I would be remiss if I didn't highlight Tom Haberstroh's rather excellent piece on the NBA's (really) slow learners:
http://espn.go.com/nba/playoffs/2015/st ... -3-pointer
Were he allocated a few more words, I would have loved to have seen an elaboration on particular historical episodes and other non-turning points, where there are very interesting stories to be told and things to be learned - the relative failure of the shortened 3-point line in advancing modern strategy, especially. (And c'mon, Tom, charts ought be shown on a per possession not per game basis!)
But overall, a wonderful piece, gently presented but absolutely brutal in its implications, illustrating the malign effects of the nexus of arrogance and incompetence at the Hall of Fame level - coaches especially.
And for all those who reflexively and instinctively espouse the view that the game has ever and always been played at the highest level of professional competence, hopefully some eyes will crack open.
How the NBA bombed the bomb
Re: How the NBA bombed the bomb
Nobody would espouse this. Only the elite teams (at most) have played at the highest level. By definition.schtevie wrote:... the view that the game has ever and always been played at the highest level of professional competence,..
Re: How the NBA bombed the bomb
Thanks, Schtevie!
(this is Tom by the way)
Great suggestions for another piece. My editors would cry if I asked for more words!
And yes, the per-100-possessions scale probably would have put smiles on the faces in this corner, but for a feature story like this for the masses, I think per-game was mostly harmless. I tried to use 3-point frequency as a % of FGA as much as I could.
Thanks for the feedback,
T
(this is Tom by the way)
Great suggestions for another piece. My editors would cry if I asked for more words!
And yes, the per-100-possessions scale probably would have put smiles on the faces in this corner, but for a feature story like this for the masses, I think per-game was mostly harmless. I tried to use 3-point frequency as a % of FGA as much as I could.
Thanks for the feedback,
T
Re: How the NBA bombed the bomb
In the article, there's a bar graph showing the number of 40% shooters (from the arc) each season. That number jumps seriously in the 3 years of the short arc ('94 to '96), and in '97 there appear to be 28 -- and this number has varied from about 15 to 40 every year since. This year, 23.
But we need to know on how many attempts (or makes). A couple paragraphs up, there's mention of the Miami '13 having 5 guys with 150 3FGA at 40% or better.
So in '97 there are 28 players with at least 100 attempts and 40%. This year, 23. So that answers that.
And the 2013 Heat had 5 of 24 players in the league with 40% on 175+ shots.
LeBron hit .406, and his next best 3fg% is .379 the next year.
Chalmers hit .409, also his best year, between 2 in the .380's.
Ray Allen's .419 was the last of many times he hit .400
Battier's .430 was off his own chart, almost .100 higher than previous and subsequent seasons.
Mike Miller's .417 was only his 7th best 3fg%
It's hard to corral one-fifth of the league's best shooters, especially when some of them are going to have their career best shooting year.
The article quotes Hubie Brown, who praises Louie Dampier of the Colonels. Hubie led Kentucky to their only title, and in so doing he cut the team's 3fga from 4.2 per game, down to 2.3. Dampier, who had attempted 550 a couple of times, was under 100.
But the bigger ABA story is that all teams cut down on their 3-pt shooting. In 9 years, it was never a big winner. From a high of 6.3 per game in 1970 to under 4.0 over the last 4 seasons. Never did the league top .300 from the arc.
That's a pretty big volume of data, in which 77 players shot at least 100 treys, another 100 shot at least 20. Only 13% of these guys managed to hit 1/3 of them.
It's not like NBA coaches hated the shot for no good reason. Only after colleges started winning with it, and lots of players knew how to use it, did it take off in the NBA.
But we need to know on how many attempts (or makes). A couple paragraphs up, there's mention of the Miami '13 having 5 guys with 150 3FGA at 40% or better.
So in '97 there are 28 players with at least 100 attempts and 40%. This year, 23. So that answers that.
And the 2013 Heat had 5 of 24 players in the league with 40% on 175+ shots.
LeBron hit .406, and his next best 3fg% is .379 the next year.
Chalmers hit .409, also his best year, between 2 in the .380's.
Ray Allen's .419 was the last of many times he hit .400
Battier's .430 was off his own chart, almost .100 higher than previous and subsequent seasons.
Mike Miller's .417 was only his 7th best 3fg%
It's hard to corral one-fifth of the league's best shooters, especially when some of them are going to have their career best shooting year.
The article quotes Hubie Brown, who praises Louie Dampier of the Colonels. Hubie led Kentucky to their only title, and in so doing he cut the team's 3fga from 4.2 per game, down to 2.3. Dampier, who had attempted 550 a couple of times, was under 100.
