taken and connected on at a higher percentage. But the net result, in terms of scoring and scoring efficiency, has been negligible. To be specific:

In 1981—the second year the NBA had the three—league eFG% was .489 and TS% was .552.
In 1986, league eFG% was .493 and TS% was .558.
In 1991, league eFG% was .487 and TS% was .550.
In 1996, league eFG% was .499 and TS% was .560. (Shortened three point line.)
In 2001, league eFG% was .473 and TS% was .534.
In 2006, league eFG% was .490 and TS% was .554.
In 2009, league eFG% was .500 and TS% was .560.

What strikes me isn't how much things have changed, but how similar they have remained. Even though we have a lot more threes taken and connected at a much higher percentage than we did in, say 1981...it hasn't resulted in greater efficiency. This is one reason I'm a little dubious when people talk about “A team from (give old NBA season) wouldn't be able to play today's game because of the three.” Sure they could. They'd have to adjust in the same way a modern team would have to adjust to the issues with more midrange shots and ensuing defensive issues...mainly more free throws, fouls, and foul related substitutions. As tsherkin notes, it's cyclical.

Ripping off a previous post of mine: Shooting the number of threes we do today is a stylistic choice. There are good and bad things about shooting more threes, and some changes that end up being effectively neutral as well. For instance:

Good
--1.5 x 3P% > Non three point shooting On the surface, it looks like you get more points shooting the three. So far this season, players are shooting .489 from inside the arc and .348 from three. If you took 1000 shots from each spot, you’d get 978 points from inside the line, and 1086 from distance…a substantial difference.

Bad
--Fewer offensive rebounds. The thing no one talks about. Three points shots are (much) less likely to have offensive rebounds. From 1979 to 1989, over 33% of rebounds were offensive rebounds. As the number of threes went up, offensive rebounding numbers went down. From 1999 to 2008, only 28.5% of rebounds were offensive rebounds.

Good
--Spreading the court more. This is the most obvious plus, at least on the surface. If you’ve got perimeter passing, you’ll get more isolation sets and better spacing. At least in theory. But this may be offset by…

Neutral
--Fewer Assists. There’s less ball movement with three point shooting. Total assist numbers are way, way down…from a peak of over 26 a game in the mid-80s to a little over 21 a game today. Pace doesn’t account for that 20% drop…the difference in pace from the mid 80s to today is around 10%. Percentage of made shots credited with assists stayed at around 61-62% from the 80s through the early 90s. It has dropped slightly in recent years. But it should probably drop more than that, because ...

Bad
--Fewer fouls/points scored at the line. Three point shots are less contested and less likely to receive fouls. Consequently, shot attempts on teams that shoot fewer threes get an extra push. It’s hard to determine the exact amount, but it’s considerable…as noted above, you get more points from shooting the three. And a smaller percentage of points comes from free throws now—from 1984-87, around 20.5% of points came from free throws. That dropped about 1 to 1.5% in the early 1990s…right when the numbers of threes attempted went up. It’s at about 19% now This offsets the idea that you get more points per shot from the three (as noted above), because the end result is…

Neutral (again)
--Virtually identical TS%. See numbers at the beginning of the post. The bottom line is that teams are not more or less efficient than they once were at producing points. The most efficient period in NBA history was, again, the early and mid-1980s. The least efficient since the three is the late 90s, by a lot. It was on an upswing in recent years, although we’re down again this season.

What does this all say? Offensive styles run in trends. Shooting more threes isn’t “better”…it’s more current. You could run a perfectly good offense that would be just as efficient and effective as any team's today and take 200 threes a season. But players are not taught to do this because--well, because they aren't. As shown above, it's not a matter of scoring effectiveness or offensive efficiency. The tendency by many posters is to think that something that is pretty much cyclical over time—like offensive style and production—is progressive. Basketball, like baseball, is not like that. There are all sorts of contributing factors to changes in style, but those changes don’t make something better—it just makes the thing different than it was.

[Sundamental]

There are more threes taken and connected on at a higher percentage. But the net result, in terms of scoring and scoring efficiency, has been negligible. To be specific:

In 1981—the second year the NBA had the three—league eFG% was .489 and TS% was .552.
In 1986, league eFG% was .493 and TS% was .558.
In 1991, league eFG% was .487 and TS% was .550.
In 1996, league eFG% was .499 and TS% was .560. (Shortened three point line.)
In 2001, league eFG% was .473 and TS% was .534.
In 2006, league eFG% was .490 and TS% was .554.
In 2009, league eFG% was .500 and TS% was .560.

