KenPom: "The Boeheim exception"

[690West]

very interesting reading. We might have a pretty good coach here.
http://kenpom.com/blog/index.php/weblog/entry/the_boeheim_exception

 

[AirForceOrange]

very interesting reading. We might have a pretty good coach here.
http://kenpom.com/blog/index.php/weblog/entry/the_boeheim_exception

Nah...he sucks...talk to me when he hits 1K wins...

 

[pfister1]

I don't think it is really endorsing JB's approach.

 

[AlaskaSU]

Ken explained it as clog the lane; force the oppositionto take more 3's, some of which are forced. Ken admits that it makes a small difference but he ignored the effect on offense. By keeping the guards out on top Syracuse gets a couple of extra high percentage runouts each game.

Suppose a team takes two extra bad three point shots and that they normally shoot 33% from three. They should have earned 2 points from those two extra threes, but since they were bad shots they earn 0 points. That gives us a two point advantage. Furthermore, assume that keeping the guards on top gives us two extra runouts, worth 4 points a game. In sum, Boeheim coaching could give us a 6 point advantage. Suppose that this estimate is 50% high: even a 3 point advantage is significant over the long run. Add a couple of points for home court advantage when we play at home and this helps explain why we have the Nations longest home court winning streak. If this is tavern league, I'll take it. Fact is that JB is perhaps the most intelligent coach (with the exception of Igor and Blue Curtain).

 

[CusefanATL]

I’m willing to concede the Boeheim zone has some influence, but it’s small.

 

[Zelda Zonk]

Ken explained it as clog the lane; force the oppositionto take more 3's, some of which are forced. Ken admits that it makes a small difference but he ignored the effect on offense. By keeping the guards out on top Syracuse gets a couple of extra high percentage runouts each game.

Suppose a team takes two extra bad three point shots and that they normally shoot 33% from three. They should have earned 2 points from those two extra threes, but since they were bad shots they earn 0 points. That gives us a two point advantage. Furthermore, assume that keeping the guards on top gives us two extra runouts, worth 4 points a game. In sum, Boeheim coaching could give us a 6 point advantage. Suppose that this estimate is 50% high: even a 3 point advantage is significant over the long run. Add a couple of points for home court advantage when we play at home and this helps explain why we have the Nations longest home court winning streak. If this is tavern league, I'll take it. Fact is that JB is perhaps the most intelligent coach (with the exception of Igor and Blue Curtain).

This is why stats suck at telling the truth.

1. We don't actually 'keep the guards on top.' They move around with the ball. They frequently are in the paint getting rebounds, or out of position when doing the fly-by thing when they're recovering toward a shooter. Every team has a 'rule' that a guard needs to be able to be the first man back on defense. Every team's guards occasionally get runouts.
2. The rebounding disadvantages outweigh the potential runouts. Not having a specific man to box out means we're at a disadvantage. And, when you start to counter this argument by saying we're outrebounding our opponents, it doesn't also mean that we're not still conceding MORE rebounds by playing zone. Even if we could outrebound a team by 2:1, you can't assume the zone isn't responsible for losing a few more rebounds, and offensive rebounds are just as/more dangerous than runouts.
3. The game isn't played like baseball. It's continuously dynamic. If you actually do get a 6 point advantage in one area, it's dissipated during the game as the other team does things to compensate, either through personnel, strategy, or gameplay. That's why you can't add each team's players' scoring averages and determine who's going to win the game.

 

[Zelda Zonk]

Here's, hopefully, a simple question, re: Pom's assertions:

"

If you look at Syracuse’s 17 NCAA tournament games since 2004, opponents have made 32.8% of their threes, about 2% below the national average over that time. I don’t have any fancy math to prove this, but I would guess that’s about the influence of the Boeheim zone on opponents shooting: somewhere around 2 to 3%. A 40% team would be expected to shoot 37% against Syracuse or Eastern Michigan."​

What part does a talent differential play in this equation? If there were a way to empirically quantify the talent level on the floor for each team, per game — minus coaching — wouldn't we almost invariably have the advantage? If we were to take our schedule, and put all of those teams on a neutral floor, wouldn't we be favored in just about every game? If that is true, why would it surprise anyone that we would be 3%+ better at defending any part of the court than the national average?