But the bigger ABA story is that all teams cut down on their 3-pt shooting. In 9 years, it was never a big winner. From a high of 6.3 per game in 1970 to under 4.0 over the last 4 seasons. Never did the league top .300 from the arc.
That's a pretty big volume of data, in which 77 players shot at least 100 treys, another 100 shot at least 20. Only 13% of these guys managed to hit 1/3 of them.
It's not like NBA coaches hated the shot for no good reason. Only after colleges started winning with it, and lots of players knew how to use it, did it take off in the NBA.
Re: How the NBA bombed the bomb
Mike's additions about the ABA are worthwhile. There is more to the story though. Hubie cut the Colonels 3pt attempts in 74-75 but mainly the attempts of guys other than Dampier. The Dampier drastic cut occurred many years before under other coaches. Perhaps tk try to curtail the unqualified copycats. Dampier was hitting almost 40% in 74-75. But Issel only hit 18% in 73-74 and most of the others taking them were weak to really awful. Dampier was way better taking 3s than 2s. His 3 pt shooting was a main strength that probably should not have been curtailed. After his rookie season he had 8 straight seasons with fine 3pt FG% by today's standards and even better by the performance level then. Hubie's clamp down was right in that its further impact on the unfit to shoot 3s but not with Dampier. The 73-74 to 74-75 clamp down was also good in that it found 58% FG% Artis Gilmore 3 more shots per game. That was an alternative that not everyone had then or now. But there were still way too many mid-range 2s.
Re: How the NBA bombed the bomb
I saw this article today http://fivethirtyeight.com/features/liv ... ly-a-myth/ and was hostile to it. Live by the three, die by the three is better judged on its own game to game than on average with everything else, imo. 3 pt attempt frequency and accuracy is a minority part of game scoring. Pace matters too. I'd rather have seen game to game over and under achievement of actual points off 3 pointers compared to expected points for that quality of shots based on league and team averages for 3s and all types of attempts. Everything else in game scoring is at least indirect and partly to mostly introduced noise in the study of the effect of using the 3.
Re: How the NBA bombed the bomb
This article by Haberstroh is a sort of micro-level, player's-eye view update of his article about the history of the (non-)usage of the 3-pointer.
Aside from putting us inside Chris Bosh's head as he tries to adjust his defense to contend with the stretch 4s and stretch 5s now swarming onto NBA courts, the article notes the rapidly rising percentage of FGAs that are 3-pointers. I.e. we've probably passed the tipping point, as it now takes willful ignorance to be unaware of the inherent superiority of the 3-pointer over anything except an inside shot.
Which is fine from a strategic viewpoint but bad from an aesthetic viewpoint. I agree 100% with what Popovich said in Haberstroh's first article: a team that wants to win will shoot more 3-pointers, because the NBA rules give you that extra point. But like Popovich I have never liked the rule in the first place, not from its NBA inception in 1979. It unbalances the game and will turn it into a 3-point shooting contest. I didn't expect it to take over 35 years, but we're about to reach peak 3-point shooting. It won't be long before we see teams shooting over half of their FGs from 3-point range, and the NBA will have to make some sort of rules adjustment.
What that adjustment will be, I don't know. I'd vote for abolishing the 3-pointer, but the NBA won't do anything that drastic. They might change the rules about fouling and free throws to increase the profitability of working the ball inside, but I don't know how effective that will be; a player with the ball and position inside still remains the most potent offensive weapon, but that's not enough to counterbalance the rain of 3s that we will increasingly see.
Aside from putting us inside Chris Bosh's head as he tries to adjust his defense to contend with the stretch 4s and stretch 5s now swarming onto NBA courts, the article notes the rapidly rising percentage of FGAs that are 3-pointers. I.e. we've probably passed the tipping point, as it now takes willful ignorance to be unaware of the inherent superiority of the 3-pointer over anything except an inside shot.
Which is fine from a strategic viewpoint but bad from an aesthetic viewpoint. I agree 100% with what Popovich said in Haberstroh's first article: a team that wants to win will shoot more 3-pointers, because the NBA rules give you that extra point. But like Popovich I have never liked the rule in the first place, not from its NBA inception in 1979. It unbalances the game and will turn it into a 3-point shooting contest. I didn't expect it to take over 35 years, but we're about to reach peak 3-point shooting. It won't be long before we see teams shooting over half of their FGs from 3-point range, and the NBA will have to make some sort of rules adjustment.
What that adjustment will be, I don't know. I'd vote for abolishing the 3-pointer, but the NBA won't do anything that drastic. They might change the rules about fouling and free throws to increase the profitability of working the ball inside, but I don't know how effective that will be; a player with the ball and position inside still remains the most potent offensive weapon, but that's not enough to counterbalance the rain of 3s that we will increasingly see.
