I haven’t run into any analysis of how volatility ought to affect 4th quarter offense as it pertains to taking 2s vs. 3s. If there is a good piece on this out there somewhere, no need to reinvent the wheel. But I would love to know thoughts here (especially from the advanced-stat gurus) about how time left on the clock should be factored into offensive decisionmaking.
It seems obvious to me that, when you’re quite a ways behind, volatility is your friend. And I assume there must be more volatility in 3-point attempts than 2s. So duh, down by 10 or more, aim at the bulls eye.
Less obviously, it seems to me that in close games, it often matters more that you do score than it does how many points you get. So just like you ought to gradually move your retirement account into bonds and away from more volatile stocks as you near your personal end, doesn’t it make sense to think an offense ought to gravitate towards the higher percentage shots—and therefore, presumably, reduce exposure to volatility—as we move into and through the 4th quarter?
Is this right? And if so, what are the implications?
I had to read this several times to make sure I understood what you say. I do...understand, and completely agree with it. Every time you don't score it is a possible 4 to 6 point swing for the other team. Why give them a gift? Score...and don't let them score. What's not to understand?
This also brings up a sore point for me. I don't believe in playing a slow down game (on offense) when you are ahead. Keep playing what put you ahead. Don't mess with a finely tuned offense. Where you take time off the clock is on defense. I am, or was, a defense oriented player. I always believed that when you get ahead--stomp the hell out of the other team! That will put a little trepidation in the other team...and the battle is almost won before it starts. I know...black and white with many variations but I hope you get the point.
Good post, by the way.
Thanks, balls, for replying to my lonely, neglected thread. Maybe I didn’t explain the idea very well. I’m flattered that you read it several times.
And thanks to Luka and Scott41 for liking so I didn’t feel like a complete idiot.
It just seems obvious to me, when you have a 2 point lead and the ball near the very end of a game, a 40-50% chance of adding 2 points is better than a 30-40% chance of adding 3. Why? Because either score increases the difficulty of catching up by a lot. And the most important thing is to avoid giving the opposition the chance to catch up or take the lead.
Now, that example is extremely narrow. And it is somewhat unique in that either score (a 2 or a 3) changes a one-possession game into a two-possession game. But the principles involved should apply more broadly. As you approach the end of a game with a lead, scoring some points (for instance, 2) rather than none becomes increasingly valuable. In the same way, scoring more points (for instance, 3) rather than just some points becomes less important.
The opposite is true early on. Whatever is most likely to increase your point total most efficiently is definitely best in the first half. And probably for awhile thereafter. But nearing the end, the point margin matters less and less. And maintaining the lead matters more and more.
Because 3-point shots are a lower-percentage shot than 2s generally speaking, the likelihood of a miss (and therefore, the likelihood of a scoreless possession) is necessarily higher when you shoot them. I believe the correct statistical term for this is concept is “volatility,” although I found a reddit thread and an article that seemed to be calling a related notion “variability.”
Volatility makes me think of the stock market and the typical advice from investment advisors that you want to decrease your exposure to volatility as you near retirement. Why is that? Because the risk of a big loss is significantly more consequential late in life, and preserving your assets, rather than necessarily growing them, becomes a more and more important goal.
This seems right on point within a basketball game. Score as much as you can early. But late, as the end approaches, preserving your lead (if you have one) becomes the more important goal.
Of course, this isn’t true in basketball or investing when you’re behind. If you don’t have enough money to retire or enough points to win, the roulette wheel becomes more and more appealing. But this is because you have less to lose when you’re already facing a likely loss. The more desperate the circumstances, the more that desperate measures are called for, right?
I also get that not all 3s and 2s are created equal. I’m not arguing that JJ Reddick should pass up open shots. What I am saying is that the calculus changes when you are protecting a lead. And I think that change (a) is due to volatility and (b) means 2-point attempts become more valuable and 3-point attempts less so.
