The Rockets and the Use of Screens

The question is do setting screens have any drawbacks? If not, then why not set more quality screens? Even if the Rockets' offense is great as it is, if doing something can improve it even more, why not?

Remember it's not just about number of screens, but number of *quality* screens that result in points (aka, non-passing assists). Heat may run total number of screens that are below the league average, but the number of points generated by their screens puts them at 8th in the league. The other teams in meh's top 10 offense list that don't run very high numbers of screens may also run more quality screens (over pure quantity) that leads to points. But this is yet to be confirmed

 

our bigs cant shoot. pick n pop is non-existent. high pnrs are invitations to trap the ball handler...

 

Analytical analysis is not the ultimate standard, and there may not be enough and perfect data to analyze and draw conclusions.

I think the issue relies in Howard's reluctance to set picks and our primary ball handler - Harden - doesn't really use screens that much. Harden's excellance in PNR relies more in his superior shooting/attacking rather than running the offense and getting teammates involved.

We can get away with winning on pure talent in regualr season, but I doubt it works against elite teams in playoff. We need strategy and screens to win against teams with more talent.

 

I'm going to take a blind guess that setting a screen for Lebron will result in a bit more efficiency than setting a screen for Brandon Knight. So my response is, "DUH!?"

The Heat doesn't have the best offense in the NBA. If setting screens is so important, why not set more? Pat Riley, Lebron, Wade, Bosh, Battier, etc. are probably smart at basketball. So the fact that they don't run as many screens as the likes of OKC or Milwaukee should indicate something?

 

My earlier post already addressed this:

"Lastly, I will say there is one criticism I have of regarding the whole points-generated-by-screens idea, which is that offensively efficient teams may simply have good players who score regardless of having screens or not. If you have a player who is offensively gifted like Tony Parker, how do you know that a good portion of his points are "generated by screens"? Maybe he was going to score regardless, which would artificially inflate the points generated by screens variable.

But then, assists are the same way. You can't really know if the assist is what *caused* a made field goal, and having offensively gifted players would naturally inflate assist numbers (while having players who can't shoot would deflate assist numbers, no matter how good your ball-movement is). But no one is arguing that assists are not important for an efficient offense. The assertion made in the vantage blog is that "Set Screen Points Per Chance" is basically a "a non-passing assist". So if we regard assists as important for an efficient offense, then the argument can be made that setting high quality screens that can lead to easy points is significant as well."

1. Maybe because, as already suggested with regards to why Morey and co (obviously a smart and competent bunch) have not put much emphasis on setting screens, there are some things they see that we don't. But that's kind of the point of this thread I think, trying to figure out why specifically certain offenses (such as ours) place so little emphasis on setting screens despite the correlation found between offensive efficiency and "Set Screen Points Per Chance". Maybe some offenses are designed such that setting screens aren't necessary, or not setting screens provides some benefit (but if that is the case, what are they specifically?). Or maybe you're right, and they have statistical evidence that screens don't really do anything, and all these vantage data are misleading. But if that is the case, what exactly about the data could be potentially misleading or wrong? We already identified one weakness which is that we are not sure if "Set Screen Points Per Chance" causes the offensive efficiency, or whether offensive efficiency (e.g., having someone like Lebron) causes more "Set Screen Points Per Chance." But this alone cannot entirely discount the possibility that setting screens can make an impact on the offense. As said earlier, assists are also inflated by having good scorers or having an overall efficient offense, but we can't really use that argument alone to say assists are consequently irrelevant. Assists are relevant, and the same is probably true of setting high quality screens (setting screens leading to points can be thought of as like non-passing assists)

2. Maybe because the data in support of setting screens has only recently been introduced by new tracking methods with the latest SportVU technology, so some teams have not fully embraced it yet. Prior to this, I think the only data available with regards to screens were the just the raw number of screens set by a team, which were not correlated with offensive efficiency. So maybe the value of setting screens was more debatable in the past, but with these new findings, more and more teams may possibly come to accept the importance of setting quality screens in the future (or find some new statistical evidence that screens are insignificant, in which case, see #1)

3. Maybe they know screens are important, but there comes a point where trying to set additional screens results in diminishing returns. As already stated, they are #8 in "Set Screen Points Per Chance." Maybe that's good enough for them? On the flip-side, one can argue, if screens are insignificant, then why run screens at all? If Lebron can do it all and needs no help from setting screens, then why even bother going through the trouble? The fact that they are #8 in these "non-passing assists" should still say something shouldn't it?

