Layups: Block Recovery %
Posted by Neil Paine on January 19, 2011
Here's a good post from frequent BBR commenter Imadogg at his blog:
Best Blockers in the League: Who Recovers the Block?
Dogg looked at the top shot-blockers in the NBA by blocks per game, and did some digging through play-by-play data to determine how often the blocker's own team recovered the ball immediately after the swat (this is basically the first half of the "Bill Russell stat" Simmons talks about in The Book of Basketball). For instance, 67% of the time league leader Andrew Bogut blocks a shot, the Bucks end up with possession; compare that rate to 44% for Pau Gasol at the low end of the spectrum.
January 19th, 2011 at 3:24 pm
Of course Dwight is at the low end of the spectrum, since he likes to go for the highlight reel swat, when with his athleticism he could easily just tap or even swallow the ball.
I think that's why the backboard pin block is so effective, because it gives your team a better chance to recover it.
January 19th, 2011 at 3:37 pm
I always wondered if anyone was tracking it. Very interesting stuff. Some advanced stats (PER, Win Shares, etc...) assign a particular value to blocks but do not adjust the value for whether a team actually recovered the ball or not (can't be found or estimated from the box score).
So Bogut's PER or WS would be slightly understated and Gasol's would be slightly overstated. Now Blocks only make up a very small piece of the calculation but it is still interesting nonetheless.
January 19th, 2011 at 4:00 pm
It doesn't look like it's being tracked, it looks like one guy reading over lots of play-by-plays and doing it by hand. Assuming he didn't mention programmatic scraping and analysis because he didn't do it.
I mean we could also assume Houstan's front office is tracking these things, so whatever.
January 19th, 2011 at 4:49 pm
It's a mistake to say some blocks are worth more than others, based on what happens AFTER the block. If the blocker's team fails to recover the ball, we see the effect in the box score; the opposing team gets an Orb(and the blocker's team doesn't get the Drb)and another chance to make a FG or FTs. So is that a "less valuable" block, or poor defensive rebounding?
January 20th, 2011 at 3:09 am
Re: AYC.
Let me put it to you this way: block recovery % is analogous to completed pass % in football (or basketball of course, if we had it). If a guy throws a perfect pass and the intended target blows it, obviously the passer should not be blamed for it - but that doesn't mean we should consider completed pass % an irrelevant stat when we rate quarterbacks. If we a general trend of block recovery with a particular player, this is probably meaningful.
Now, could one player be at a disadvantage due to the rebounding situation that is the actual cause of a decreased performance in this stat? Sure. For that matter, it's probably likely that the type of block (weak side, etc) is a factor too. However, I'll still keep the knowledge we have here stored away when making other observations because it seems likely to of some use.
Last, we should of course remember that quite possibly the most important impact of a shotblocker is not in his blocked shots but in his ability to force low percentage shots to avoid the block - and I don't think anyone would argue that players show less fear of Howard because of their innate knowledge of what's described in this stat.
January 20th, 2011 at 8:44 am
@AYC... You fail to discuss when someone blocks a shot out of bounds. By doing that they really haven't accomplished much because they gave the ball right back to the opposition. These instances should certainly be worth less than when a player blocks a shot and keeps it in bounds (then the rebounding part comes in).
So I think that some blocks are worth more than others.
January 20th, 2011 at 8:46 am
This is only a little related to this post...
Is there anyone keeping up with how many times players score on dunks compared to layups?
This past week I have seen a couple players blow wide-open dunks and have been wondering if they would be better served just going for the layup.
January 20th, 2011 at 11:29 am
Doesn't this stat ignore the value of the shot block out of bounds at the end of the play clock? If the shot is blocked with 3 seconds left on the clock, chances are that a lower percentage shot will be taken due to limited time to create a play. Compare this to a block out of bounds with 15 seconds left, where a team has time to run a full play.
