By JOHN JEANSONNE
Before Vancouver's Olympic figure skating is over, the viewing public - and maybe even the defeated men's gold-medal favorite, Yevgeni Plushenko - might want to turn to economists and statisticians to help explain the sport's judging scheme. What counts most? How is it counted? Who's counting?
In fact, there are two such professionals - Dartmouth economics professor Eric Zitzewitz and Yale statistics professor Jay Emerson - already years into analyses of the sport's scoring methodology that has been rendered even more mysterious and secretive since the 2002 Olympic judging scandal. And the consensus, though Zitzewitz has concentrated on bias and Emerson on what he calls "the phantom judge," is that there remain some major bugs in the system.
Certainly Plushenko, the 2006 Olympic champ from Russia, was questioning how the numbers had been crunched after American Evan Lysacek was awarded the gold Thursday night. Plushenko claimed that "it raises questions" when one skater performs a quadruple jump (Plushenko) while the other settles for a triple (Lysacek) but the difference isn't reflected in the result.
Was there an influence of politics? Personal preference? Judging incompetence? Some form of skulduggery?
Zitzewitz began studying the process because of the 2002 Salt Lake City incident, in which French judge Marie Reine Le Gougne was fingered for favoring the Russians over the Canadians in pairs - in exchange for having a Russian judge vote for the Canadian team in ice dance.
The uproar caused the International Skating Union to dismantle a century-old scoring code, producing in its place a doozy of a math game in which skaters top out with scores about 200. Worse, the ISU's new system introduced the kind of anonymity and randomness that makes fellows like Zitzewitz and Emerson crazy.
In a 2002 paper, Zitzewitz found "lots of evidence" of bias in the old 6.0 system, when judges' names still were attached to their scores. Now, all nine judges on a panel score anonymously, and two of those judges' scores are eliminated randomly by a computer.
"The argument was that [anonymity] would make it hard to do vote-trading deals," Zitzewitz said, "because neither party would know if the partner had followed through on the deal. But the audience has more trouble looking for bias, and I have found bias has gotten a little bit worse than in 2002."
He argued that anonymous judging not only makes his own investigative chore more difficult but also has eliminated the ability of judges to monitor their peers, "because now they can't know who gave which score."
Emerson's calculations, though they aren't meant to look for bias, nevertheless take the ISU to task. "I'm just saying they have this screwy system with nine judges, presumably nine experts who all think they're scoring the competition, so they should use all nine," he said. "And, at least, if they don't, they should acknowledge that the random selection of a subset [two of nine judges not counted] introduces noise to the process."
In examining last [week's Olympic] pairs competition, Emerson concluded that, had all scores been processed, the fourth- and fifth-place positions would have been reversed, so the same potential flip-flop exists for gold and silver medalists.
"In a close competition," he said, "the skaters should feel it is the judges' evaluation that awarded medals, not the possibility of a sub-panel made by a computer in a back room." The phantom judge.
Neither man claims to be a skating expert (though Zitzewitz said his 5-year-old daughter is showing interest in the sport). And both acknowledge that any judging done is, by definition, subjective.
"The best you can do," Zitzewitz said, "is try to come up with a system that gets the obvious biases out" - which is why he recommends the ski-jumping model, in which the international federation picks its judges. (National federations choose their own for skating's judging pool.)
"People say the new system has turned skating into a jumping competition, with less artistry," Zitzewitz said. "[The ISU] wanted to make it more objective, and I don't have a strong view whether that's right or wrong. But that was the trade-off."
For Emerson, "there is a whole side to this that I'm unqualified to comment on - artistic freedom and so on. I'm not even sure I know enough about it to say which [skater] is better. You don't want to ask a statistician that question."
Actually, sometimes you do.
Back to Jay's Homepage