Hack-a-Why

Several weeks ago, a colleague turned me on to David Freedman's 20 Second Time Out blog about the NBA. I don't really agree with all the themes Freedman develops, but one very solid point he's been making is that despite Greg Popovich's sterling reputation and solid track-record of success, his love of hack-a-X plays -- whether "X" is Shaw or Tyson Chandler or whomever -- doesn't make a great deal of sense.

To get it down to cold, hard math the NBA's top team in terms of offensive efficiency this year, Phoenix, scored 1.11 points per possession. Even if you assume no chance of offensive rebound that's equivalent to giving up two shots to a 55 percent free throw shooter. And that's the average for the best offense in the league. San Antonio only gives up 0.95 points on its average possession which is equivalent (again, wrongly ignoring the possibility of an offensive rebound) to giving up two free throws to a 48 percent free throw shooter. In general, bad as guys like Shaq and Chandler are at shooting free throws, for either of them two free shots is still a more-efficient-than-average offensive possession.

As I say, Popovich's love of this tactic is unusually odd since the Spurs are a very good defensive team. The Wizards, by contrast, are a not-so-hot defense that was at times facing off against Ben Wallace who's even worse than Shaq or Chandler, creating a situation where hack-a-Ben really might have been a good idea.

Comments

you should take into account the benefit derived from disrupting the other team's flow as well.

Right, in the short term, on those possessions, a team could actually end up conceding more points, as you aptly point out that Pt/possession averages are still worse than the worst free throw shooters.

But by giving up those points, you can take a team out of it's rhythm, and get in their head even, and that could be more important. Especially if that the team that's getting hacked is on a run or has good momentum.

Posted by nattyb | May 5, 2008 11:21 AM

nattyb - isn't this the flaw of the whole hardcore stathead analysis of basketball - that it utterly ignores imponderables like psychology and the like?

That whole way of analyzing things always strikes me as deeply autistic - if it doesn't show up in some quantifiable way, it must be meaningless.

Posted by John | May 5, 2008 11:30 AM

There's also the benefit of giving your players some rest without having to rotate them in and out of the game.

Posted by Alex | May 5, 2008 11:36 AM

Question for the NBA historians here: what was the origin of the 2-to-make-1 and 3-to-make-2 rule? I assume it was to discourage fouls, maybe in particular at the end of the game, although it only applied to shooting fouls. Did it stem from a particular incident or deliberate strategy?

I think they abandoned the rule in the early 80s, although the NBA has maintained the 2-shot non-shooting foul in the penalty (rather than the 1-and-1). Does anyone remember why it was rescinded?

Posted by Pesto | May 5, 2008 11:37 AM

And the truly amazing thing is, I suspect if you tried to analyze politics or foreign policy the same way, Matt would laugh you out of the room.

Steve Nash himself said that the hack-a-Shaq took him out of the rhythm of the Suns' offense. But what does Nash know -- there are numbers to be crunched!

Posted by Swopa | May 5, 2008 11:37 AM

There is something you neglected to mention. Phoenix scores 1.11 points/possession. Look at the subset of possessions where the play develops to the point that Shaq has the ball close to the basket; the points/possession is probably higher. I don't know if this changes things much, but should be kept in mind. If you only foul when Shaq looks about to score, that is better than fouling EVERY possession, even when some guy jacks up a 3.

Posted by guy | May 5, 2008 11:38 AM

While I would like to agree as tactics like hack-a-whoever are part of what has made the NBA unwatchable in the past few years, it seems there is one flaw.

As you note the best offenses have around a 55% chance of scoring on any given possesion, however I am not sure this is the right metric. When a guy as bug as Shaq gets the ball close to the basket it seems to me that, barring a foul, his chances of scoring are better than 55% so puttin him at the line seems like a reasonable calculation. Plus you get the added bonus of frustrating him, disrupting the other team's flow (as noted above) and making him more sore the next day.

Posted by DP | May 5, 2008 11:39 AM

That whole way of analyzing things always strikes me as deeply autistic - if it doesn't show up in some quantifiable way, it must be meaningless.

This is a valid critique of many stat-influenced analyses, but I don't think it fits here.