And most importantly, I’m asking if anyone knows of any literature on this concept or has any insight. All that I can find deals with whole game/whole season analytics. None of it seems to address this idea that what is more valuable changes the closer you get to the finish.
(05-18-2021, 04:09 PM)Jommybone Wrote: [ -> ]I haven’t run into any analysis of how volatility ought to affect 4th quarter offense as it pertains to taking 2s vs. 3s. If there is a good piece on this out there somewhere, no need to reinvent the wheel. But I would love to know thoughts here (especially from the advanced-stat gurus) about how time left on the clock should be factored into offensive decision making.
It seems obvious to me that, when you’re quite a ways behind, volatility is your friend. And I assume there must be more volatility in 3-point attempts than 2s. So duh, down by 10 or more, aim at the bulls eye.
Less obviously, it seems to me that in close games, it often matters more that you do score than it does how many points you get. So just like you ought to gradually move your retirement account into bonds and away from more volatile stocks as you near your personal end, doesn’t it make sense to think an offense ought to gravitate towards the higher percentage shots—and therefore, presumably, reduce exposure to volatility—as we move into and through the 4th quarter?
Is this right? And if so, what are the implications?
Excellent points well expressed!
Your reasoning has a prima facie strength to it, with or without advanced stats. From the point of view of the team coming back, opponents that are behind by big numbers have to both get stops on defense and score themselves
I've made the point repeatedly as it related to the Carlisle Mavs who at times were known to blow huge leads in very short time periods as their high powered offense last season could let a team back in as fast as it could put them in a hole.
In the
Boban Watch thread we discussed some of the worst cases as Marjanovic can be very efficient even over periods from an offensive perspective, if used consistently which is rarely the case.
This season the Mavs by my eye test appear to be doing a much better job overall of holding their share of leads or at least not having as many
late game epic fails thanks in large part to Luka's elite ability to score inside or generate high percentage looks for his teammates and his general ability to perform in the clutch. Luka doing much better from the 3 makes his threat inside that much harder to reliably defend.
(05-19-2021, 02:18 PM)Jommybone Wrote: [ -> ]Thanks, balls, for replying to my lonely, neglected thread. Maybe I didn’t explain the idea very well. I’m flattered that you read it several times.
And thanks to Luka and Scott41 for liking so I didn’t feel like a complete idiot.
It just seems obvious to me, when you have a 2 point lead and the ball near the very end of a game, a 40-50% chance of adding 2 points is better than a 30-40% chance of adding 3. Why? Because either score increases the difficulty of catching up by a lot. And the most important thing is to avoid giving the opposition the chance to catch up or take the lead.
Now, that example is extremely narrow. And it is somewhat unique in that either score (a 2 or a 3) changes a one-possession game into a two-possession game. But the principles involved should apply more broadly. As you approach the end of a game with a lead, scoring some points (for instance, 2) rather than none becomes increasingly valuable. In the same way, scoring more points (for instance, 3) rather than just some points becomes less important.
The opposite is true early on. Whatever is most likely to increase your point total most efficiently is definitely best in the first half. And probably for awhile thereafter. But nearing the end, the point margin matters less and less. And maintaining the lead matters more and more.
Because 3-point shots are a lower-percentage shot than 2s generally speaking, the likelihood of a miss (and therefore, the likelihood of a scoreless possession) is necessarily higher when you shoot them. I believe the correct statistical term for this is concept is “volatility,” although I found a reddit thread and an article that seemed to be calling a related notion “variability.”
Volatility makes me think of the stock market and the typical advice from investment advisors that you want to decrease your exposure to volatility as you near retirement. Why is that? Because the risk of a big loss is significantly more consequential late in life, and preserving your assets, rather than necessarily growing them, becomes a more and more important goal.
This seems right on point within a basketball game. Score as much as you can early. But late, as the end approaches, preserving your lead (if you have one) becomes the more important goal.
Of course, this isn’t true in basketball or investing when you’re behind. If you don’t have enough money to retire or enough points to win, the roulette wheel becomes more and more appealing. But this is because you have less to lose when you’re already facing a likely loss. The more desperate the circumstances, the more that desperate measures are called for, right?