Re: How the NBA bombed the bomb
That saturation point of 3fga/fga may or may not be reached quite yet. And even if it's not, there may be an offsetting feature, independent of the relative efficiency.
Pounding the ball inside, getting layups and FT and putbacks -- when a team is good at this -- may also be more consistent than shooting 3's.
When a 3-point shooter misses 3 in a row, that doesn't really mean he's having a bad shooting night. It's par for the course: A 38% shooter who takes 10 a night may well miss 3 straight. But if he keeps missing, and his sidekick is also not connecting, that team is in for some hurt.
In another thread, we saw that older teams tend to win more 1- and 2-point games than they should. This results in their winning more than their MOV says they should. Is it possible that 3-pointer-centric teams tend to win more or less than they should?
From the last 3 years, 90 team-seasons, I ranked teams in order of their 3FGA/FGA ratio. Dividing this sample into 5ths -- 18 team-seasons each -- I also calculate the games they 'should have won' based on their avg margin of victory.
The averages:Teams that shoot the most 3's also tend to be the better teams; and they probably do other things right.
Teams that shoot fewer 3's get progressively weaker -- and weaker yet in the close games?
Suddenly that bottom quintile reverses things: They tend to win 1.25 more games than others, given their MOV.
EDIT -- The Memphis Grizzlies own 3 of the 4 lowest spots on the 3FGA/FGA continuum. They're also 25-11 in decisions of 3 pts or less; AND I estimate they had about 14 'extra' wins in the interval.
Without the Grizz, what's left of the bottom quartile averages just .55 more wins than expected.
Pounding the ball inside, getting layups and FT and putbacks -- when a team is good at this -- may also be more consistent than shooting 3's.
When a 3-point shooter misses 3 in a row, that doesn't really mean he's having a bad shooting night. It's par for the course: A 38% shooter who takes 10 a night may well miss 3 straight. But if he keeps missing, and his sidekick is also not connecting, that team is in for some hurt.
In another thread, we saw that older teams tend to win more 1- and 2-point games than they should. This results in their winning more than their MOV says they should. Is it possible that 3-pointer-centric teams tend to win more or less than they should?
From the last 3 years, 90 team-seasons, I ranked teams in order of their 3FGA/FGA ratio. Dividing this sample into 5ths -- 18 team-seasons each -- I also calculate the games they 'should have won' based on their avg margin of victory.
The averages:
Code: Select all
3/fga W+ expW W
.320 .45 48.4 48.8
.282 -.21 45.1 44.8
.252 -.68 42.2 41.5
.229 -.82 32.7 31.8
.202 1.25 36.7 37.9
Teams that shoot fewer 3's get progressively weaker -- and weaker yet in the close games?
Suddenly that bottom quintile reverses things: They tend to win 1.25 more games than others, given their MOV.
EDIT -- The Memphis Grizzlies own 3 of the 4 lowest spots on the 3FGA/FGA continuum. They're also 25-11 in decisions of 3 pts or less; AND I estimate they had about 14 'extra' wins in the interval.
Without the Grizz, what's left of the bottom quartile averages just .55 more wins than expected.
Re: How the NBA bombed the bomb
Mike T, I thank you for revisiting this topic, for the timeliness of the comment as well as the lovely coincidence. What I mean by the latter is that your notification in late July of old friend Dean's parting of ways with the Kings is what got me thinking that a "celebration" of the anti-analytic/progressive NBA would be in order - front offices, coaches, players...the whole lot.
In that vein, I began thinking about what besides the slow adoption of three-point shooting has all along been hiding in plain view in the empirical record? Said another way, what box score-availabe stats (with perhaps minor input from more involved analytic results) offer prima facie evidence of sub-optimal play, with significant implications for competitive performance?
I will offer some thoughts on this topic another day, however, my implicit assumption had been that the long, twilight struggle with 3s was over and won and that all that was left was mopping up operations. But then I thought it worth actually revisiting the empirical record, and now I am not so sure that the anticipated 3PA long-run "equilibrium" will arrive any year soon.
In fact, I had been planning on starting a new post, titled "NBAARP", as I wouldn't be surprised if it will be a half century before an expected stability in terms of 3PAr is realized. Stated otherwise (in a way that is amusing to me anyway) it could well be the case that a child born in 1979-80 will get her AARP card before the NBA has finally taken maximum, collective advantage of the three-point shot.
But in the interest of brevity, rather than elaborate on the evidence regarding the on-going slow adoption of the 3, let me address some of the issues you raise as well as the recent Tom Haberstroh article cited. First, the latter.