At the end of the day, it doesn't really matter why you have a positive defensive metric. You use it and move on. But, it still bothers me when this 3% is trumpeted as some great achievement or a validation for the zone. It means something. It doesn't mean everything. And, just as there are benefits, there are detriments. The zone routinely gives big outside scoring nights to nobodies. A 2-star guard can have a 25-point catch-and-shoot day against us, where if he were playing against a man defense, he'd have to do more than just stand there and wait for an opportunity to chuck. We just made Darrun Hilliard look like All World. Same thing happens against even our early season cupcakes. Some schmo from South by Southeast Bumblefrick College will go off for 20-something, and then never be heard from again. We can't actually say we're conceding this, because it weakens the other players, right? The math/stats don't back that up.

I still don't get a sense of what Pom's conclusion is. And along with him, none of the other media heads/ex-players/ex-coaches who so often ramble on about the zone — none of them actually go so far as to endorse the zone. Why the disconnect?

 

[TinyManInside]

This is why stats suck at telling the truth.

1. We don't actually 'keep the guards on top.' They move around with the ball. They frequently are in the paint getting rebounds, or out of position when doing the fly-by thing when they're recovering toward a shooter. Every team has a 'rule' that a guard needs to be able to be the first man back on defense. Every team's guards occasionally get runouts.
2. The rebounding disadvantages outweigh the potential runouts. Not having a specific man to box out means we're at a disadvantage. And, when you start to counter this argument by saying we're outrebounding our opponents, it doesn't also mean that we're not still conceding MORE rebounds by playing zone. Even if we could outrebound a team by 2:1, you can't assume the zone isn't responsible for losing a few more rebounds, and offensive rebounds are just as/more dangerous than runouts.
3. The game isn't played like baseball. It's continuously dynamic. If you actually do get a 6 point advantage in one area, it's dissipated during the game as the other team does things to compensate, either through personnel, strategy, or gameplay. That's why you can't add each team's players' scoring averages and determine who's going to win the game.

Great post.

 

[IthacaMatt]

I’m willing to concede the Boeheim zone has some influence, but it’s small.

KenPom said it was about a 3% reduction in the opponents' 3 point shooting percentage when they face our zone. So if a team ordinarily shoots 37% (about the national average), and they attempt 16 threes a game, if they play us, they can be expected to score about a half a 3 per game less (slightly under 5.5 makes, versus nearly 6 makes), or a 1.5 point difference in our favor.

Not a lot, but a bit of an edge. And of course, that's only 1 aspect of the zone.

KenPom's article doesn't take into account the number of turnover numbers forced by the zone versus the national average, nor the zone's ability to launch us into the fast break off a turnover.

Yet another factor is that the zone takes away much of the offense of opposing teams, making them game plan for us and ditch 75% of their offensive playbook.

When you figure in how much we generate more turnovers, and convert more turnovers into fast break points than other teams, plus our ability to force the opponent out of its comfort zone and ordinary offense, I would guess that the zone is probably worth a 5 point difference against a random opponent. Thoughts?

 

[BuffOrange]

Ken explained it as clog the lane; force the oppositionto take more 3's, some of which are forced. Ken admits that it makes a small difference but he ignored the effect on offense. By keeping the guards out on top Syracuse gets a couple of extra high percentage runouts each game.

Suppose a team takes two extra bad three point shots and that they normally shoot 33% from three. They should have earned 2 points from those two extra threes, but since they were bad shots they earn 0 points. That gives us a two point advantage. Furthermore, assume that keeping the guards on top gives us two extra runouts, worth 4 points a game. In sum, Boeheim coaching could give us a 6 point advantage. Suppose that this estimate is 50% high: even a 3 point advantage is significant over the long run. Add a couple of points for home court advantage when we play at home and this helps explain why we have the Nations longest home court winning streak. If this is tavern league, I'll take it. Fact is that JB is perhaps the most intelligent coach (with the exception of Igor and Blue Curtain).

50% is a big understatement. The difference between 33% and 30% on two three's isn't 2pts and 0pts, it's 2pts and 1.8pts on average. And as pointed out the runouts we get are mitigated by the boards we sacrifice.

 

[Knicks411]

So if a team ordinarily shoots 37% (about the national average), and they attempt 16 threes a game, if they play us, they can be expected to score about a half a 3 per game less, or a 1.5 point difference in our favor. Not a lot, but a bit of an edge.