I also get that not all 3s and 2s are created equal. I’m not arguing that JJ Reddick should pass up open shots. What I am saying is that the calculus changes when you are protecting a lead. And I think that change (a) is due to volatility and (b) means 2-point attempts become more valuable and 3-point attempts less so.
And most importantly, I’m asking if anyone knows of any literature on this concept or has any insight. All that I can find deals with whole game/whole season analytics. None of it seems to address this idea that what is more valuable changes the closer you get to the finish.
Sorry about your thread being neglected, jommy. Sounded like you wanted a reference to an article, and I'm guessing no one knew of one.
Sounds like your issue is shot selection in the clutch. Might be a good subject for a Sloane Conference presentation, but I haven't seen anything specifically on that to date.
Might be because the answer is so situation-dependent.
When your team has a lead, agree that it can be more important to score than to specifically hit threes. However, in the last minutes of a close game, the defenses have usually clamped down, and the offenses often have to take whatever shot the defense allows, be it a two or a three.
I should say that my view on this was somewhat influenced by the Monta Ellis days, when the team blew a bunch of leads in the last few minutes even though they were so far ahead that if they had just run out the shot clock on each possession, they were far enough ahead to have won. He played the same fast pace no matter what. This resulted in more turnovers, giving up offensive rebounds, and more long shots with long rebounds enabling the other team to run on them. Just gifted the opponent the opportunity to win. Sometimes, discretion is the better part of valor.
Of course, when your team is behind, they have to take more risks, and threes might be the only way they can catch up, especially if the leading team is still scoring.
I can't claim any particular insight on the issue, but the question is interesting.
Mavsluvr, you are certainly right about the D clamping down near the end. Lots of reasons that happens. And it makes clutch scoring an enormously important thing—one that is significantly different from scoring in general.
But if I’m right about the volatility issue, that knowledge (about 2s being more important than 3s late in games) is much more useful for deciding who should be on the floor than it is for deciding whether a given open look should be taken. Which makes me think Dwight over Maxi and Jalen over, well, anyone not named Luka.
Appreciate your thoughts. Sorry to be whining over the lack of responses.
Dahlism, gracias. Yes, blowing big leads looks like volatility at its most obvious. Or as the TNT crew likes to say, live by the 3, die by the 3.
Dare I go one step further and call 2-point shots “conservative” and 3 pointers, um, “liberal”? (I know, I know. Bad idea on a discussion board.)
(05-19-2021, 05:09 PM)Jommybone Wrote: [ -> ]Mavsluvr, you are certainly right about the D clamping down near the end. Lots of reasons that happens. And it makes clutch scoring an enormously important thing—one that is significantly different from scoring in general.
But if I’m right about the volatility issue, that knowledge (about 2s being more important than 3s late in games) is much more useful for deciding who should be on the floor than it is for deciding whether a given open look should be taken. Which makes me think Dwight over Maxi and Jalen over, well, anyone not named Luka.
Ok, I see where you are coming from. I think dahlsim has often expressed the view that the emphasis when the team is leading in the last few minutes should be more toward scoring than defending. I think his position has a lot of merit. However, I think coaches more tend to think in terms of a defensive lineup in that situation, because that's just the track their minds run on.
But to your point, I think that is why they had Luka develop a mid-range shot in the offseason. Because that may be all the defense gives you in tight situations. On the offensive end, coaches might want to give the nod to guys who can get to the basket and score, or who can hit from mid-range. Moreso than guys who pretty much just hit threes, at least if they're not indispensable for defensive purposes.
Appreciate your thoughts. Sorry to be whining over the lack of responses.
No worries. This is an interesting issue that offers plenty of food for thought.
(05-19-2021, 05:16 PM)Jommybone Wrote: [ -> ]Dahlism, gracias. Yes, blowing big leads looks like volatility at its most obvious. Or as the TNT crew likes to say, live by the 3, die by the 3.