I am much less pleased with this offering of Tom's than the earlier piece of work that motivated the praise in my originating comments. It's less the anecdata and more its fourth paragraph - what provides numbers - that is very confusing to me. I cannot reference his stated percentages, and then there is the big no-no-no-no-no of using a strike-shortened year as a basis of comparison for current 3PAs.
But no matter. Underlying my displeasure is my long-standing skepticism of the "stretchiness" meme. Not that I don't believe that there are more 3PA taken by "stretchy" players, rather that there's some special story to be told about this phenomenon. What I want to know is how the increase in 3PA by "stretchies" has compared to the general increase seen in the NBA. That is what matters, no?
So, I decided to query the database at Basketball Reference to that effect.
I define "stretchies" as centers (C, C-F, and F-C) and 6'9" and above forwards (per B-R and Tom Haberstroh specification) who have realized 3P% of 0.267 and above. Why 0.267? Well, this corresponds to a 2P% equivalent of 0.400, what is the approximate, average value of all (crappy) mid-range shots. Said another way, any tall guy shooting less than this percentage isn't stretching any defense.
So, what do you find if you compare 1999-2000 to 2014-15 three-point shooting for this group, compared to the remainder of the NBA? Let me phrase this in the form of a quiz. Of the increase in 3PAs per game realized between 1999-2000 and 2014-15, how many can be attributed to increased "stretchiness"? That is what is the increase in 3PAs per game taken by "stretchies" above the general increase observed by "non-stretchies" (aka everyone else)?
Well, the answer I got (and please someone check this) is....wait for it...1.7 3PAs per game. A realized gain of about 0.1 3PAs per year. Wow?
So, assuming this is correct, what is going on, and why the attention paid?
Well, my conjecture is that folks are simply getting (over) excited about the final frontier. Bigs are the last category of players that have been encouraged to adapt to this now age-old rule. (After this there aren't any position/minutes really left to change over, unless one has high hopes for true centers and Ricky Rubio.) As such, this phenomenon is confirmation of the historical failure. And so what remains is "stretchies" (and their shorter colleagues) increasing their average 3PAs per game until this whole sorry adaptation process comes to its conclusion...some year.
OK. Finally a comment on Mike T's comment about aesthetics: be careful what you wish for, as it may be very, very ugly.
By this I mean that whatever one's views on the "prettiness" of the 3, it is important to recognize that the NBA has evolved since the so-called Golden Age of the 80s. In particular, there is very good evidence that defenses have substantively improved over time, and an NBA without the space-creating features of the 3 could be a very, very unaesthetic game.
In that vein, I began thinking about what besides the slow adoption of three-point shooting has all along been hiding in plain view in the empirical record? Said another way, what box score-availabe stats (with perhaps minor input from more involved analytic results) offer prima facie evidence of sub-optimal play, with significant implications for competitive performance?
I will offer some thoughts on this topic another day, however, my implicit assumption had been that the long, twilight struggle with 3s was over and won and that all that was left was mopping up operations. But then I thought it worth actually revisiting the empirical record, and now I am not so sure that the anticipated 3PA long-run "equilibrium" will arrive any year soon.
In fact, I had been planning on starting a new post, titled "NBAARP", as I wouldn't be surprised if it will be a half century before an expected stability in terms of 3PAr is realized. Stated otherwise (in a way that is amusing to me anyway) it could well be the case that a child born in 1979-80 will get her AARP card before the NBA has finally taken maximum, collective advantage of the three-point shot.
But in the interest of brevity, rather than elaborate on the evidence regarding the on-going slow adoption of the 3, let me address some of the issues you raise as well as the recent Tom Haberstroh article cited. First, the latter.
I am much less pleased with this offering of Tom's than the earlier piece of work that motivated the praise in my originating comments. It's less the anecdata and more its fourth paragraph - what provides numbers - that is very confusing to me. I cannot reference his stated percentages, and then there is the big no-no-no-no-no of using a strike-shortened year as a basis of comparison for current 3PAs.
But no matter. Underlying my displeasure is my long-standing skepticism of the "stretchiness" meme. Not that I don't believe that there are more 3PA taken by "stretchy" players, rather that there's some special story to be told about this phenomenon. What I want to know is how the increase in 3PA by "stretchies" has compared to the general increase seen in the NBA. That is what matters, no?
So, I decided to query the database at Basketball Reference to that effect.
I define "stretchies" as centers (C, C-F, and F-C) and 6'9" and above forwards (per B-R and Tom Haberstroh specification) who have realized 3P% of 0.267 and above. Why 0.267? Well, this corresponds to a 2P% equivalent of 0.400, what is the approximate, average value of all (crappy) mid-range shots. Said another way, any tall guy shooting less than this percentage isn't stretching any defense.
