Sometimes Help Hurts (Post Play and Defensive Schemes)
Take your time, Dwight On the podcast on Tuesday, we talked briefly about posting up, in general terms, as an offensive strategy. The subject of post play has become a hill upon which a lot of the "back in my day" types have chosen to die - you can't win in the NBA without a post presence, because, PRESENCE. Ignoring, of course, the changes in both the rules in place and the skillset of the modern big man. Taking into account those factors, many people smarter than me have wondered aloud what's so special about posting up?
The median outcome of a non-mismatched post play seems to be a jump hook or contested turnaround jumper from about 8 feet, smack in the heart of the "deadzone" between rim attempts and 3 pointers. This isn't to say that posting up has no value, however.
Wise basketball heads both young and old have talked recently about why feeding the post isn't necessarily done with the intention of the post player scoring, but more to initiate the offense by getting the ball into the middle of the floor and forcing the defense to react. In effect, the threat of the post up score is meant to open up other high value shots either at the basket or from 3. What happens if this threat is not seen as credible by the defense? When talking about Blake Griffin's lack of post up prowess, I've talked about how the recognition that he is not at his best on the block can make those post touches ineffective for the Clippers:
Wise basketball heads both young and old have talked recently about why feeding the post isn't necessarily done with the intention of the post player scoring, but more to initiate the offense by getting the ball into the middle of the floor and forcing the defense to react. In effect, the threat of the post up score is meant to open up other high value shots either at the basket or from 3. What happens if this threat is not seen as credible by the defense? When talking about Blake Griffin's lack of post up prowess, I've talked about how the recognition that he is not at his best on the block can make those post touches ineffective for the Clippers:
An interesting upshot some teams becoming more analytically-minded is the degree to which they respond to the threat of post scoring. There is a growing recognition, even among advocates of post play that low post scoring is not particularly efficient (depending on the data, possessions which end with a post up shot, shooting foul or turnover are between 2 and 4 points/100 less efficient than a league average possession - though post up possessions are probably slightly above league average in terms of "planned" possessions, non-fast break, non-put back situations). This recognition has led some teams to be very hesitant about doubling or helping on the post player.
One such team is Portland, who not only lead the league in exciting games, but also in things I find interesting to talk about. Last Thursday, Dwight Howard "destroyed" Portland's single coverage to the tune of 32 and 17 with only one turnover. However, Houston lost the game fairly comfortably in the end, shooting 5-20 from 3. That is well below their season averages of over 27 attempts (leading the league), and nearly 10 makes (2nd) per game, converting at over 10% worse than their seasonal average of 35.7%.
The following night, Houston played Golden State, and despite the Warriors having a post defender of similar ilk to Portland's Robin Lopez in Andrew Bogut, the Warriors elected to double Dwight Howard as soon as he started dribbling into his move in the post. Howard produced a much more modest 18 and 11 with 5 turnovers. However, the Rockets prevailed, shooting 12-29 from 3, bettering their season averages in makes, attempts and percentage (41.4%).
I don't think these two outcomes are unrelated. While I'm hesitant to declare the coverage of Howard the sole or even primary reason for the differing outcomes, (Make or Miss League is still a primary factor on a game-to-game basis) it has a lot to do with it. Posting up has two possible positive outcomes - either the guy who gets the ball scores or gets fouled, or the defense reacts to make him pass the ball, but in so doing puts other defenders at a disadvantage. Portland essentially asks the question "can you score, because we aren't going to help?"
One such team is Portland, who not only lead the league in exciting games, but also in things I find interesting to talk about. Last Thursday, Dwight Howard "destroyed" Portland's single coverage to the tune of 32 and 17 with only one turnover. However, Houston lost the game fairly comfortably in the end, shooting 5-20 from 3. That is well below their season averages of over 27 attempts (leading the league), and nearly 10 makes (2nd) per game, converting at over 10% worse than their seasonal average of 35.7%.
The following night, Houston played Golden State, and despite the Warriors having a post defender of similar ilk to Portland's Robin Lopez in Andrew Bogut, the Warriors elected to double Dwight Howard as soon as he started dribbling into his move in the post. Howard produced a much more modest 18 and 11 with 5 turnovers. However, the Rockets prevailed, shooting 12-29 from 3, bettering their season averages in makes, attempts and percentage (41.4%).
I don't think these two outcomes are unrelated. While I'm hesitant to declare the coverage of Howard the sole or even primary reason for the differing outcomes, (Make or Miss League is still a primary factor on a game-to-game basis) it has a lot to do with it. Posting up has two possible positive outcomes - either the guy who gets the ball scores or gets fouled, or the defense reacts to make him pass the ball, but in so doing puts other defenders at a disadvantage. Portland essentially asks the question "can you score, because we aren't going to help?"