I do wonder how much of that edge can be wiped away by the fact they when they do miss, they are more likely to get the rebound against our zone

 

[Sgt Cuse]

Here's, hopefully, a simple question, re: Pom's assertions:

"

If you look at Syracuse’s 17 NCAA tournament games since 2004, opponents have made 32.8% of their threes, about 2% below the national average over that time. I don’t have any fancy math to prove this, but I would guess that’s about the influence of the Boeheim zone on opponents shooting: somewhere around 2 to 3%. A 40% team would be expected to shoot 37% against Syracuse or Eastern Michigan."​

What part does a talent differential play in this equation? If there were a way to empirically quantify the talent level on the floor for each team, per game — minus coaching — wouldn't we almost invariably have the advantage? If we were to take our schedule, and put all of those teams on a neutral floor, wouldn't we be favored in just about every game? If that is true, why would it surprise anyone that we would be 3%+ better at defending any part of the court than the national average?

At the end of the day, it doesn't really matter why you have a positive defensive metric. You use it and move on. But, it still bothers me when this 3% is trumpeted as some great achievement or a validation for the zone. It means something. It doesn't mean everything. And, just as there are benefits, there are detriments. The zone routinely gives big outside scoring nights to nobodies. A 2-star guard can have a 25-point catch-and-shoot day against us, where if he were playing against a man defense, he'd have to do more than just stand there and wait for an opportunity to chuck. We just made Darrun Hilliard look like All World. Same thing happens against even our early season cupcakes. Some schmo from South by Southeast Bumblefrick College will go off for 20-something, and then never be heard from again. We can't actually say we're conceding this, because it weakens the other players, right? The math/stats don't back that up.

I still don't get a sense of what Pom's conclusion is. And along with him, none of the other media heads/ex-players/ex-coaches who so often ramble on about the zone — none of them actually go so far as to endorse the zone. Why the disconnect?

Gee, I almost don't know who to believe now, Pomeroy's exhaustive statistical analysis or your anecdote on some player having a good night. If the zone routinely gave up big outside scoring night to nobodies, as you claim, then the rest of the teams must be doing it more, because as Pom is showing, SU is routinely ranked high in the defensive percentage of 3s category. Did you somehow miss that? (the point of the article?)

Pomeroy originally was under the impression that 3 point percentage defense was basically random. He analyzed several years of data and saw that JB/SU routinely outperformed the average. He reluctantly adjusted his theory. I suggest you do the same.

 

[IthacaMatt]

Gee, I almost don't know who to believe now, Pomeroy's exhaustive statistical analysis or your anecdote on some player having a good night. If the zone routinely gave up big outside scoring night to nobodies, as you claim, then the rest of the teams must be doing it more, because as Pom is showing, SU is routinely ranked high in the defensive percentage of 3s category. Did you somehow miss that? (the point of the article?)

Pomeroy originally was under the impression that 3 point percentage defense was basically random. He analyzed several years of data and saw that JB/SU routinely outperformed the average. He reluctantly adjusted his theory. I suggest you do the same.

Historically, we have seen random players go off on the zone for big nights, especially in NCAA tournament games, it seems (or at least, those are the ones that are most memorable). But if you think about it, over the past 4 years, there have been fewer and fewer of those games where some guy goes nuts against us for a career high.

I think the key is that after Jonny Flynn and Paul Harris left town, we groomed a team who accepted the zone and respected what it had to offer. When you have players who buy into the defense, that makes a huge difference. We are so much better at our slides and rotations, it seems, which is probably why our turnovers generated by the zone seem to be higher than ever these past few years.

Are Scoop Jardine, Dion Waiters, Andy Rautins and Michael Carter-Williams such great defenders that they would still be among the national leaders in steals if we played man-to-man, or have they simply learned the tendencies of what opponents do against our defense, and know where the passes are going?

 

[Sgt Cuse]

Historically, we have seen random players go off on the zone for big nights, especially in NCAA tournament games, it seems (or at least, those are the ones that are most memorable). But if you think about it, over the past 4 years, there have been fewer and fewer of those games where some guy goes nuts against us for a career high.

I think the key is that after Jonny Flynn and Paul Harris left town, we groomed a team who accepted the zone and respected what it had to offer. When you have players who buy into the defense, that makes a huge difference. We are so much better at our slides and rotations, it seems, which is probably why our turnovers generated by the zone seem to be higher than ever these past few years.

Are Scoop Jardine, Dion Waiters, Andy Rautins and Michael Carter-Williams such great defenders that they would still be among the national leaders in steals if we played man-to-man, or have they simply learned the tendencies of what opponents do against our defense, and know where the passes are going?

