Jonah Lehrer is still arguing that the Dallas Mavericks beat the Miami Heat by ignoring statistics and turning loose the intangible powers of J.J. Barea. Or at least that stat-heads can't prove that wasn't the case:
Maybe the Mavs are relying on statistical models that exceed what Beech put together at 82games.com, and that these pro models (unlike every model used by us rank amateurs) statistically demonstrated the value of starting Barea despite his mediocre numbers. But that strikes me as a bit of a black box argument. If you believe in stats, it's not enough to say that Barea must have been the statistically-minded call, when we have no idea what those stats might be.
Huh. Who might be able to illuminate the black box surrounding the Mavericks' lineup decisions? We e-mailed Mark Cuban: Was any statistical analysis employed by Carlisle to start Barea?
Cuban writes (punctuation from original):
you use everything.
the stats they used in the article dont show match ups by lineup against lineup. IN a playoff series, while the information isnt perfect, the longer the series, the better the information. So we knew which lineups worked against which of their lineups. We also then had to build in substitution patterns and expected responses to change. Things that Coach Carlisle is great at. Its a chess match.
Plus, Using stats arent always about putting your best combination out there, but sometimes about trying to get them to put their worst combination of players on the floor and then getting them to do things they are least comfortable with.
net, net, as I always have said, the one thing that stats, no matter which stats you use, dont take into account is coaching. Making the right decisions with the stats is more important than knowing what the numbers are .
m
[Note: We believe that's stats expert Roland Beech—who sits on the Mavs bench during games—above, at far right, celebrating the title.]