This crazy defense

[Catledge]

rcklss and tooler made some noteworthy points in the game thread that I thought was worthy of further discussion:

Not sure what to make of this yet but we allow a lot of 3s but teams are shooting a low % against us.

Pelton has said that opponent 3PT% is one of the most random things that define a team. It's completely unreliable in terms of predictive power. So we're running the risk of a regression to the mean.

Is it that we're tempting teams into taking threes that they think are open but turn out to be contested because of great athleticism and hustle on our closeouts?

Is it that we're doing such a good job of scrambling to contest everything that teams are having to settle for contested threes late in the shot clock?

Is it that we're just getting lucky?

I'm not at all sure, but I'm leaning toward the first option, which seems like something that teams will eventually scout and plan for.

[tooler]

I submitted a question to Kevin Pelton for clarification and what to expect from these stats. He's pretty helpful on things like that.

More generally, I find it weird that Skiles focuses so much on FG% defense rather than overall efficiency. There are other ways for teams to score, such as free throws and 3-point attempts. We happen to suck at both.

But so far it's working. We're a top 10 defense.

[CraZyPraiZ]

We allow a lot but I also think we wear teams down at the end. Hence maybe why they are not going in late as much as they are early in the games. We do spend an awful lot of energy on the closeouts but Skiles is doing a good job with the rotation and keeping guys fresh. We don't look as tired late in games.

[MagicFan32]

Here is what i notice..the magic are very good on their rotations and going to the right spot of the floor. when we contest the 3's, we have the defenders to cause shooters problems, it could be random but then again skiles teams as a whole usually defend well

[GameOver25]

rcklss and tooler made some noteworthy points in the game thread that I thought was worthy of further discussion:

Not sure what to make of this yet but we allow a lot of 3s but teams are shooting a low % against us.

Pelton has said that opponent 3PT% is one of the most random things that define a team. It's completely unreliable in terms of predictive power. So we're running the risk of a regression to the mean.

Is it that we're tempting teams into taking threes that they think are open but turn out to be contested because of great athleticism and hustle on our closeouts?

Is it that we're doing such a good job of scrambling to contest everything that teams are having to settle for contested threes late in the shot clock?

Is it that we're just getting lucky?

I'm not at all sure, but I'm leaning toward the first option, which seems like something that teams will eventually scout and plan for.

How would a team even scout or plan for something like that? I don't think that's possible.

[NBlue]

I've heard Amin El Hassan espouse the same theory on percentage being a less reflective measure than attempts. I'm not certain I agree with that though, certainly, Pelton and El Hassan generally know their ****. It would seem to me that 3p attempts would tend to correlate more with 3pt percentage if that were the case http://espn.go.com/nba/statistics/team/_/stat/offense-per-game/sort/threePointFieldGoalPct

For example, the Rockets have shot the most 3s in the league despite having one of the worst shooting percentages. OTOH the Bucks and Spurs are two of the best percentage 3pt shooting teams in the league but average shooting under 20 per game. Presumably, Pelton's theory is that when a player is not open he is less likely to shoot the 3 -- meaning good defense results in less shots and not as much less of a percentage made. However, teams like the Rockets seem to shoot them regardless of the situation and, conversely, teams like the Bucks and Spurs seem to be more selective even though they shoot at a relatively high percentage.

I guess, at the end of the day, I'm not sure that theory pans out. I think if you are looking at a limited sample size of a handful of games its hard to get predictive information due to the variability; however, over the long haul with a larger sample size I still would tend to think that percentage is actually the best indicator.

I would like to hear an explanation of it from Pelton though as I have not heard it fully explained.

[NBlue]

rcklss and tooler made some noteworthy points in the game thread that I thought was worthy of further discussion:

Pelton has said that opponent 3PT% is one of the most random things that define a team. It's completely unreliable in terms of predictive power. So we're running the risk of a regression to the mean.

Is it that we're tempting teams into taking threes that they think are open but turn out to be contested because of great athleticism and hustle on our closeouts?

Is it that we're doing such a good job of scrambling to contest everything that teams are having to settle for contested threes late in the shot clock?

Is it that we're just getting lucky?

I'm not at all sure, but I'm leaning toward the first option, which seems like something that teams will eventually scout and plan for.

How would a team even scout or plan for something like that? I don't think that's possible.

Pretty simple, actually. If you have a defense closing out extra hard on 3s (which I agree we do) then you pump fake and drive off them. You can check out Nicholson against the Celtics for the perfect example.

[tooler]

Actually, I've read the reason some teams have a good percentage but low number of attempts is that it's hard work to generate a 3PT attempt.

I don't think there's much theory behind it. Rather, they've seen opponent 3PT% change pretty significantly over the course of a season or between seasons.

But I'm not sure, and unfortunately Pelton didn't answer me on that one.

[Skin]

Well Utah hit 9 3pointers in the first half to get the lead back. 14-33 on the game for 42.4%. We kinda sucked yesterday.

I'm sure Skiles is not unlike every other NBA coach... wants to defend the perimeter and the post and force teams to take 2pt jumpers.

[tooler]

Actually he did answer me when I asked a second time.

Tim (Florida)

How worried should I be that the Magic have one of the lowest opponent 3-point percentages but allow one of the highest number of attempts? You've said opponent 3PT% is pretty random.

Kevin Pelton (1:01 PM)

Due for a little regression there but other teams are due for more.

Tim (Florida)

Sorry, I'm a little confused when you say the Magic are due for regression in opponent 3PT% but other teams are due for more. What did you mean by that?