Matt isn't saying that it has been demonstrated that hack-a-X is a bad strategy. That would be to miss the thus-far unquantified effects of the strategy.

Rather, the point is that we do have some good information to create a baseline for measuring the strategy's effectiveness. If it adds ~.15 points per possession when used, then it has to, in some fashion, increase the team's success on offense and on other defensive possessions by an equivalent factor to be worth it.

I have to say, I'm skeptical that the effect is quite that large. But if someone were to study the issue and make an argument, it would be worth seeing.

The numbers should function as a good baseline for discussion and a place from which we can begin to create more and better ways of understanding sport. They should enable rather than constrain discussion and study, and I think that's all that Matt's post here is doing.

Posted by DivGuy | May 5, 2008 11:41 AM

Taking Phoenix out of an offensive rhythm: Priceless.

Posted by Eric | May 5, 2008 11:45 AM

Popovich's love of this tactic is unusually odd

Not like it's won him any championships or anything.

If any coach should change what he's doing, it's Popovich.

Posted by SoCalJustice | May 5, 2008 11:50 AM

What is interesting to me is that I don't remember the Spurs ever doing this before.

The Chandler foul the other night was especially bewildering because it came at a time in the quarter where there was very little to be gained by doing it - something like 30 seconds left in the half. Bottom line is that each team was going to get one possession, and instead each team got one possession plus New Orleans got two free throws, which they made.

The times that the hacking was especially successful in the Phoenix series were when the Phoenix defense was really poor against the pick and roll, and Parker and Ginobili had great rhythm, so they packed in extra possessions while our guys were scoring well.

Agreed with the posters above about clobbering Shaq when he gets the ball five feet from the rim. But these fouls occurred when he was not yet across half court. And somehow Jacque Vaughn grabbing his jersey isn't going to leave him sore the next morning.

Posted by Curtis | May 5, 2008 12:02 PM

When a guy as bug as Shaq gets the ball close to the basket it seems to me that, barring a foul, his chances of scoring are better than 55% so puttin him at the line seems like a reasonable calculation.

But that's not hack-a-Shaq, that's just "fouling Shaq." Hack-a-Shaq typically refers to fouling Shaq off-ball when you're in the penalty to put him on the line before his team has a chance to set up their offense.

Posted by Jake H. | May 5, 2008 12:02 PM

With trepidation, I'll invoke Gregg Easterbrook of Tuesday Morning Quarterback, who constantly argues that many coaching decisions in the NFL (like punting in the opponent's territory) actually go against the numbers, but they are nevertheless done to shield the coach from criticism of tactics. The Hack-a-X strategy makes it look like you're doing something significant even if you aren't (which actually seems kind of odd from Popovich, who has made it clear by his actions that he's immune from complaints that he puts an un-entertaining form of basketball out on the court).

Posted by Indecisive | May 5, 2008 12:07 PM

First of all, Shaq is a career 52% FT shooter. Since 2003 he has only broken 50% once. It makes sense!

Second, I think this has to do with the discrete probability of individual outcomes.

... Wait, it doesn't even make sense there either. You have a 24.7% chance of losing if you just let the Suns shoot (that's "Suns score"*"Spurs miss"), but a a 42.3% chance of losing if you foul Shaq. To believe that this make sense requires you to believe that there's something fundamental about end-game situations that makes it much more likely that the spurs will score.

I suppose that if Shaq is in foul trouble, that might be enough to take the Spurs effective FG% much higher such that this strategy makes sense.

Posted by Nicholas Beaudrot | May 5, 2008 12:18 PM

Socaljustice, so you're saying: if the Spurs use Hack-a-Shaq, and the Spurs win titles, therefore Hack-a-Shaq helps them win titles. I don't think that's right.

Just because someone is a good coach doesn't mean every decision they make is perfect. I think the Hack-a-Shaq was defensible for the Spurs (although the league should put a stop to it), but fouling Tyson Chandler away from the ball is bizarre.

Posted by too many steves | May 5, 2008 12:37 PM

Probably makes sense for a player who shoots ~50%. The difference between the expected result from the free throws and the expected result from playing straight up defense is nominal, and you do get the intangible benefits mentioned above (slowing tempo, whatever). However, when you get to ~60% free throw shooters, the difference between the expected result from the free throws and the expected result from playing straight up defense is too great and outweighs the other benefits.