So, what do you find if you compare 1999-2000 to 2014-15 three-point shooting for this group, compared to the remainder of the NBA? Let me phrase this in the form of a quiz. Of the increase in 3PAs per game realized between 1999-2000 and 2014-15, how many can be attributed to increased "stretchiness"? That is what is the increase in 3PAs per game taken by "stretchies" above the general increase observed by "non-stretchies" (aka everyone else)?
Well, the answer I got (and please someone check this) is....wait for it...1.7 3PAs per game. A realized gain of about 0.1 3PAs per year. Wow?
So, assuming this is correct, what is going on, and why the attention paid?
Well, my conjecture is that folks are simply getting (over) excited about the final frontier. Bigs are the last category of players that have been encouraged to adapt to this now age-old rule. (After this there aren't any position/minutes really left to change over, unless one has high hopes for true centers and Ricky Rubio.) As such, this phenomenon is confirmation of the historical failure. And so what remains is "stretchies" (and their shorter colleagues) increasing their average 3PAs per game until this whole sorry adaptation process comes to its conclusion...some year.
OK. Finally a comment on Mike T's comment about aesthetics: be careful what you wish for, as it may be very, very ugly.
By this I mean that whatever one's views on the "prettiness" of the 3, it is important to recognize that the NBA has evolved since the so-called Golden Age of the 80s. In particular, there is very good evidence that defenses have substantively improved over time, and an NBA without the space-creating features of the 3 could be a very, very unaesthetic game.
Re: How the NBA bombed the bomb
That is a very good point, indeed we could see some signs of the NBA moving in that direction even in 1979. And even with the 3-pointer, we witnessed what Keith Ellis on the old APBR email list liked to call the deadball era of the NBA, the excruciating slow-down late 1990s.schtevie wrote:[...] it is important to recognize that the NBA has evolved since the so-called Golden Age of the 80s. In particular, there is very good evidence that defenses have substantively improved over time, and an NBA without the space-creating features of the 3 could be a very, very unaesthetic game.
To my mind the best solution would've been not the 3-pointer, but the changes in hand-checking rules which gave quick dribble-drive penetrators an advantage. Would that have been enough? Maybe not. But perhaps in conjunction with other speed-up rules such as the 8-second backcourt rule, the combination would've been enough.
That still leaves the question of what should the NBA do a few years from now when the number of 3-point FGAs gets too out of hand. I think this will take some brainstorming and some experimentation in the NBDL and other leagues. If we're going to keep having 3-point FGs, I'd like to see the line moved out so that the shot is more difficult. But theres not enough space in the corners as it is, so another useful step that would open up the game would be to make the court wider. But that's a drastic step that won't happen for years, if ever. So I do not have a definitive suggestion.
Re: How the NBA bombed the bomb
3 points for dunks would balance the opportunity locations for that extra point.
4 on 4 would help remedy the space issue for a court that is ridiculously crowded and has a lot of pretty extraneous population for a game with one ball. We have gotten use to seeing basketball 5 on 5 but imo you don't NEED five to play a functional, diverse, beautiful game. I'd argue 4 on 4 would be more enjoyable to watch. It was more fun to me than 5 on 5 when I played. An increase in the depth of the baseline behind the basket would help open up more drive opportunities.
I would only award 2 free throws on a foul of a 3 pt shot. 3 is an over-reward given the expected value of a 3 pt shot vs the expected value of 3 free throws. Fouling a 3pt play should be a legitimate, effective defensive option similar to reducing the expected points of a driver at the rim if you foul him in time compared to him scoring the fga. The tradeoff should not go one way in one area of the court and the opposite elsewhere. I would not give a foul for light arm contact with the shooter after the shot or well after. I would for body contact or use a technical for wanton foot endangerment. 4 pt plays are another issue. I would probably non call all but strong body contact. Interior plays and buckets often face a lot of contact that is not called a foul. 3 pt shooters don't deserve an unfair, pristine level of non-contact.
4 on 4 would help remedy the space issue for a court that is ridiculously crowded and has a lot of pretty extraneous population for a game with one ball. We have gotten use to seeing basketball 5 on 5 but imo you don't NEED five to play a functional, diverse, beautiful game. I'd argue 4 on 4 would be more enjoyable to watch. It was more fun to me than 5 on 5 when I played. An increase in the depth of the baseline behind the basket would help open up more drive opportunities.
I would only award 2 free throws on a foul of a 3 pt shot. 3 is an over-reward given the expected value of a 3 pt shot vs the expected value of 3 free throws. Fouling a 3pt play should be a legitimate, effective defensive option similar to reducing the expected points of a driver at the rim if you foul him in time compared to him scoring the fga. The tradeoff should not go one way in one area of the court and the opposite elsewhere. I would not give a foul for light arm contact with the shooter after the shot or well after. I would for body contact or use a technical for wanton foot endangerment. 4 pt plays are another issue. I would probably non call all but strong body contact. Interior plays and buckets often face a lot of contact that is not called a foul. 3 pt shooters don't deserve an unfair, pristine level of non-contact.
