In a talk on Rockets.com with Jason Friedman, Daryl Morey gives the best explanation I've seen as to why so many teams are content to have well-covered superstars take incredibly difficult shots with the game on the line.
We’ve talked so much over the last few years about having that star come crunch time and all the mythology behind that. This team in particular though seems to fare pretty well and execute at an above average rate in clutch situations, even without having that traditional “closer.”
I think the reality is that no one is any good at crunch time. I think if you’ve got a guy who can create his own shot then you’re better off than not.
I think the biggest misnomer people have … I’ve seen a lot of things like, ‘You should run a play. You should just do your normal things.’ Well, the reason why teams go with a particular isolation play, even though that often has a low efficiency because it’s just hard to score for anybody, I don’t care how good you are, is not because teams think that’s optimal for scoring, it’s because it’s optimal for controlling the amount of time the other team has after the play. If you’re just running a set and a team jumps it or tries to disrupt it, it can really change the timing of when your shot goes off and it’s a massive, massive difference how many ticks are left when the other team gets the ball.
So a lot of what people want to criticize coaches for which is ‘Don’t they know that guy is bad in isolation; don’t they know this?’ -- it’s really because they’re not, in my opinion, thinking about the big picture which is controlling the clock the other way in terms of when your opponent gets the ball back. Even three seconds with an advance of the ball is a huge difference versus only having one second. The efficiency drop based on you controlling the clock the other way is a massive difference.
Fascinating. Would love to know the value of those three seconds.
This explains only some of the downsides of hero ball, not all. On video, many (although certainly not all, and probably not even most) of those "hero ball" shots come with enough time, and an open passing lane to an open teammate in scoring position who doesn't get the ball. (For instance, Pau Gasol or Pau Gasol.) This is, let's say, sup-optimal decision-making, even if it does have an underappreciated freebie clock-burning kickback.
In deference to Morey's point about the offense suffering in the name of the defense ... plus/minus would sniff this out. Basically, you want to see which teams, if any, consistently outscore opponents over the last few minutes of close games. Having messed with the NBA's new stats tool in this regard, I can tell you that over the last decade the Lakers have had, on average, the 10th best plus/minus in the last three minutes, trailing by three or fewer or tied. The rankings are all over the place, though, and might just be random. They had the best plus/minus in 2003-2004, and the 27th best plus/minus the next season.
Statistical support for this story from NBA.com.