What strikes me isn't how much things have changed, but how similar they have remained. Even though we have a lot more threes taken and connected at a much higher percentage than we did in, say 1981...it hasn't resulted in greater efficiency. This is one reason I'm a little dubious when people talk about “A team from (give old NBA season) wouldn't be able to play today's game because of the three.” Sure they could. They'd have to adjust in the same way a modern team would have to adjust to the issues with more midrange shots and ensuing defensive issues...mainly more free throws, fouls, and foul related substitutions. As tsherkin notes, it's cyclical.

Ripping off a previous post of mine: Shooting the number of threes we do today is a stylistic choice. There are good and bad things about shooting more threes, and some changes that end up being effectively neutral as well. For instance:

Good
--1.5 x 3P% > Non three point shooting On the surface, it looks like you get more points shooting the three. So far this season, players are shooting .489 from inside the arc and .348 from three. If you took 1000 shots from each spot, you’d get 978 points from inside the line, and 1086 from distance…a substantial difference.

Bad
--Fewer offensive rebounds. The thing no one talks about. Three points shots are (much) less likely to have offensive rebounds. From 1979 to 1989, over 33% of rebounds were offensive rebounds. As the number of threes went up, offensive rebounding numbers went down. From 1999 to 2008, only 28.5% of rebounds were offensive rebounds.

Good
--Spreading the court more. This is the most obvious plus, at least on the surface. If you’ve got perimeter passing, you’ll get more isolation sets and better spacing. At least in theory. But this may be offset by…

Neutral
--Fewer Assists. There’s less ball movement with three point shooting. Total assist numbers are way, way down…from a peak of over 26 a game in the mid-80s to a little over 21 a game today. Pace doesn’t account for that 20% drop…the difference in pace from the mid 80s to today is around 10%. Percentage of made shots credited with assists stayed at around 61-62% from the 80s through the early 90s. It has dropped slightly in recent years. But it should probably drop more than that, because ...

Bad
--Fewer fouls/points scored at the line. Three point shots are less contested and less likely to receive fouls. Consequently, shot attempts on teams that shoot fewer threes get an extra push. It’s hard to determine the exact amount, but it’s considerable…as noted above, you get more points from shooting the three. And a smaller percentage of points comes from free throws now—from 1984-87, around 20.5% of points came from free throws. That dropped about 1 to 1.5% in the early 1990s…right when the numbers of threes attempted went up. It’s at about 19% now This offsets the idea that you get more points per shot from the three (as noted above), because the end result is…

Neutral (again)
--Virtually identical TS%. See numbers at the beginning of the post. The bottom line is that teams are not more or less efficient than they once were at producing points. The most efficient period in NBA history was, again, the early and mid-1980s. The least efficient since the three is the late 90s, by a lot. It was on an upswing in recent years, although we’re down again this season.

What does this all say? Offensive styles run in trends. Shooting more threes isn’t “better”…it’s more current. You could run a perfectly good offense that would be just as efficient and effective as any team's today and take 200 threes a season. But players are not taught to do this because--well, because they aren't. As shown above, it's not a matter of scoring effectiveness or offensive efficiency. The tendency by many posters is to think that something that is pretty much cyclical over time—like offensive style and production—is progressive. Basketball, like baseball, is not like that. There are all sorts of contributing factors to changes in style, but those changes don’t make something better—it just makes the thing different than it was.

Old thread I know but still relevant. I haven't bothered to consider all your points but your analysis re: offensive rebounds is seriously flawed. review the data more closely and you will see there is no relationship showing three point shots reduce offense rebounds. None. The conclusion you draw may be correct (although I think it is not) but not for the reasons suggested. Specifically go to the period around the time when the three point line was reduced. Review this and you will clearly see the relationship you suggest is not there. While offensive rebounds as a percentage of shots has fallen consistently, there is some other reason.

I'm not trying to bust any chops here but would like to discuss the benefits of a system like Phoenix' during the D' Antoni years and the three was central to that strategy.

[Nivek]

Good post TrueLAFan.

Sundamental: Roland at 82games took a look at offensive rebounding rates by FGA back in 2004.

Code: Select all

TYPE    ORB%
Jumper  28.5%
Close   34.4%
Dunks   29.6%
FTA     13.9%


Suggesting that TrueLAFan's point that 3pt attempts are offensive rebounded less frequently is accurate.

[Sundamental]

Good post TrueLAFan.

Sundamental: Roland at 82games took a look at offensive rebounding rates by FGA back in 2004.

Code: Select all

TYPE    ORB%
Jumper  28.5%
Close   34.4%
Dunks   29.6%
FTA     13.9%


Suggesting that TrueLAFan's point that 3pt attempts are offensive rebounded less frequently is accurate.