Re: How the NBA bombed the bomb
So, some empirical notes on the subject of achieving a long-run "equilibrium" 3PAr in the NBA.
Last year, the NBA saw 44.3% of shots taken in the mid-range, defined in B-R as 3 ft out to the 3-point line, with only 39.2% completed - far less rewarding than the 26.8% of shots taken beyond the arc that yielded an eFG% of 52.5. This is noted (not to argue that a shot from 3.1 feet is unquestionably mid-range but) solely for general perspective on the potential share of shots that could be profitably shifted to beyond the arc.
What might one imagine that the pace of such change will be? Over the 36 year life of the rule, the increase in 3PAr has averaged an absurdly slow 0.7%age points per year. Ignoring the first five year period of retrogression doesn't alter the picture much, it's been 0.8%age points per year since then. But now we stand at an all-time high average 3PAr of 0.268, so how much more can it go up?
Well, the pace of change has quickened over the last five years, and now that it is generally recognized that the cool kids are the ones shooting 3s, perhaps this recent "surge" to a rate of convergence of 0.9%age points per year is a reasonable estimate of what to expect?
But then there is the very open question as to what the long-run equilibrium value of 3PAr might be (assuming no subsequent rule changes). The gap between average and "best" practice (i.e. NBA average 3PAr and that of the team with the highest ratio) has not begun to diminish over recent years, averaging over 10&age points over the past 15 years, with last year's push by the Rockets showing an all-time high of 12.4%age points.
So, if Rockets-ball is the correct vision of the (average) future (what still leaves 29.9% of shots in the mid-range) and the rate of convergence is what we've seen over the (faster) last five years, then we are looking at the long-run equilibrium arriving about 16 years from now. AARP it is! (Or at least I think it would be unwise to believe that it will be anything much less than that.)
One objection, informing a belief that the end might arrive sooner rather than later, is that the NBA is soon to become supply constrained. This is to say that, over time, the increased share of 3PAr has depended, in part, on new "3P able" shooters entering the league, and per my previous remark about centers being centers and Ricky Rubio being Ricky Rubio, there aren't many additional minutes that remain available for such players, implying that the equilibrium target isn't last year's Rockets but something less.
Well, this is a serious argument, but I don't think it has that much empirical bite. Taken at face value, suppose that the NBA is converging, not on last year's Rockets, but on the average of last year's Top 5 3PAr teams. This would imply an increase in the NBA average of 7.4%age points, what at the recent pace of convergence would take "but" 8 years. However a corollary of the "supply-constrained" argument might also be a decrease in the rate of convergence.
But then there are two other objections. First, supply is endogenous; call this the Larry Bird effect. As one may recall, he entered the league in the inaugural year of the 3-point shot, and hit 0.406 on 143 attempts, but both stats plummeted the following two years, owing to indifference and what must have been much poorer shot selection/opportunity cost. So, the point is that there are many current "insufficiently good" shooters in the NBA, owing only to lack of direction and opportunity.
But the other objection to the "supply-constrained" argument is that there remains lots of room to improve even assuming supply is fixed at current levels of effective 3-point shooters: who last year used 81.7% of NBA minutes.
Last year, there was only a moderate negative correlation between the fraction of team minutes played by such players (previously defined as those realizing a 3P% less than 0.267) and 3PAr: -0.46. Based on this (assumed linear relationship) were the average team to eliminate all it's "sub-par" minutes (i.e. having all its players realizing 3p% of 0.267 or greater - something akin to what the Atlanta Hawks realized) the suggested corresponding increase in 3PAr is only 3.9%age points - what is considerably less than the observed range. So, even assuming (incorrectly) a hard and binding supply constraint, there's a lot of room left for strategic adoption.
It looks like it's gonna be years and years...
Last year, the NBA saw 44.3% of shots taken in the mid-range, defined in B-R as 3 ft out to the 3-point line, with only 39.2% completed - far less rewarding than the 26.8% of shots taken beyond the arc that yielded an eFG% of 52.5. This is noted (not to argue that a shot from 3.1 feet is unquestionably mid-range but) solely for general perspective on the potential share of shots that could be profitably shifted to beyond the arc.
What might one imagine that the pace of such change will be? Over the 36 year life of the rule, the increase in 3PAr has averaged an absurdly slow 0.7%age points per year. Ignoring the first five year period of retrogression doesn't alter the picture much, it's been 0.8%age points per year since then. But now we stand at an all-time high average 3PAr of 0.268, so how much more can it go up?
