During U.S. Nationals, while being interviewed by Peter Carruthers, David Kirby described his participation in the development of IJS. One aspect of IJS that he found appealing, he said, was that the system was developed to measure a skating performance, instead of rank the skaters, as is done under 6.0.

The score in IJS is intended to be an absolute measure of what the skaters have accomplished in their performances taking into account both the difficulty and quality of the performances. There are some who say this is a futile effort, since skating is both sport and art, and one cannot "measure" art. Perhaps that is true, perhaps not. But without venturing into that philosophical discussion, IJS has a more practical problem. The scores in IJS are simply not a measure of the absolute accomplishments of the skaters in their performances. Not even close. The IJS scoring method is riddled with arbitrary point values and mathematical inconsistencies that guarantee that the relation between scores and program quality/difficulty is limited at best, and calls into question the correctness of the exact placements calculated using IJS.

As the Winter Olympics approach, a common
concern of fans, and others, is that gold medals may end up being determined by
Technical Panel calls of dubious accuracy
-- perhaps due to marginal or inconsistent downgrades, levels or edge
calls. With results under IJS often determined by small fractions of a
points, *perfection* of judgment is required; something that human judgment
rarely achieves. But even if perfection of judgment was achieved in the
calls, the arbitrary nature of the points values still acts to frustrate the
validity of results. The edge call in flips and Lutzs is just one example
of this.

The IJS Scale of Values assigns a base value of 5.5 to a flip jump and 6.0 to a Lutz. This in itself is nonsense -- to think that the triple jumps are equally differentiated in difficulty by exactly 0.5 for each triple jump for toe loop through Lutz, or that a Lutz is only 9 percent more difficult that a flip. Base values of jumps simply do not reflect the true relative difficulty of the jumps and thus cannot provide a true measure of the skater's jumping accomplishments in a program. The errors in the base values for jumps are compounded further with the handling of edge calls for the flip and Lutz.

While, it makes no practical sense for the Technical Panel to agonize at great length over edge calls, since edge calls do not affect the base value of the jumps, they do make these calls. Once they make the edge calls, the judges will then go down 1 to 2 in GoE. The result of this approach is that point values for the flip and Lutz with edge calls do not accurately reflect what the skaters actually accomplished.

For an obvious change of edge in the Lutz, the skater will receive a base value of 6.0 and then typically lose up to two more points on the GoE, ending up with 4.0 points for the jump. But with a change of edge in the Lutz, what the skater actually accomplished was a perfectly good flip, and a correctly executed flip is worth 5.5 points according to the IJS Scale of Values. The point value for the jump, then, is not based on what the skater accomplished but on what they intended to accomplish. Hardly a correct approach for a system whose purpose is to measure what the skaters actually accomplish. It were as though in a basketball game a player attempts a three point shot, but releases inside the three point circle and is told, 'since you intended a three point shot, but completed a two point shot, we will give you one point!' Utter nonsense.

A Lutz with a change of edge should not get the full value for a Lutz, as it really isn't a Lutz. But it should also not get less points than the value of what it really is, a flip. If a 0.5 loss of points due to the change of edge is viewed as too little a reduction, that simply supports the view that the base values for the flip and Lutz are incorrectly set.

Consider now the lip. With an obvious change in edge on a flip, the skater currently gets a 5.5 base value for the flip, and then typically loses another two points in GoE, for a total of 3.5 points. But what the skater accomplished was the more difficult Lutz. Nevertheless, instead of getting the greater point value (6.0) for accomplishing something more difficult, the skater gets 2.5 points less for not accomplishing what they intended. Again, the points scored are arbitrary nonsense and not a measure of what the skater actually achieved, defeating the entire purpose of IJS measuring what the skaters accomplish.

From a pure absolute measurement accuracy point of view, if the intrinsic difficulty of flip is given one set base value, and the intrinsic difficulty of a Lutz is given another higher set base value, then a Lutz taking off from the wrong edge has the difficulty of a flip and should received the base value of a flip, with no loss of GoE if the jump has no other problems. A flip taking off from the wrong edge has the difficulty of a Lutz (in fact is a Lutz) and should receive the base value of a Lutz, with no loss of GoE if the jump has no other problems. A similar argument can be made for toe-Axels, where a true measure of what was accomplished would demand a base value of the corresponding Axel, and not the base value of a toe loop of one less rotation. A double toe-Axel, while less difficult that a true double toe loop, is certainly intrinsically more difficult than its typical 0.2 point score would indicate.

In terms of measuring the technical difficulty and quality of a skating performance, IJS remains deeply flawed mathematically and is too often penalty driven. When it comes to measuring the difficulty of elements, the scores in an absolute system must be about what is actually accomplished, even by accident, and not about intentions. Currently, however, IJS scores are little more than a nearly meaningless sum of arbitrary and inconsistently assigned point values, that have little to do with the true value of a skating performance in an absolute repeatable sense.

*Copyright 2010 by George S. Rossano*