Posted by Al | May 5, 2008 12:44 PM

Just allow the hacked team to transfer the personal foul to a player of their choice on the hacking team. Your scrubs hack --->> your star fouls out.

Posted by milo | May 5, 2008 12:47 PM

The last time the Hack-a-Shaq was used so prominently were the championship teams of the Lakers. The Pistons beat the Lakers by playing Shaq straight up, not by hacking him.

Posted by Mo | May 5, 2008 12:48 PM

Steve Nash himself said that the hack-a-Shaq took him out of the rhythm of the Suns' offense. But what does Nash know -- there are numbers to be crunched!

While it is worth considering what Nash says about the tactic, it's also worth noting that they were disrupted sounds a lot better than saying:

"Despite being a two time MVP, I've never been able to take a team to the NBA finals, let alone a championship, despite having some genuinely talented teammates. I really am beginning to wonder whether my entire approach to the game is sound in a playoff environment."

But hey, why consider stats when there is anecdotal evidence to hear!

Posted by Brad L | May 5, 2008 12:54 PM

Shaq is a streaky foul shooter, as are many with bad mechanics. Sending him to the line early and often lets you see if he is "on" or "off". If he's on, drop the plan. If he's off, he'll be shooting below 50%.

This tactic improves if you have critical information. Perhaps he does something that belies some pain that affects his foul shooting. For most players, a small affect would drop them from 70%+ to 60%+ - not worth hacking. For those mentioned, if you reliably know that they won't be up to their normal low standards, it is worth it.

Posted by Njorl | May 5, 2008 1:26 PM

The Utah Jazz were the most efficient offensive team this year, not the Suns. They scored `113.8 points per 100 possessions, to the Suns 113.3.

I see your point here, and I don't know I disagree with it. But I think the strategy worked fairly well against the Suns. Shaq shot 50% on his free throws in the series, so the Suns did score less than they would have otherwise.

Posted by Flint | May 5, 2008 1:28 PM

sometimes you do things in basketball (well, other sports, too) as much for their psychological impact as for their tactical outcome, which is to say that hack-a-shaq tends to frustrate the big guy and his teammates and that has some value too.

but really, i'm only bothering to comment because it gives me a chance to reference an article that sports illustrated recently put online penned by bill russell in 1966, "the psych and my other tricks:"

http://vault.sportsillustrated.cnn.com/vault/article/magazine/MAG1077812/index.htm

here's a little sample:

The first thing I am not about to do is look up the definition of psychology in the dictionary. Why bother? I mean, dictionaries are nice and all that, but did old Daniel Webster ever have to stand there at the top of the key and define five sweating monsters rushing down at him? He did not. Well, then.

I will not confuse you with Webster's words, because my definition of psychology is something else again, and I have been practicing it for a whole flock of years now and I ought to know. In my psychology you wear short pants and tape and sneakers, and this is the kind of thing you do:

Say I am standing next to a rookie who has just come into the game—some hotshot college All-America who is not yet used to his rookie role. The action is swirling all around him, and I say to him, casually, "Hey, what's the matter with you, baby? Don't they ever pass that ball to you? What are you, a nothing on this club?" Oh, yeah, they laugh it off. But you can see them thinking about what you said.

Or I find someone who is new in the league, and I stand next to him and hack and cough it up. Sometimes I feel I should get an Oscar for this. I know they're watching me out of the edge of their eyes, and they are figuring, "So this is the great Bill Russell. Hell, he's just a tired old cat. And here I am, as fresh as can be." They don't know that I have a reserve tank.

You say these are minor league tricks? Maybe. But you'd be surprised at how often they work. The thing is, you have to pick your spots. Let's say you are playing center opposite Wilt Chamberlain of the Philadelphia 76ers, and it is one hot and heavy game. The score is just about even, and it is the middle of the second quarter—the time when you're most tired before getting your second wind. Tired? Listen, you are so tired that your leg muscles burn, and you know in your heart that Wilt is as tired as you are. But you are both breathing shallowly so as not to give any sign of how you really feel. Now. Wilt is on defense, and he is leaning on you with all of his 250 pounds and you have your mouth up close to his ear and you say to him, pleasantly, "Hey, baby. I never thought I'd see the day when a great big guy like you would be pushing an old man like me around."