I agree it seems like there have been times that people have gone off. I just think this helps to realize it probably happens everywhere. Duke's equivalent to Igor is probably saying "what's it going to take for K to go zone, Lehigh was killing us from deep". And if JB did pull out and switch it up, how many of those games with the opponent getting off to a hot early start that falters badly in the second half would we have lost out on?

I'm a little surprised by the numbers because as you mention, it really seems like only the last 4 years that the defense has been a point of emphasis. And its only been the last 4 years where JB has felt comfortable enough with his depth to sub in for players he felt weren't playing D the way he wanted.

I hope with our new lack of depth that this team doesn't abandon the defensive focus.

 

[CusefanATL]

KenPom said it was about a 3% reduction in the opponents' 3 point shooting percentage when they face our zone. So if a team ordinarily shoots 37% (about the national average), and they attempt 16 threes a game, if they play us, they can be expected to score about a half a 3 per game less (slightly under 5.5 makes, versus nearly 6 makes), or a 1.5 point difference in our favor.

Not a lot, but a bit of an edge. And of course, that's only 1 aspect of the zone.

KenPom's article doesn't take into account the number of turnover numbers forced by the zone versus the national average, nor the zone's ability to launch us into the fast break off a turnover.

Yet another factor is that the zone takes away much of the offense of opposing teams, making them game plan for us and ditch 75% of their offensive playbook.

When you figure in how much we generate more turnovers, and convert more turnovers into fast break points than other teams, plus our ability to force the opponent out of its comfort zone and ordinary offense, I would guess that the zone is probably worth a 5 point difference against a random opponent. Thoughts?

you may be right... I was just quoting KenPom from the end of his article.

 

[Zelda Zonk]

Gee, I almost don't know who to believe now, Pomeroy's exhaustive statistical analysis or your anecdote on some player having a good night. If the zone routinely gave up big outside scoring night to nobodies, as you claim, then the rest of the teams must be doing it more, because as Pom is showing, SU is routinely ranked high in the defensive percentage of 3s category. Did you somehow miss that? (the point of the article?)

Pomeroy originally was under the impression that 3 point percentage defense was basically random. He analyzed several years of data and saw that JB/SU routinely outperformed the average. He reluctantly adjusted his theory. I suggest you do the same.

You don't have to decide who to believe. I'm not contradicting Pomeroy's "exhaustive statistical analysis." I'm discussing the relevance and significance of those stats. Maybe i should have been clearer. As many people have said before me: you can make statistics say anything you want.

If you haven't noticed "nobodies" 'going off' against us, you may not have been watching SU ball for very long. My "anecdote" was merely a specific mention of the most recent incident. I didn't think i needed to list every scrub from one of our 'warm up games' going back 25 years in order to make the point to those who are familiar with the occurrence.

Personally, i kinda feel 3pt percentage is a moot statistic, on its own. If you force a team to take more 3s, and they shoot a slightly lower percentage than versus the national average, that doesn't mean you're not giving up more points via the 3. I also dispute the relevance of the "national average." We don't play a "national average schedule," for better or worse. Teams have to play differently versus us than they do against other teams. We encourage certain teams to take more threes than other teams, or at least, we should. National average is simply the easiest type of metric to discuss. If you want it to have more meaning, cross-factor it with RPI or some such, but even that's sorta on the moot side. What is more significant to me is FG efficiency.

 

[SWC75]

If you force a team to take more 3s, and they shoot a slightly lower percentage than versus the national average, that doesn't mean you're not giving up more points via the 3.

More points than what?

 

[AlpharettaCuse]

At the end of the day, it doesn't really matter why you have a positive defensive metric. You use it and move on. But, it still bothers me when this 3% is trumpeted as some great achievement or a validation for the zone. It means something. It doesn't mean everything. And, just as there are benefits, there are detriments. The zone routinely gives big outside scoring nights to nobodies. A 2-star guard can have a 25-point catch-and-shoot day against us, where if he were playing against a man defense, he'd have to do more than just stand there and wait for an opportunity to chuck. We just made Darrun Hilliard look like All World. Same thing happens against even our early season cupcakes. Some schmo from South by Southeast Bumblefrick College will go off for 20-something, and then never be heard from again. We can't actually say we're conceding this, because it weakens the other players, right? The math/stats don't back that up.

I still don't get a sense of what Pom's conclusion is. And along with him, none of the other media heads/ex-players/ex-coaches who so often ramble on about the zone — none of them actually go so far as to endorse the zone. Why the disconnect?