Dare I go one step further and call 2-point shots “conservative” and 3 pointers, um, “liberal”? (I know, I know. Bad idea on a discussion board.)
De nada, mi amigo.
To add, blowing big leads with time running out is only one of the most obvious symptoms of what you well termed as a
volatility issue a team may have. In general the top teams, the real contenders have less bust and boom throughout their games than the rest of the field.
One of the common elements I've seen overall is that despite the 'small ball' era where guards and wings generally became more valuable than Bigs at C and PF, the top teams
still generally have superior play inside from their big men. Last year the Lakers had not only AD for example but had former Allstar Dwight Howard and former champion and ex-Mav Javale McGee among others playing key roles in putting their roster over the top. This year they made sure to get even stronger inside with another veteran NBA champ in Gasol, former 6th man Montrez etc. then going out and adding yet another All star quality big in Drummond!
The
Toronto Raptors before that had Gasol and Ibaka next to Siakam, GSW always had quality big guys like David West, Kevon ex-mavs McGee and Pachulia surrounding stars Draymond and Durant.
This year at 2 centers are in the top 5 MVP running for front leading teams, etc. etc.
Superior play from a teams
front line bigs reduces volatility not only with inside offensive options but with tough rebounding, offensive boards with high percentage put backs and 2nd chance offensive opportunities, free throw line advantages and the obvious bigger bodies at the rim on defense.
(05-19-2021, 06:32 PM)Dahlsim Wrote: [ -> ]One of the common elements I've seen overall is that despite the 'small ball' era where guards and wings generally became more valuable than Bigs at C and PF, the top teams still generally have superior play inside from their big men. Last year the Lakers had not only AD for example but had former Allstar Dwight Howard and former champion and ex-Mav Javale McGee among others playing key roles in putting their roster over the top. This year they made sure to get even stronger inside with another veteran NBA champ in Gasol, former 6th man Montrez etc. then going out and adding yet another All star quality big in Drummond!
The Toronto Raptors before that had Gasol and Ibaka next to Siakam, GSW always had quality big guys like David West, Kevon ex-mavs McGee and Pachulia surrounding stars Draymond and Durant.
This year at 2 centers are in the top 5 MVP running for front leading teams, etc. etc.
Probably not the best examples. Both Lakers and Warriors went small when it mattered most. Reduced the minutes of their centers in the playoffs. And it´s not like they used any capspace for them. Basically all vet min signings or MLE guys.
(05-19-2021, 04:24 PM)mavsluvr Wrote: [ -> ]Might be because the answer is so situation-dependent. ...
I can't claim any particular insight on the issue, but the question is interesting.
Very true as there is a number of ways to look at the issue as raised.
Volatility as it relates to the 2 point shot is the thread title. The 2 pointer is compared to the 3 pointer which I think is relevant to closing games yes but is also relevant to team consistency vs volatility throughout the game. Many times an unanswered scoring outburst for example can practically all but win a game even before the clutch time minutes arrive.
So a lot of stats can be relevant there since its really tied to the
huge increase in the role of the 3 pointer across the league.
A point I find particularly interesting for example is that we often assume that NBA coaches have it all figured out so if we see something odd we figure they're the pro's and if they aren't doing what looks obvious it much be for good reasons we just don't see.
That is often true no doubt but its also true that coaches follow trends and mindsets as you pointed out. It can take years for them to adjust as a whole and that is what happened in fact with the 3 point shot.
Quote:When the NBA slapped a 3-point line on the court and declared certain shots worth more than other shots—the single most consequential rule change in modern sports—it would take years for basketball teams to realize they had different incentives. Slowly at first, and then with astonishing speed, they shot more and more 3-pointers. But for almost four decades, there was a limit. No team attempted 40% of its shots as threes.
Old habits are hard to break and it turns out to be true for NBA coaches as well.
Eventually though, the strategies for using the 3 point shot came around.
