All the algorithms in the world can't predict who will succeed and fail in pro football.
For the last couple of months, scores of the best college football players in the country have been made to jump through hoops like trained animals and undergo psychological tests like the replicants in Bladerunner, all to determine whether or not they are suitable fodder for the National Football League.
Now the draft is under way, the first step in a process that will prove, over the next couple of seasons, that despite the most sophisticated and rigorous scouting and testing techniques, selecting players for the NFL is still pretty much what it has always been: a crap shoot.
Writing for The Atlantic in 2006, I asked a pro football scout:
Why, despite the enormous amounts of time and money devoted to it by NFL think tanks, the annual college draft produces such erratic results. Nobody, I noted, had been anxious to select Johnny Unitas or Bart Starr or Joe Montana—by my measure the three greatest quarterbacks of the last 50 years and winners of a combined 11 NFL championships. By NFL standards, they weren't very big and didn't have great arms. Unitas was the 102nd player drafted in 1955, Starr was the 200th in 1956, and Montana the 82nd in 1979.
The scout told me that even as recent as 1979 football scouting was "in the stone age." Scouting, he said, had become so sophisticated since Joe Montana was drafted that it was unlikely such a great player would go ever go that low in the draft again. A few weeks after our conversation, the New England Patriots drafted a relatively unheralded quarterback out of Michigan named Tom Brady. He was chosen in the 6th round, the 199th player selected overall.