Well, the pace of change has quickened over the last five years, and now that it is generally recognized that the cool kids are the ones shooting 3s, perhaps this recent "surge" to a rate of convergence of 0.9%age points per year is a reasonable estimate of what to expect?
But then there is the very open question as to what the long-run equilibrium value of 3PAr might be (assuming no subsequent rule changes). The gap between average and "best" practice (i.e. NBA average 3PAr and that of the team with the highest ratio) has not begun to diminish over recent years, averaging over 10&age points over the past 15 years, with last year's push by the Rockets showing an all-time high of 12.4%age points.
So, if Rockets-ball is the correct vision of the (average) future (what still leaves 29.9% of shots in the mid-range) and the rate of convergence is what we've seen over the (faster) last five years, then we are looking at the long-run equilibrium arriving about 16 years from now. AARP it is! (Or at least I think it would be unwise to believe that it will be anything much less than that.)
One objection, informing a belief that the end might arrive sooner rather than later, is that the NBA is soon to become supply constrained. This is to say that, over time, the increased share of 3PAr has depended, in part, on new "3P able" shooters entering the league, and per my previous remark about centers being centers and Ricky Rubio being Ricky Rubio, there aren't many additional minutes that remain available for such players, implying that the equilibrium target isn't last year's Rockets but something less.
Well, this is a serious argument, but I don't think it has that much empirical bite. Taken at face value, suppose that the NBA is converging, not on last year's Rockets, but on the average of last year's Top 5 3PAr teams. This would imply an increase in the NBA average of 7.4%age points, what at the recent pace of convergence would take "but" 8 years. However a corollary of the "supply-constrained" argument might also be a decrease in the rate of convergence.
But then there are two other objections. First, supply is endogenous; call this the Larry Bird effect. As one may recall, he entered the league in the inaugural year of the 3-point shot, and hit 0.406 on 143 attempts, but both stats plummeted the following two years, owing to indifference and what must have been much poorer shot selection/opportunity cost. So, the point is that there are many current "insufficiently good" shooters in the NBA, owing only to lack of direction and opportunity.
But the other objection to the "supply-constrained" argument is that there remains lots of room to improve even assuming supply is fixed at current levels of effective 3-point shooters: who last year used 81.7% of NBA minutes.
Last year, there was only a moderate negative correlation between the fraction of team minutes played by such players (previously defined as those realizing a 3P% less than 0.267) and 3PAr: -0.46. Based on this (assumed linear relationship) were the average team to eliminate all it's "sub-par" minutes (i.e. having all its players realizing 3p% of 0.267 or greater - something akin to what the Atlanta Hawks realized) the suggested corresponding increase in 3PAr is only 3.9%age points - what is considerably less than the observed range. So, even assuming (incorrectly) a hard and binding supply constraint, there's a lot of room left for strategic adoption.
It looks like it's gonna be years and years...
Re: How the NBA bombed the bomb
Last season, the NBA had an aggregate TS% of .534
This of course includes FT, layups, "midrange", and 3's.
If you hit less than .356 from the arc, you're taking a below-avg shot out there.
You'll occasionally get a FT (or 3) from the arc; and the league hit .350 on avg.
Has equilibrium been reached?
About 40% of NBA minutes were by players who hit .350 or better from the arc. But some took only a few.
And 31% of minutes were from players who hit .350 while at least 25% of their FGA were 3s.
Not quite 40% of player minutes were with players who hit 1/3 of their 3s and who opted to take 1/4 or more from the arc.
Some 85% of minutes were by players who hit any 3s last year.
This of course includes FT, layups, "midrange", and 3's.
If you hit less than .356 from the arc, you're taking a below-avg shot out there.
You'll occasionally get a FT (or 3) from the arc; and the league hit .350 on avg.
Has equilibrium been reached?
About 40% of NBA minutes were by players who hit .350 or better from the arc. But some took only a few.
And 31% of minutes were from players who hit .350 while at least 25% of their FGA were 3s.
Not quite 40% of player minutes were with players who hit 1/3 of their 3s and who opted to take 1/4 or more from the arc.
Ranking players by 3fg%, regardless of frequency, it looks like that 81.7% lines up with players who shot .167 or better from the arc....effective 3-point shooters: who last year used 81.7% of NBA minutes.
Some 85% of minutes were by players who hit any 3s last year.
Re: How the NBA bombed the bomb
Mike G, I was indeed a bit loosey-goosey with my specification of the average mid-range shooting percentage that should serve as the guide for equivalently or more valuable three-point shots. Not only did I not include an augment owing to mid-range shot attempts that are fouled. I also didn't discuss the comparative risk of turnovers - what is also a relevant criterion.