January 20th, 2011 at 11:48 am
I didn't say block recovery % isn't a worthwhile stat. My point was that the value of a block isn't contingent upon what happens after the block. The shotblocker is like a goaltender in soccer; the value in a block comes from preventing an attempted field goal from becoming a made field goal. If you block it out of bounds, you have prevented a shot that might have otherwise gone in. Even if you block it directly into the hands of an opposing player, you have still prevented a potential basket.
Your logic can be applied to other stats. I could say "some defensive rebounds are more valuable than others", based on whether or not the the next recorded credit is a made FG, a missed FGA, a steal, or a turnover. But we don't need to muck with the value of Drb, because the next recorded credit speaks for itself. How are blocks any different?
January 20th, 2011 at 11:54 am
Some blocks are definitely worth more than others, but how can you say that a player hasn't accomplished much when he blocks a ball out of bounds? Not only has he prevented a potentially made shot, but he establishes that there is an interior defense which makes the opposing team less likely to take shots near the basket, in other words forcing the opponent into lower percentage shots.
This would be a great stat to track, but a pain to do by hand. Isn't there a way to create a program that would pull digital play-by-play? Problem is that pbp has errors once in a while, too.
January 20th, 2011 at 11:58 am
Actually, all defensive rebounds are worth the same amount. Each one creates one possession for your team. What happens afterwards in that possession is irrelevant to the fact that you've created another opportunity for your team to score.
January 20th, 2011 at 12:32 pm
Per player, it would be cool to see percentages of blocks followed by a FG by opposing team. Or LVBs.
January 20th, 2011 at 12:32 pm
If each FG attempt qualifies as a possession, than a block creates a new possession too, even if the shooting team recovers.
January 20th, 2011 at 2:58 pm
@AYC: Yeah, but if the shooting team recovers it creates a new possession for the opposing team, which is clearly less valuable than creating one for your own team. Blocks that go to your team are essentially steals, and really ought to be treated as such. Blocks that don't are really just general defense that runs down the shot clock, not any more valuable than tight defense that prevents the shot from being taken at all.
January 20th, 2011 at 3:24 pm
You're still missing my point. The negative effect of not recovering a block is already reflected in the box score, in the form of an offensive rebound and whatever points are scored off the Orb. But the value of a block in itself doesn't change either way.
January 20th, 2011 at 3:58 pm
Re: AYC.
The reason the name "Bill Russell statistic" has been used is because Russell expressed that when he blocked shots he made a point to block them to teammates to kickstart fast breaks. How can it be debatable whether Russell's single action here is more valuable than if he simply blocked the ball back to an opponent?
If you want to say that Russell's action is really a block + a defensive rebound + a good outlet pass and that his block component is no different from anyone else's, okay - but you and I both know that Russell's only going to get scorekeeper credit for the block, and thus not all occasions where the scorekeeper marks "block" indicate identical value contributed.
January 20th, 2011 at 4:50 pm
@AYC:
If a ball is blocked out of bounds it is absolutely not recorded in the box score in any way, shape, or form beyond as a block.
January 20th, 2011 at 7:06 pm
One of the tenets of advanced statdom is that when two teams play they use basically the same number of possessions. That's why shooting efficiency matters so much; if the number of possessions is the same, then winning is largely determined by how efficiently you shoot the ball relative to your opponent. The value in a block doesn't come from creating a new possession, it comes from ensuring an opponent's field goal attempt doesn't become a made field goal.
January 20th, 2011 at 7:51 pm
If it is blocked out of bounds it does not become a team offensive rebound?
January 20th, 2011 at 11:22 pm
P Middy is right. It is a team rebound for the team that shot the ball.
January 20th, 2011 at 11:28 pm
Matt,
Russell might not get credit beyond a block, but his team (a teammate) will get credit for a defensive rebound and if it is a good outlet pass that leads to a basket, then a teammate also gets an assist. AYC is arguing that the team effort is reflected accurately in the statistics, and I guess he's right. However, it's true that Russell's individual contribution to the team effort is not accurately reflected by these stats.