by George Rossano
(31 March 2013) Following the recent World Championships, skating blogger Monica Friedlander launched a petition drive on Facebook to have the ISU award Denis Ten a gold medal for the Men's event for the following reason:
"In light of the unjust results in the men's competition at the 2013 World Figure Skating Championships, we request that a second gold medal be awarded to the true winner of the competition, Denis Ten, the way a second gold medal was awarded at the 2002 Olympics to the pair team of Jamie Sale and David Peletier."
As we begin writing this commentary some two weeks after the close of the competition, 765 people have thus far signed the petition.
Fundamentally, the second medal in 2002 was a response to the fact the result then was suspect due to the clear likelihood that vote trading had taken place. The results at Worlds, however, were the result of field of play decisions that some fans/observers do not agree with. In my view, litigating field of play decisions and changing results after the fact is a road down which no sport should ever go. So as much as I love to tweak the noses of the skating establishment, I'm afraid I cannot go along with this one.
Working rink side at the competition, as I have for the last 27 World Championships, there was no doubt in my mind that Patrick Chan won the Short Program. There was also no doubt in my mind that Ten won the Free Skate. But I would have been hard pressed then or now, to combine the two programs in my mind (with the Short Program two days earlier) and decide which combination of the two programs together was best -- and I strenuously doubt anyone else could accomplish those mental gymnastics either, then or now.
For me, the results in the Men's event and the resulting controversy, illustrate two aspects (defects?) in the scoring system that have been present since its inception. The first of these is the method of combining the results of the short and long programs.
In the former system, the results of the short and long where combined through total factored place, with the place in the free skate having twice the weight (importance) of the short program. This put the main weight of the total result on the more important of the two programs, but made it difficult (rare) for a skater with a poor showing in the short to win the event, even if they won the long by a substantial margin. This came to be viewed as a defect, and the new system was designed to allow a skater with any place in the short a better chance of still winning the event. The method chosen to do this however had unintended consequences.
In the current system, the points from the short and the long are added to determine the final result, and the number of points available in the long is about twice those available in the short.
By combining points instead of places, skaters have a much greater opportunity to move up in the standings from the short to the long, but the 1:2 weight of the two programs was lost. As a result, skaters who win the short and not the long win events about half the time. The reason for this follows.
The free skate has twice the point value of the short program, but the points in the short and long now have equally weight, whereas before the places in the two programs did not have equal weight. A skater trailing by two points in the short program currently needs to lead by two points in the long to catch up. But for the free skate to really have twice the weight of the short program, the skater in this example would only need a one point lead in the long to catch up.
My suspicion is that people who think Ten should have won the whole event were, and are, reacting to the old concept that Ten won the more important of the two segments, and thus should have been the champion. But that concept no longer exists in the current system. If one wants to bring that concept back into the scoring system there are several things one could do.
First, one could return to using total factored place to combine the results of the two segments. This, however, would eliminate the mobility to move up after the short program that people prefer.
Second, one could apply a factor of 0.5 to the short program score to make the points in the short program half the weight of the points in the free skate. For every point a skater trails in the short program they would need only 1/2 point in the free skate to make up the difference. This would truly give the free skate twice the weight of the short program and still allow mobility to move up after the short.
Third, one could eliminate the short program entirely. Problem solved. Instead of having a best one-out-of-two playoff-like system we would have a one game super bowl, winner takes all.
The short program no longer serves the purpose it did in competition as when it was first introduced, so why keep it? Particularly when the short and long are now pretty much the same animal, with just different time limits. If one is to have two programs to determine the winner, it would be more interesting, in my opinion, to have a technical-only program and an artistic-only program to make up an event.
The second troublesome aspect of the current system that is highlighted by the Men's result, is the lack of precision in determining what a program is really worth in absolute value, and thus what the order of finish should be.
Chan won overall with a 1.3 point lead, but as we have written about on many prior occasions, the mathematical accuracy of the system is only about 3 points, at best. Between the uncertainty for what the elements should really be worth, the unknown consistency for calling levels, the limitations to accurately assigning levels and choosing features, the random errors of judgement of the judges, the uncertain method and questionable consistency of the judges in assigning component scores, the differences of opinion over the proper penalty for falls and under-rotations, and on and on, any point difference between skaters of less than 3 points is statistically meaningless. With a 1.3 point margin of victory the system has only a limited certainty who really won the Men's event, though it decides who to give the gold medal to regardless.
For example, the margin of victory for Chan was a difference of 0.25 points in only three of the five Program Components. If Chan was over-scored by this small amount, as some assert, or if Ten was underscored by this small amount, the result is reversed. (Though to be fair, if Ten had not doubled his flip in element seven, he would have won, and no one would be having this discussion either.)
Fixing the statistical accuracy of the system will be more difficult than fixing how the segment scores are combined, but is probably fixable if a serious effort was made to address it.
As for Ten getting a gold medal, he got one for winning the free skate, and rightly so.
Copyright 2013 by George S. Rossano