These omissions were (admittedly a bit lazy but) primarily for expositional clarity and reflected, in the first instance, my recollection that mid-range shots are simply not fouled that much to matter. (Makes intuitive sense, no? Why foul someone taking a terrible shot, except "accidentally"?) And in the second instance, the recollection was that turnovers were more likely inside the three-point line than out, so omitting mention of this factor was implicitly yielding a conservative estimate.
What is certainly the case is that your suggestion that league-wide TS% should inform the threshold is incorrect. The issue/choice at hand is notionally transferring shot attempts from the mid-range to behind the arc and NOT those from close-in, where the vast majority of fouled shot attempts take place.
But as my faith in my memory decreases by the day, I thought it a good idea to revisit the document informing my views, an old game-charting article in 82games: http://www.82games.com/locations.htm. Now, there are new and improved data on this very question, but I will be surprised if these substantively modify the picture painted therein.
In that article, the relevant mid-range comparisons are the zones listed as "6 to 10" within which it is noted that 7% of all shooting fouls take place. Now, within that 7% (as I understand the text) are also included shooting fouls occurring within the bonus that are not on field goal attempts. Accordingly, the 7% should be interpreted as an upper-bound augment to the returns on mid-range shot attempts.
So, to cut to the chase, when I take that assumed 7% and apply that to the 2014-15 data for shots between 10 feet out and the arc, what I get as an estimate of the relevant TS% is 0.409. And given that this is an upper-bound (both in terms of the applied free throw premium as well as the increased danger of turnover, relative to three-point shots) I feel quite comfortable with the numbers I provided: 0.400 for the mid-range, with a 0.267 three-point equivalent.
Does this explanation make sense, and if not, with what do you disagree?
These omissions were (admittedly a bit lazy but) primarily for expositional clarity and reflected, in the first instance, my recollection that mid-range shots are simply not fouled that much to matter. (Makes intuitive sense, no? Why foul someone taking a terrible shot, except "accidentally"?) And in the second instance, the recollection was that turnovers were more likely inside the three-point line than out, so omitting mention of this factor was implicitly yielding a conservative estimate.
What is certainly the case is that your suggestion that league-wide TS% should inform the threshold is incorrect. The issue/choice at hand is notionally transferring shot attempts from the mid-range to behind the arc and NOT those from close-in, where the vast majority of fouled shot attempts take place.
But as my faith in my memory decreases by the day, I thought it a good idea to revisit the document informing my views, an old game-charting article in 82games: http://www.82games.com/locations.htm. Now, there are new and improved data on this very question, but I will be surprised if these substantively modify the picture painted therein.
In that article, the relevant mid-range comparisons are the zones listed as "6 to 10" within which it is noted that 7% of all shooting fouls take place. Now, within that 7% (as I understand the text) are also included shooting fouls occurring within the bonus that are not on field goal attempts. Accordingly, the 7% should be interpreted as an upper-bound augment to the returns on mid-range shot attempts.
So, to cut to the chase, when I take that assumed 7% and apply that to the 2014-15 data for shots between 10 feet out and the arc, what I get as an estimate of the relevant TS% is 0.409. And given that this is an upper-bound (both in terms of the applied free throw premium as well as the increased danger of turnover, relative to three-point shots) I feel quite comfortable with the numbers I provided: 0.400 for the mid-range, with a 0.267 three-point equivalent.
Does this explanation make sense, and if not, with what do you disagree?
Re: How the NBA bombed the bomb
Well, a 40% eFG shot is either a good shot, or it is not. If your objective is to park beyond the arc and shoot, then you're subject to that limitation. If you're driving or otherwise working to get closer to the rim, then you have to include your TS% from everywhere inside the arc. Almost nobody insists on taking a 15' shot if he has a chance to get closer, for a layup, dunk, or and-1.
With the clock about to expire, of course a 40% shot is not bad. If you can rebound 1/4 of that other 60%, then you have a 55% chance of either scoring or having the ball again.
Turnovers are a good consideration. Drawing a foul on the opponent is also worthwhile, and yet not well described by the box score.
Are teams getting better at defending the 3? League 3fg% has dropped from .367 to .350 in the last couple of years. Any predetermination of the offense invites predatory action in the passing lanes. I don't know the tradeoff.
With the clock about to expire, of course a 40% shot is not bad. If you can rebound 1/4 of that other 60%, then you have a 55% chance of either scoring or having the ball again.
Turnovers are a good consideration. Drawing a foul on the opponent is also worthwhile, and yet not well described by the box score.
Are teams getting better at defending the 3? League 3fg% has dropped from .367 to .350 in the last couple of years. Any predetermination of the offense invites predatory action in the passing lanes. I don't know the tradeoff.