Quote:Then it happened in 2017. [color=var(--color-blue)]It happened again in 2018—but with two teams above the 40% threshold. That number would balloon to five in 2019 and nine in 2020. [/color]
But what’s happening this season is a rapid acceleration of a trend that has reshaped the game. There are currently 15 teams with 3-point rates above 40%. Half the league is now doing something that until recently no team in the history of the league had ever done.
So who is to say that right now teams aren't still figuring out how and when to balance the best use of 2 point offense in order to get an advantage over the 3 point dominant offenses they all face?
(05-19-2021, 06:54 PM)dirkfansince1998 Wrote: [ -> ]Probably not the best examples. Both Lakers and Warriors went small when it mattered most. Reduced the minutes of their centers in the playoffs. And it´s not like they used any capspace for them. Basically all vet min signings or MLE guys.
Mostly true (it also
mattered a few times when they used their bigs) but it still serves my example. If all the teams are doing the same thing then where is the advantage for one team over the other?
The point is, the Lakers were able to play big when they needed too, even if it was only for certain matchups and certain stretches of certain games. The Lakers
had a number of ways to get consistent scoring when another had found an opening, partly because they could use size advantages when needed at several positions including among the Bigs. It became a difference that could put them over the top and turn close games or possible L's into W's.
(BTW, Watching the Spurs get destroyed right now by Memphis and the commentators point out the Memphis front line led by Valanciunas is something the Spurs don't seem to have an answer for).
(05-19-2021, 05:09 PM)Jommybone Wrote: [ -> ]Mavsluvr, you are certainly right about the D clamping down near the end. Lots of reasons that happens. And it makes clutch scoring an enormously important thing—one that is significantly different from scoring in general.
But if I’m right about the volatility issue, that knowledge (about 2s being more important than 3s late in games) is much more useful for deciding who should be on the floor than it is for deciding whether a given open look should be taken. Which makes me think Dwight over Maxi and Jalen over, well, anyone not named Luka.
Appreciate your thoughts. Sorry to be whining over the lack of responses.
I just hope your agenda isn't a push for us to get Demarr, when 1) he's going to be costly, and 2) scoring inside the arc is his sole skill set, so he binarily doesn't deserve such a contract.
I guess Valanciunas is making @"Dahlsim" case right now. We all know about his short comings on defense but he is forcing the Spurs to go big. Basically the reverse RC. Bigmen offense dictates the matchups. Not the ability to defend.
I love this topic. The logic seems like it should be so clear. And it also seems like you should be able to find the data to sort out some tests.
It seems like a no brainer that if you are up by 2 possessions with a minute left, you put in Boban for a 2 pointer, and you have 60% possibility of being up by 3 possessions with 50 seconds left. It leaves a 40% chance of being up by 2 possessions with 50 seconds left. If the planned volatility by the other team works in their favor, they hit a 3 pointer within 10 seconds and your original lead has decreased - either by 1 or by 3 points. I think the Mavs stall-the-ball crunch time offense (that I do hate) follows a different logic. Up by 2 possessions with a minute left, dribble in place for 20 seconds before jacking up whatever shot you can get in the remaining shot clock. We often see a contested 3 or a turnover. Worst case is you are still up by 2 possessions with 40 seconds left. So the question is how valuable are those 10 seconds? You might say the Mavs opt for 0 volatility for those 10 seconds. In terms of investing, it's like telling your 75 year old uncle to not play the ponies with his ss check this weekend.
Other arguments that I can think of for going for the 3 pointer that dont have to do with clock management include staying within the flow of the offense you've built or that you might have players on the court who are perimeter players (theirs because they need the 3 pointers and yours because you'll be defending the 3 point line)
Now this thread has taken off! Nice commentary. All that’s left is for a math guy to reduce it to a formula.
Scott41, no conscious agenda here. But I’m a homer. So there might be some unconscious pro-current roster underpinnings.
Let me test one more piece of logic here. I believe it’s obvious that, by its very nature, volatility should be helpful, generally, when you’re the underdog and harmful, generally, when you’re the fav. See where I’m going?
