Somehow Arnold Barnett manages to massage the subject of airline accidents into a breezy and sometimes comforting talk on statistical probabilities. In decades of research, he has taken firm hold of the metrics of measuring mortality in flight. While there are many ways of looking at the grim numbers, Barnett has developed his own preferred ratio, which looks at “death risk per randomly chosen flight.” Applying this approach, Barnett has come up with very reassuring statistics: The death risk per flight on first world domestic jet services, for the period of 1990-1999, was 1 in 13 million. To the air averse, Barnett offers that “a citizen is 2.5 times as likely to win the jackpot of the Massachusetts state lottery as to perish on his or her next flight.” For the four years between 2000 and 2004, there were zero accidental deaths in 70 million first world flights. Airline safety has tangibly improved, says Barnett. But security is another matter entirely: “We lost it all on a Tuesday in September,” he says. While we’ve “brought accidents to the brink of extinction”, we haven’t solved our problems “dealing with the forces of evil.” He strongly urges the reintroduction of positive passenger-baggage match, which he believes will deter terrorists who may use flawed explosive detection devices “as roulette wheels.”

As for fixing the Electoral College, which he likens to tilting at windmills, Barnett proposes applying a weighted average. This would “all but eliminate the worse consequences of the winner take all rule.” The biggest drawback? “People have difficulty with mathematical ideas. And this sounds complicated.”