Kevin Pelton (1:45 PM)

I have an adjusted defensive rating in my spreadsheet that assumes league-average 3P% and FT% for opponents. The Magic's adjusted defensive rating is 0.6 points per 100 possessions worse than their actual defensive rating, which is a difference but not a particularly exceptional one. Chicago, for example, has been 1.8 points better in defensive rating than with average opponent shooting.

I'm half-asleep today so I don't exactly understand the implications of this, but it looks like we should be okay IF our defended 3PT% regresses to the league average, whether it's random or for any other reason. A decrease in defensive efficiency of 0.6 points per 100 possession moves us from 7th to... uh, 7th.

Maybe it means teams have had an abnormally high FT% against us so it evens out??

[GameOver25]

rcklss and tooler made some noteworthy points in the game thread that I thought was worthy of further discussion:



Is it that we're tempting teams into taking threes that they think are open but turn out to be contested because of great athleticism and hustle on our closeouts?

Is it that we're doing such a good job of scrambling to contest everything that teams are having to settle for contested threes late in the shot clock?

Is it that we're just getting lucky?

I'm not at all sure, but I'm leaning toward the first option, which seems like something that teams will eventually scout and plan for.

How would a team even scout or plan for something like that? I don't think that's possible.

Pretty simple, actually. If you have a defense closing out extra hard on 3s (which I agree we do) then you pump fake and drive off them. You can check out Nicholson against the Celtics for the perfect example.

I think I saw the Jazz try that a once yesterday (I think it was Alec Burks in the corner) and we didn't fall for it.

[Orlwillbeback]

Actually he did answer me when I asked a second time.

Tim (Florida)

How worried should I be that the Magic have one of the lowest opponent 3-point percentages but allow one of the highest number of attempts? You've said opponent 3PT% is pretty random.

Kevin Pelton (1:01 PM)

Due for a little regression there but other teams are due for more.

Tim (Florida)

Sorry, I'm a little confused when you say the Magic are due for regression in opponent 3PT% but other teams are due for more. What did you mean by that?

Kevin Pelton (1:45 PM)

I have an adjusted defensive rating in my spreadsheet that assumes league-average 3P% and FT% for opponents. The Magic's adjusted defensive rating is 0.6 points per 100 possessions worse than their actual defensive rating, which is a difference but not a particularly exceptional one. Chicago, for example, has been 1.8 points better in defensive rating than with average opponent shooting.

I'm half-asleep today so I don't exactly understand the implications of this, but it looks like we should be okay IF our defended 3PT% regresses to the league average, whether it's random or for any other reason. A decrease in defensive efficiency of 0.6 points per 100 possession moves us from 7th to... uh, 7th.

Maybe it means teams have had an abnormally high FT% against us so it evens out??

Quality analysis from our biggest fan, Mr. Kevin Pelton. Sheesh. Did Mickey Mouse sleep with his wife or something?

[MagicStarwipe]

When we're really on, I love watching our closeouts. Especially from Payton, Oladipo and Gordon.

[Catledge]

How would a team even scout or plan for something like that? I don't think that's possible.

Pretty simple, actually. If you have a defense closing out extra hard on 3s (which I agree we do) then you pump fake and drive off them. You can check out Nicholson against the Celtics for the perfect example.

I think I saw the Jazz try that a once yesterday (I think it was Alec Burks in the corner) and we didn't fall for it.

Beyond the pump fake specifically, it seems like coaches would start giving the general instruction to not settle for 3s because these guys are really good at contesting.

In any event, Pelton's explanation that the difference in drtg is only .7 eases my mind quite a bit. Thanks to tooler for getting that answer.

[TreasureCoast]

Kevin Pelton hates us

[tooler]

https://thecauldron.si.com/there-s-a-d-in-orlando-47cdd3c64405#.20x8u34kd

This thread is a better place for it than that locked speculation thread.

Super important section given our concern about giving up 3-point shots:

For a lot of those young guys on the team, having a more defined set of rules in place has helped simplify things and given them the certainty of knowing right where they should be at all times. The Vaughn-coached Magic would often change up their coverages based on that night’s opponent, but Skiles has them playing things the same way, every single game.

“It gives us an identity defensively, which is one thing we didn’t have last year,” Gordon says. “We were trying to do 100 different things according to the team that we were playing. Now teams have to make adjustments to what we’re doing.”

For this season, we probably need to just run the risk. Maybe next season we can change up coverages once the guys have the basics down (plus a few obvious adjustments we're already trying to do, like not leaving Korver).

Perhaps that's why Skiles teams reach their defensive peak in their second year.

[tooler]

While I'm here, anyone curious about why our defense has improved should watch this practice interview with Skiles around the 5:45 mark:

http://www.nba.com/magic/video/2015/12/19/1450553490973-skiles_121915.mp4-182414

I've transcribed some quotes that might help.

We attempt to pressure the ball without gambling. We try to keep everything out of the middle of the floor and down to the baseline. And we try to contest almost everything. If a team starts their offense [a certain way] we try to deny that.

Generally we want everyone on the strong side, we want everyone over top of us, and we want to defend the rim. We've had some really nice verticals there lately.

We want to be moving on the pass (run while the ball is in the air, since a pass across the court will take a couple seconds).

No matter what the "Scott Skiles principles" are, if people are hanging 120 on us every night the guys will lose faith. The guys bought in and for the most part it's working.

You do have to have some individual athletes [on the perimeter] that can pull it off. If you're not, you have to be really sharp with where you're supposed to be. Dedmon in the past has bailed the team out and we challenge our perimeter guys to do their job.

Our goal is to fight over the top of all screens. Victor is absorbing some punishment but he's also dishing it out to the big guys setting screens. [Victor's] defense has a trickle-down effect because you stand out if you don't join in.

Bonus: Skiles admits it was his fault to play Victor and Evan so much early in the year with that weird lineup. So he does take the blame for something!