I acknowledged that it is possible the conclusion is correct but TrueLAfans reasoning is incorrect. I'm trying to be gentle but there is no other way to state it. I told you what to look for in my previous post but obviously you just accept what people tell you. I don't do that so here's what I did to see if TrueLAfans point made sense. I went back twenty years and collected data for the average number of threes and the percentage of total rebounds that are offensive. As TLF said, the number of offensive rebounds as a percentage of the total has steadily dropped and the total number of threes has steadily increased. They seem to be a direct correlation. But heres the critical piece he missed. From the seasons ended 1995 to 1997, the line was moved in and the number of threes exploded. But guess what? When the line was moved back out in the season ended 1998 the number of threes dropped substantially (as would be expected) but the offensive rebound rate didn't increase as TLF would assume. It just kept dropping. So clearly something else here is at play. I have all this data on a spreadsheet but I'm terrible with computers so it's hard to post. I'll do it if necessary.

So this gets back to my problem. I'm trying to look at the efficiency of threes besides just the numbers. I thought TLF arguments were well presented so I gave one of them some thought and review and found it to be incorrect. How many more will also be incorrect upon further review? Don't know but I'd still like to discuss this topic of three efficiency further if anyone is interested. Thanks for the link to 82games. I will review it.

[Nivek]

Gotcha. I do agree that the rise in 3pt attempts is not necessarily related to the drop in offensive rebounding. I was mainly pointing out that jumpers do get offensive rebounded less frequently than other kinds of shots. I think the biggest reason offensive rebounding has dropped off is coaching. Most coaches would rather their guys get back on defense to stop easy transition baskets than commit extra resources for a marginally better chance at getting an offensive rebound. The percentage play is in getting back on defense.

[TrueLAfan]

I am sure that there are other factors that have affected offensive rebounding. I am equally sure that one of the main factors factor, is the increased number of threes. The percentage of offensive rebounds as a percentage of shots missed (and total team rebounds) stayed pretty consistent between 1977 and 1991. The number of threes taken was minute in that period...the percentage of shot attempts that were threes didn't go over 4% until 1987. Since 1990, offensive rebound percentages have dropped sharply and, largely, consistently since then. There have been all sorts of shifts since then, most of which have self-corrected at some point. Three point attempts are one of the few things that continued to increase in correlation with the decrease in offensive rebounding.

And while you can find individual teams that buck the trend, the general assessment seems fairly clear. I looked at the top 5 teams in 3PA in 2010, 2007, 2005, 2002, 2000, 1998, 1994, and 1990. The correlation certainly seems pretty strong and consistent. There have been 18 seasons with the “normal” three since 1991, and I looked at eight. In seven of the eight, the 5 teams at the bottom in 3PA had a higher percentage of offensive rebounds than the five teams at the top in 3PA. The top five teams in three point attempts grab, on average, about 2.3% fewer offensive rebounds as a percentage of total team rebounds, than the bottom five teams in 3PA. (That may not seem like, much, but it's really a substantial amount and difference). In other words, take fewer threes—get more offensive rebounds. More threes taken—fewer offensive rebounds.

While it's true that the 1995-7 periods didn't have a great effect as might be expected, there's a couple of things to consider. One is that 1998, the year the line was pushed back out, was one of two years in the past 25 when there was a rise of greater than .3% in percentage of offensive rebounds. So there did seem to be some effect. The second is that when the line was pushed back, the number of threes taken really didn't drop by much. The percentage of 3PA jumped from under 12% to 19% in 1995. When the line was pushed back, the percentage only dropped to about 17%, and continued to climb after that. It's close to 22.5% now.

I think teams have adjusted to changes in offense produced by threes, and this has included changes in rebounding. I think changes in offensive and defensive styles and rule changes have exacerbated the effect on offensive rebounding. And, like I said, individual teams certainly can and have worked around this. But I think it's pretty clear that the shift to taking 13+ threes a game has significantly reduced the percentage of offensive rebounds as opposed to when teams took 5-7 threes a game. Other factors have surely been involved but, like I said, the evidence (including info at 82games.com) points toward the increase in threes being a major contributor.

[Sundamental]

Just look at the data as I suggested. You are using incomplete data and therefore get an incomplete conclusion. Do as I suggest and you will no longer be so sure. Yes, offensive rebounds have been on the decline but not because of three pointers.

[floppymoose]

Regarding the overall efficiency remaining pretty constant… I have a theory on this (for which I have no data to support, but hey, I *could* be right):

The NBA adjusts the rules, and the refs adjust their style, to preserve the desired balance between offense and defense. The balance is not always the goal, but it has tended to work out that way.

Years ago the NBA wanted more scoring, so they made the dumb illegal defense rules outlawing zones. The result was that the offense could force four defenders to one side of the floor to play one on one with their best player. This didn't really increase scoring, but it did increase snoring. So eventually the league figured out that team play vs a real defense was more fun to watch and adjusted the rules.

Ok so I'm rambling away from my original thesis. But that doesn't mean I don't believe it. The basic balance between O and D is central to the entertainment factor, and the league adjusts to preserve this.