So what does Wilt say to you? Wilt says, "Don't give me that old psych, baby." (I have cleaned up that quote. I have also shown that psychology does not work every time. The trick is in knowing who to talk to under the basket.)

Posted by howard | May 5, 2008 1:44 PM

a chance to reference an article that sports illustrated recently put online penned by bill russell in 1966,

That is a great article!

Posted by NickS | May 5, 2008 2:00 PM

The big flaw in all Matt's talk about basketball is that basketball ain't baseball. It can't all be broken down into stats, it just doesn't work that way.

If it could, you'd believe KG isn't even close to a Top 5 player based on his stats this year, but basketball is a game where fundamentals often can't be measured, and the naked eye is necessary to asses players.

This year, the Hack-A-Shaq worked because Shaq wasn't really the target of it. The Suns offense was. The Spurs got in their heads (again!) and it totally took them off their game. It has nothing to do with stats. As your first commenter pointed out, it has to do with flow and psychology.

And it was BRILLIANT strategy, even though the stat geeks can't see it. But anyone who watched the games saw what it did. Turned it into a half court game, because the Suns couldn't build any momentum, which is what their running game is built on. Without that, they are not a very good team. Certainly not a great one.

And plus, Shaq is old and sort of sucks now, certainly compared to the Shaq of the Lakers.

Posted by Hack-A-Shakira | May 5, 2008 2:07 PM

Really people, go read that article. I can't imagine a player of that stature writing something like that today.

Posted by nolaboyd | May 5, 2008 2:19 PM

Too many steves,

I'm not saying every decision he has ever made is right.

I'm saying they beat the Suns in 5. He's built a dynasty by NBA standards. His strategy is pretty good.

And I don't expect him to change it based on season average statistics that don't really take into account head-to-head matchups with the individuals players in question into account.

Obviously the strategy is not perfect and can't/shouldn't be used across the board. And if he's relying (doubtful) on that to get by New Orleans, it would be a dumb move.

But it worked against Phoenix, despite their points per possession season average statistics.

Posted by SoCalJustice | May 5, 2008 2:26 PM

nick s, nolaboyd: being an old geezer, i actually remember that article from when it first appeared, so you can imagine how excited i was when sports illustrated put it online.

and the second thing i said to myself (after "wow, this is great!") was "i'll use that article more than once during the playoffs over at matthew's, where other people might enjoy it too."

and nolaboyd, no, i can't imagine a big-name athlete in 2008 writing something like that for publication, and even if one did, the front office would never allow it....

Posted by howard | May 5, 2008 2:33 PM

Matt, like everyone else in the world who didn't live in Los Angeles during the three-peat, you don't understand the actual point of Hack-A-Shaq.

Most people think hack-a-shaq is a good method for catching up when you're behind, but that's not true. As you noted, making a point on every possession is enough to hold a lead so long as you play good enough defense so the other team isn't scoring on every possession.

BUT look at how Popovich, a basketball genius used it. He only rarely used it when the Spurs were behind, instead, using it after they'd gotten a lead to randomly disrupt the flow of the Suns offense. When the Suns were behind by ten, they couldn't catch up with Shaq getting one point a possession, and he didn't do it every possession, just a few at a time to disrupt and confuse the Suns.

This is the lesson from Shaq with the Lakers. Hack-a-Shaq only works if you're already winning but worried that Shaq's team is about to go on a big offensive run.