And I get that maybe the Rockets also have an optimal number of setting screens or something, but the reason this case is so interesting is that we are dead last in setting screens, and even the 2nd to last team in terms number of screens set is 4% higher. If this is the "optimal number," then why specifically is it so much lower? What is it about the Rocket's system exactly that necessitates this?

Additionally, Howard is not known for setting the highest quality screens, so while it is understandable for teams like the Heat to set fewer screens than average (because their screens are better and results in more points generated), it is more puzzling in the case of the Rockets.

 

My point is that there seems to be nothing publically available which seems capable of isolating this one area of basketball in terms of any causation. Which you seem to basically agree with. It's a maybe-maybe no thing. So why make a fuss about it?

This isn't like, say, 3 pointers. That is something that have been studied and proven through large sample sizes that having 3 point shooters to spread the floor is very advantageous to your offense. And something that one can see both using the eye and through stats.

 

The only problem with your raw approach here is there's a 3rd factor, which is % of half court offenses. The screens only pertain to half-court offense, not transition offense. Ex: Houston has an awesome transition offense that more than mitigates their half-court offense, thus the efficiency.

 

You are grossly over-simplifying things. Basketball offense even in its most simplistic breakdown can be split into 3 parts. Primary transition, or breakaway layups/dunks, secondary break, where the offense tries to get a good shot before opposing offense is set, and true half-court sets.

The Rockets because they push the pace not only get a lot of fast break points, but also shots when defense is still scrambling to get set. This I'm certain is factored into the stats given by the OP, because secondary breaks in most people's minds are not transition baskets. But to equate them with true half court sets is also just plain wrong.

 

someone already mentioned Rockets offense is predictable. It very much is and there's not enough helping out in terms of screens or team defense. All they rely on is outscoring the opponent which won't always work. They'll never get that coveted trophy like this. The staff are too stupid to see that.

 

That's true. I guess where we differ is for me, from reading the vantage data it seems more like a "maybe yes" to me than "maybe no". But point taken, the evidence isn't clear cut just yet, so I guess we will see in the future as more studies are done. I'm not really making a fuss about it, stuff like this just kinda interests me so I like to talk about it

 

Oh, I definitely don't mind talking about it. I just find it odd that many here uses the BR article as a sign that we should screen more. I feel the Rockets can do a lot of things better, to be sure. There are many things the Rockets can do that are CERTAIN to improve their team. Most of them on defense. But I've never watched a game and thought, "If only the Rockets set more screens, they'd be winning more."

 

I see our players on defense fighting thru picks like chickens getting dropped from a skyscraper therefore resulting an easy midrange bucket.

Do you not see that MEH?

 

I saw even Bev can't fight over a lot of screens, resulting easy buckets for opponents. And we become stagnant on offense without screens.

Our record is just fine with our talent level(no team 5-8 in the west has 2 legit near superstar level players like we do).

 

Where do you actually get vantage data on screens in terms of individual team stats (on a per game basis)? This would be an opportunity to test out with a regression model how much setting effective screens actually affects a game outcome for the Rockets (in comparison with other factors like defense, and 3 point shooting). Maybe meh has a point, that there's the possibility that other factors such as defense may be much more important and maybe setting screens is trivial in comparison, but one can't really be sure about anything without actually testing this out. As of now, it's all kind of a matter of opinion or conjecture (with the one exception of the correlational finding published on the vantage site, although not everyone finds it compelling).

Making such a model shouldn't be too bad, but I don't know where to get this kind of data (is it even publicly available at this moment?)

 

“Their screens hurt. They actually set real screens,” Lillard told CSNNW.com. “They do a great job of setting and holding screens. It wears you down. Chasing Tony Parker is one thing. Getting hit every single time is another thing. It takes a toll on you.”...

“It makes you tired,” Lillard said. “When you’re tired from chasing him (Parker) around down there, it’s tough to come down on the offensive end and get much done when you’re wore down. It’s hard. I’ve got to do a better job at avoiding screens. That’s a challenge that we’re all going to have to accept.”

Interesting Blazers went at Harden every opportunity they could. Its a pity Harden's our weakest defender too.

 

Common sense. Just like playing more mins mean a bit less energy to perform at your peak. Both of which seems somewhat loss with the Rockets

 

We have two of the best PNR guards in Harden and Lin, and one of the best roll man in Howard. Yet most of our offense is Harden ISO or Howard ISO in low post. I always wonder what would happen if the most athletic and gifted players play within a PNR heavy system...nobody would be able to defend them.

Basically Howard and Harden are using the season to develop skills, rather than incorporate their best available skills into the system. The only problem is we are not a rebuilding team so the young stars can take seasons to expand their skillset for contending years later.

 

Wish we had a good PnR point guard.