Feeds:
Posts
Comments

Archive for the ‘Statistics’ Category

Apparently horses in races are almost always (98%) whipped.*  The main reason is to make them go faster.

Congratulations to the scientists from the University of Sydney who won a prize for discovering that “whipping does not increase horses’ chances of finishing in the top three and that they actually run faster when they are not being hit.”

So it seems that overwhelmingly, the horse racing experts – the jockeys and trainers particularly – have for decades followed a practice which:

  1. Hurts their chances of winning.
  2. Hurts their beloved animals.
  3. Hurts their standing in the wider community.

Given how much money is at stake in horse racing, this is remarkable.  It calls out for explanation.  How can these experts have persisted for so long in such self-destructive behavior?   Here are some possible explanations.

First, the idea that whipping makes a horse go faster has a certain “truthiness.”  This truthiness made questioning the practice seem otiose.  Further, belief in the truthy proposition became a perceptual filter through which they “saw” that whipping made horses go faster.

Second, the omnipresence of whipping prevented the possibility of observing the weak correlation between whipping and losing.   If only about half of the jockeys whipped their horses, then over time people might have started to notice that the unwhipped horses tended to win a bit more often.  Or at least didn’t lose more often.

Third, since everyone else was whipping, everyone naturally assumed that whipping was the “right” thing to do.   Failing to whip the horse would look rather odd – especially if you didn’t win.   As Mark Twain wrote: “We are creatures of outside influences; as a rule we do not think, we only imitate.”

* I don’t know anything about horse racing.  This claim seems implausible, but here is my source.

Read Full Post »

A version of  my Quadrant essay ” The Wise Delinquency of Decision Makers” was recently broadcast on ABC Radio National’s Ockham’s Razor program.  Audio and transcript available here.

Robyn Williams, the great science journalist  and host of Ockham’s Razor, introduced it thusly:

I remember a few years ago being on a committee choosing some prize winners, for innovation, or some such recognition. There was a new system in front of us, with grids, numbers out of ten in about twelve categories, cross-over elimination processes, all in all the same kind of chart you imagine The Pentagon using to invade Iraq in 2003.  Now I happened to know who should have won the main prize. It was obvious; Malcolm Gladwell wrote all about this in Blink, you just know. But the process didn’t allow any of that. We ground through the chart system methodically. Hours passed.  My winner wasn’t even in the shortlist. A list of the usual suspects was signed off. My guy later went on to win three national awards.  I quietly resigned from the committee.  So much for the gossip about decision making. Now the theory.

Robyn’s anecdote, and his instinctive antipathy to the officially mandated, matrix-based method, are certainly consonant with the themes of Wise Delinquency.  It is just a pity about the passing reference to Blink, a book whose beguiling readability concealed the simplistic and misleading nature of its main message.   My point  was not the Gladwellian idea that rigorous and painstaking methods should give way to gut feelings or intuitive “blinks”.   Rather, it was the certain kinds of rigorous and painstaking methods should not be applied by force in situations where qualitative deliberation is more appropriate.  And there are lots of such situations.

A recent article in Slate describes how one of the flagship examples in Gladwell’s book, the supposed predictive ability of marriage researcher John Gottman, only seems impressive because of egregiously lousy statistical methods.   These problems had been made public well before Blink was published.  I’m told that Gladwell had been informed of these problems, but apparently chose not to mention them; the truth of the matter would have interfered with a good story.   That seems like delinquency, but not of the wise kind.

Read Full Post »