Posted by Simon | May 5, 2008 2:33 PM

The answer is: the Spurs are old and slower, so it's a bigger benefit on their other defensive possessions, with a low expectation of damage for the hack-a-possession. Against running teams they are better served by limiting the damage of their fast break over the course of the game. This is weighed against the risk-reward proposition of allowing a bad free throw shooter, called out public, to attempt to make both shots during multiple trips to the line and burn you on the strategy. The odds are that you wind up basically even in terms of points scored (see Shaq's 50% FT shooting in the series). The gain is realized 1) on later possessions with your preserved energy to get back on defense and 2) if you are losing and/or in an end of quarter scenario by providing more pace and clock management capability. In that way, it's loosely pareto-optimal--the expected damage is not much worse on the hack-a-possession than any other given posession (you expect to be no worse off or only slightly worse off b/c of the foul), but you are certainly better off if you prevent a fast break chances at easier baskets during later possessions and/or if you manage the clock so that you end up with more possessions played at your pace. Also, it provides for more half-court strategy, given the normal course of line up switches and possession by possession match-up gamesmanship that usually develop at the end of a quarter when fouls allow for substitutions. The subbing by itself can be enough to disrupt the other team's rhythm. B/c you've fouled you expect to go on offense and so can try to force can or else force the other team to accept a defensive possession with an obvious mismatch. And the poor free throw shooter (always a big guy) can't be removed if he's shooting on the free throw dead ball (unless he makes his second shot, or is that only in college?) so if he has fouls you also get the benefit of keeping him in the game.

Honestly, while I understand that people like to see the ball whip around and generally prefer up tempo basketball, I think the hack-a-shaq adds a huge element of strategy to the game. It forces the fast team to play their fast break possessions with greater urgency, and it provides more opportunities for all sorts of chess match stuff in the half court. It leads to more mismatches and double teams/defensive rotations that end up with the offensive team moving the ball around looking for the right shot. And I definitely enjoy watching Duncan, Parker and Ginobili zip the ball around until they find the right open shooter, forcing the defense to scramble, always anticipating and a step ahead. To me, that's really fun stuff--more so than watching Nash and Amare, Deron and Boozer, Paul and X run the same pick and roll 50 times a game.

Posted by MSB | May 5, 2008 2:40 PM

It's not that basketball can't be captured by statistics. It's that we don't have the statistics yet to capture it. For example, KG. The Celtics have proprietary stats that show that he holds the guy he's guarding to some ridiculous number like 14% below his usual fg%. That's a hugely valuable stat. Factor that into KG's other numbers, and now he's near the top of the league, just like we all think he should be.

That's just man-to-man defense. I'd also like to know how good a guy is at setting picks, at running off of them, at rotating to help on D, at making the "hockey assist." It would be possible to measure all of this if you charted every possession, but you'd either have to have a team of 30 stat guys working every game or you'd have to do it with fancy-schmancy computers.

Posted by too many steves | May 5, 2008 4:04 PM

too many steves, i recall that when riley coached showtime, they had some kind of stat they kept on rebound effort or something along those lines, which is what convinced them to give rambis minutes.

the problems of basketball stats are, of course, similar to those of soccer, which are also slowly, steadily improving (for example, simply tracking the percentage of completed passes is a very useful data point) and now people are starting to track things like "sucessful dribbles" (although i don't know the criteria).

Posted by howard | May 5, 2008 4:11 PM

"Hack-a-why?"

Are you making fun of me, Yglesias?

You'll pay for this - more Corrs references, more Terminator references!

I got a million of 'em! A million of 'em!

Posted by Richard Steven Hack | May 5, 2008 5:11 PM

I'm not sure I buy Nash's claim that H-a-S took them out of their offensive 'rhythm'. It's just an excuse for losing. Maybe it did have a psychological impact. I don't think it effects Shaq, b/c he's gotten really good at dealing with that kind of crap over the years. But if Nash couldn't handle additional breaks in the flow of the game, then he isn't as good as advertised. It's really no different than a few extra timeouts. The Suns weren't running anyone of the court with Shaq regardless.

Posted by mpowell | May 5, 2008 5:35 PM

Sharing any wisdom from Tuesday Morning Quarterback writer Gregg Easterbrooke should be done with trepidation. The man is a hack, rarely knows what he's talking about, and comes up with ridiculous theories nearly every article. He was good for the first year or so, and it's been all downhill, to the point where I rarely read his stuff (and if I do, it's mostly to ridicule on other sites).

Posted by LnGrrrR | May 5, 2008 11:35 PM

Good article. Of course, these days, it would be a blog post (see, e.g., Gilbert Arenas's blog) rather than SI article.

Posted by Al | May 5, 2008 11:45 PM