Plenty of media people talk about how great our zone is and why. A lot of them are former coaches. They also talk about how Boeheim adjusts it and traps out of it. Bobby Knight talks about how teams pass the ball around looking for a shot, often to find the shot clock running out. The fact that we've been so good defensively is the reason we've won so many games. How good is our defense? What are teams shooting overall against our zone? How many points do we give up per game? The fact some no-name has a career night on it occaisionally means very little. Our defense is very good. As far as I can see we defend the 3 very well. 16th in the country at 28.5%. Overall, we are 3rd in fg defense at 36%. Only Kansas and Texas are better. Most importantly we are 18-2.

 

[Knicks411]

Is there any reason to think we have a greater number of "nobodies" going off against us than any other good team>

 

[Sgt Cuse]

As many people have said before me: you can make statistics say anything you want.

If you haven't noticed "nobodies" 'going off' against us, you may not have been watching SU ball for very long. My "anecdote" was merely a specific mention of the most recent incident. I didn't think i needed to list every scrub from one of our 'warm up games' going back 25 years in order to make the point to those who are familiar with the occurrence.

Personally, i kinda feel 3pt percentage is a moot statistic, on its own. .

You are confusing having a "nobody" going off with an overall percentage. These nobodies are not killing us as regularly as you make it out. Pomeroy did not care who makes the 3, just the overall percentage. Suppose JB/SU did identify shooters and purposefully baited lesser shooters to attempt the 3s. People would then generally be talking about the small percentage of times the poor shooter actually hit (yes, that would be you). But after a large sample size revealed the benefits of this plan wouldn't people appreciate it and support it? - apparently not yet.

Pomeroy was not making the statistics say anything. He started similar to you, believing 3pt % D was random. He noticed a counter to this, thus the "JB exception". It is a complement from an independent statistician.

 

[Knicks411]

Let's also realize that the sample size here is still relatively small on this, so while I would buy this theory, it's not like I'm 100% sold on it.

 

[JLine]

 

[LennX6]

We just made Darrun Hilliard look like All World. Same thing happens against even our early season cupcakes. Some schmo from South by Southeast Bumblefrick College will go off for 20-something, and then never be heard from again.

The all time low was the Vermont nobody who has ,seriously, the game of his life, never had one like if before playing us or after playing us. I couldn't even tell you his name, nor do I really feel like looking his name up. All I know is he went 7-7 and shouldn't have.

 

[AlaskaSU]

A more careful reading of Kens article and my

50% is a big understatement. The difference between 33% and 30% on two three's isn't 2pts and 0pts, it's 2pts and 1.8pts on average. And as pointed out the runouts we get are mitigated by the boards we sacrifice.

A more careful reading of Ken shows why I estimated a 2 point advantage on offense. Ken's rationale is that some of the extra threes are bad threes, ie threesw that would normally not be taken. If they were typical threes there would be no advantage. Forcing teams to take just two bad threes a game does give us an advantage. Two points is a reasonable estimate.

As to offense. Although our guards are not always on top, they are on top far more than the guards of the team we are playing. Except for some sideline traps they are on top.

 

[AlaskaSU]

You don't have to decide who to believe. I'm not contradicting Pomeroy's "exhaustive statistical analysis." I'm discussing the relevance and significance of those stats. Maybe i should have been clearer. As many people have said before me: you can make statistics say anything you want.

If you haven't noticed "nobodies" 'going off' against us, you may not have been watching SU ball for very long. My "anecdote" was merely a specific mention of the most recent incident. I didn't think i needed to list every scrub from one of our 'warm up games' going back 25 years in order to make the point to those who are familiar with the occurrence.

Personally, i kinda feel 3pt percentage is a moot statistic, on its own. If you force a team to take more 3s, and they shoot a slightly lower percentage than versus the national average, that doesn't mean you're not giving up more points via the 3. I also dispute the relevance of the "national average." We don't play a "national average schedule," for better or worse. Teams have to play differently versus us than they do against other teams. We encourage certain teams to take more threes than other teams, or at least, we should. National average is simply the easiest type of metric to discuss. If you want it to have more meaning, cross-factor it with RPI or some such, but even that's sorta on the moot side. What is more significant to me is FG efficiency.

Once again, the point is not how many 3 point shots. The point is whether we force some very low percentage 3's. I think we do. Some teams get frustrated by their inability to drive. On the other hand, some teams get hot from three. Most do not sustain it for 2 halves. There are exceptions, but they are the minority. The zone does give us a small advantage. We have not always had the best talent, but we consistently win regardless. Year to year we are the most consistent team in history.