And Howard did pretty well against single coverage, getting 17 points on 15 possessions against Lopez with only a single turnover. But that's well above average for him this season, where even after a recent surge in production, he's scoring a paltry 73pts/100 possessions from the post, where he's turning the ball over more than 1 of every 5 times. And it wasn't enough. As detailed above, he is created zero in the way of opportunities for others as the Blazers blanketed the Houston shooters instead of doubling or digging down onto Howard.
On top of the possible mono-dimensional nature of a post play allowed to progress without the defense reacting or helping, these plays chew up shot clock. Especially for guys with a somewhat lumbering style based on sheer brawn in the post like Howard or Andrew Bynum, this approach is very much like the running game in football, with the offense executing fewer plays in the same amount of time. For example:
On top of the possible mono-dimensional nature of a post play allowed to progress without the defense reacting or helping, these plays chew up shot clock. Especially for guys with a somewhat lumbering style based on sheer brawn in the post like Howard or Andrew Bynum, this approach is very much like the running game in football, with the offense executing fewer plays in the same amount of time. For example:
Houston walks the ball up and decides to post Howard
They use almost 5 seconds clearing space to get him the ball
It takes Howard another 5 seconds to shoot. Houston has used more than half of the clock available on this possession to create this shot, and since Portland never helped, they've used that time to create this shot only. Seconds on the shot clock are a precious commodity, and unless you can generate consistent high value looks, spending 10-12 seconds to run a post up does not seem like a great investment of time.
Of course, it is possible to generate quick post looks without running so much time off the clock, as in the following play where DeAndre Jordan executes a "rim run" in transition, burying Kevin Love under the hoop and scoring a layup in 6 seconds of shot clock time:
Of course, it is possible to generate quick post looks without running so much time off the clock, as in the following play where DeAndre Jordan executes a "rim run" in transition, burying Kevin Love under the hoop and scoring a layup in 6 seconds of shot clock time:
Getting back to the quandry of whether to double Howard or not, sometimes Dwight makes a good move, or the defender makes a mistake and Howard dunks and it looks great (complete with magazine cover flashbulbs):
But other times, he shoots a janky spinning lefty hook shot.
It's easy to remember the dunks, but we tend to forgot the more forgettable possessions as viewers. Adding them all together, we get possessions that are just alright but with limited upside, because the post player shooting has been the only option. By comparison, Golden State decided to double when Howard started his move with a dribble:
The Warriors are completely focused on Howard, to the point that there are three Rockets spacing the floor who will either have open 3s our will require tough close outs. This strategy successfully takes the ball out of Howard's hands but leaves them open to penetration:
Given Houston's ability to hit 3s not to mention the ability of a dribble penetration play to put fouls on defenders, this outcome is probably more dangerous that letting Howard work. Of course, in the above example, Howard is being guarded by the clearly overmatched Speights.
Size and strength mismatches are a big reasons why Portland's "never double" strategy is not always correct. Sometimes personnel requires an adjustment. On Sunday, Josh Smith had his best game as a Piston by taking advantage of the matchup advantages created by playing him at small forward. He used his combination of strength and skill to dominate Batum and Matthews for 18 points on 10 post-up possessions (including 3 and 1s!)
Of course, posting Smith in the monster lineup has obvious downsides for Detroit. If the threat of Smith posting is taken appropriately seriously and help does come, Monroe and Drummond are not exactly floor spacers or guys who thrive on attacking closeouts (as Parsons does in the Houston-Golden State example above). I think Pistons' opponents would be happy to live with the endless series of Brandon Jennings isolations which would be created by doubling Smith.
Of course, posting Smith in the monster lineup has obvious downsides for Detroit. If the threat of Smith posting is taken appropriately seriously and help does come, Monroe and Drummond are not exactly floor spacers or guys who thrive on attacking closeouts (as Parsons does in the Houston-Golden State example above). I think Pistons' opponents would be happy to live with the endless series of Brandon Jennings isolations which would be created by doubling Smith.
Deciding how often to look for post ups is part of the same calculation as any offensive set - what's the matchup, how will the defense likely respond, and most importantly, where can we get our man the ball. Despite the fact that good post moves are dead sexy, there is nothing inherently special about back to the basket play that makes an 8 foot shot better than a four foot shot. The best part about the DAJ rim run possession? Where he caught the ball:
It's hard to come up with a BAD shot if you're starting from there.
Mainly though, a dissertation on post play is really just an excuse to post this:
RSS Feed