Much has been said this week about the Mavs performing better than expected vs. winning teams and not so hot compared to expectations vs. losing teams. So maybe we’re on to something here?
But my original thesis isn’t that we should aim for less volatility. It’s that the relative volatility inherent in taking 3s vs. 2s* likely means an offense should try to shoot fewer 3s and more 2s in late-game situations when they have the lead—and maybe even when they don’t, if it’s close. (My guess would be the last half quarter or so.)
* The same logic probably holds for free throws, as even a 70% shooter has a 91% chance of making 1 of 2 freebies. (Math guys, feel free to correct me.)
(05-19-2021, 08:37 PM)Jommybone Wrote: [ -> ]* The same logic probably holds for free throws, as even a 70% shooter has a 91% chance of making 1 of 2 freebies. (Math guys, feel free to correct me.)
Yes, a 70% free throw shooter has a 9% [30% X 30%] chance of missing two out of two free throws, a 42% chance of making exactly one free throw [30% X 70% for miss-the-first/make-the-second, and 70% X 30% for make-the-first/miss-the-second], a 49% [70% X 70%] chance of making two out of two free throws, and a 91% [42% + 49%] chance of making
at least one of two free throws.
So ...
Please explain what you mean by "the same logic probably holds for free throws."
(05-19-2021, 10:19 PM)Hogmelon Wrote: [ -> ]Yes, a 70% free throw shooter has a 9% [30% X 30%] chance of missing two out of two free throws, a 42% chance of making exactly one free throw [30% X 70% for miss-the-first/make-the-second, and 70% X 30% for make-the-first/miss-the-second], a 49% [70% X 70%] chance of making two out of two free throws, and a 91% [42% + 49%] chance of making at least one of two free throws.
So ...
Please explain what you mean by "the same logic probably holds for free throws."
In the OP, I wrote,
“Less obviously, it seems to me that in close games, it often matters more that you do score than it does how many points you get. So just like you ought to gradually move your retirement account into bonds and away from more volatile stocks as you near your personal end, doesn’t it make sense to think an offense ought to gravitate towards the higher percentage shots—and therefore, presumably, reduce exposure to volatility—as we move into and through the 4th quarter?”
What I mean by “same logic . . . free throws” is that, as the end of the game nears, FTs likely become even more valuable relative to 3-point attempts.
Let’s compare league average from 3 (36.7%), from 2 (53.0%) and from the free throw line (77.8%). On average, a 3PA has an expected value of 1.10 points (3 x .367), a 2PA 1.06, and a pair of FTs 1.56.
We could reasonably express the relative values of these shot attempts by saying, on average, the 3-point attempts are worth 104% of the value—and free throw pairs 147% of the value—of the 2-point attempts.
As you know, this is exactly why teams are taking fewer 2-point attempts these days and hoisting more and more 3s. And it’s why free throws (and the ability to get them) have always been a key factor in the outcome of games.
But let’s imagine that a hypothetical average 3 or average 2 or average pair of FTs is available on 4 consecutive possessions in a hypothetical game. Expected values for the shots stay the same, but because there are 4 possessions, we multiply by 4:
Expected value of 4 average 3PAs = 4.4
Expected value of 4 average 2PAs = 4.24
Expected value of 4 average pairs of FTs = 6.24
Now finally, let’s talk volatility. Here are the chances of going
scoreless on those 4 consecutive possessions:
3PAs: 16.1% (63.3% miss rate x
63.3% x 63.3% x 63.3%)
2PAs: 4.9% (47% miss rate x . . .)
FTs: .00017% (22.2% miss rate x . . .)
Let’s state the same thing in slightly different terms. A team that shoots 3s on 4 consecutive possessions in the NBA will go scoreless on those 4 possessions about 1 out of 6 times. Shooting 2s instead will result in going scoreless on 4 consecutive possessions about 1 out of 20 times. Shooting FTs instead will result in going scoreless on 4 consecutive possessions about 1 out of 6,000 times.
This math seems to support what I was trying to say.