Tie-breakers in figure skating: stupid, irrelevant, ineffectual, biased, and counterproductive

Free Web Hosting Provider - Web Hosting - E-commerce - High Speed Internet - Free Web Page

Dirk L. Schaeffer

Vancouver, Canada

In a recent posting to the Internet (“Why do we have judges in figure skating competitions?”), I presented the findings of a statistical analysis of three recent figure skating competitions. At several points in those analyses I came up against the problem of tie-breakers and, each time, they seemed to be doing more harm than good. This led to some general reflections on the use of tie-breakers in this sport, but since these seemed a sideline to my main purpose in the other posting, they are presented separately here.

Tie-breakers turn up in three places in the scoring of figure skating competitions. First, when each judge’s summed raw marks are transformed into ordinals, where the rule of assigning a higher rank to the score with the higher technical mark in the short program, performance mark in the long program is applied. Second, when the judges’ ordinals are converted into standings at that event, according to one-by-one (or, earlier, best-of-majority) rules, which resolve all – or virtually all - “ties” among skaters receiving the same total ordinal marks. And third, by the weighting (currently, two-thirds for the long program, one-third for the short, in the men’s, ladies’, and pairs’ competitions) that combines the results of individual events into a final score. While the last two of these may not immediately look like tie-breakers, that is, in fact, how they function, and at least one of the effects they have.

Clearly then, tie-breakers can play a large role in determining a skater final standing at any competition. Nevertheless, if we back away from the traditional view of basing our thinking about skaters’ abilities on the ordinals provided by the judges, and focus instead simply on the raw marks each skater receives, it becomes possible to look at tie-breakers simply as one choice we make among many potentially different ways of scoring these competitions, and to consider their effects simply as they are. The resulting picture is not pretty, suggesting that tie-breakers are for various reasons, stupid, irrelevant, ineffective, biased, and counterproductive. Let me deal with these one at a time.

1. Stupid.

Tie-breakers in figure skating consist of a set of often incomprehensible rules applied by anonymous accountants to the raw scores generated by the judges. I don’t know of any sport that has won widespread respect that uses such a cockamamie technique. Respectable sports either live with ties (as in timed and measured events) or add on extra competitive round (tenth innings, overtimes, etc.) to decide the issue.

In contrast, the system used in figure skating is about as sensible as it would be to decide the winner of a tied baseball game on the basis of which team scored the most home runs, or fewest errors.

Since there is not earthly reason why figure skating could not adopt either of the two saner procedures, it’s use of a nutty alternative, which confuses audiences, alienates sportswriters, and generally looks like something that could only have been dreamed up by a committee, not only seems stupid, but really is.

2. Irrelevant.

In the research cited above, I investigated the nine separate marks – technical, performance, and total – generated at three separate competitions, and in none of them could I find any instance of ties except at the level of the individual judge. That is, once you add up all the judges’ marks in any event, there is no longer any need for tie-breakers. While this example does not prove, of course, that ties will never occur, my best guess is that they will be quite rare. Thus, tie-breakers seem largely irrelevant: a cure for a problem that does not exist.

3. Ineffective.

In that research also, I began by noting that since there are no objective standards that allow us to determine that one skater “really” is better than another, we need to use judges. And since there are no objective standards to allow us to determine that one judge “really” is better than another, we have to treat all judges as equal. Consequently, the most meaningful criterion for any scoring system is that it produce the most agreement among all judges – rather than total agreement among the most judges , as the present systems do.

Within that framework, it turns out that tie-breaking rules not only do not increase agreement, they actually diminish it, by artificially creating differences were none existed, or converting very small differences into far larger ones. I’ll let one example suffice.

What started me on that investigation was the attempt to find out what “went wrong” in the scoring of the 2000 Men’s World Championships, as most clearly demonstrated in the case of one skater ranked sixth by one judge, fifteenth by another. Part of what went wrong, then, turned out to be the scoring system: when you looked just at the total scores assigned to this skater by these two judges, one gave him 11.0, the other 10.7. What was “wrong” then was far less the judge’s judgments than their difficulties in using the system in a close competition.

4. Biased.

The long (or “artistic”) program is weighted twice as heavily as the short (or “technical”) program in the determination of final standings. In the long program, the performance score is the tie-breaker, while in the short program the technical score is the tie-breaker. Thus, performance tie-breakers count twice as much as technical tie-breakers, which sort of contradicts the idea that both are weighted equally in determining the total scores on which the ordinals are based.

Some years ago, when compulsory figures were still used, these could compensate (or over-compensate, depending on your perspective) for this imbalance. Now all we’re left with is the bias.

5. Counter-productive.

Suppose for the moment that we had no tie-breakers in figure-skating, and then came upon the fairly rare situation where a tie actually occurred. What would we do about it?

Well, there seem to be two options. First, we could do nothing---just let it stand. For skaters below some arbitrary point in the distribution – say, fifth place – this seems like a perfectly adequate solution. Who, other than some of the agencies that dispense money or other favors based on the minutiae of standing differences, really cares? And it seems not just stupid, but downright addle-pated to design a scoring system solely to please a few bureaucrats.

But for skaters scoring fifth or higher, some kind of tie-breaker seems in order, and that leads to two new options.

First, some kind of accountant’s rule similar to what we have now. This will, if nothing else, ensure that skating never gets the kind of respect in the sports community that most skaters and fans would dearly welcome. But it would, of course, resolve the tie, fairly or arbitrarily.

Second, some kind of tie-breaking contest – re-skating the short program, or a third program now that these are mandated by the Four Championships’ rules – could be held. It’s more or less what everybody else does; there’s no reason it shouldn’t work for skaters. But that raises the problem of when to hold this tie-breaking contest, and here again there are two option.

First, it could be held at the conclusion of the event at which it occurred. What would this involve? Only going over-time on television. And as we all know, going overtime, actually pre-empting another program, is probably the best way there is of winning respect in sports and television studios. It would probably be the best thing to happen to skating since Tonya’s people took after Nancy.

Alternatively, it could be held at a later time, with the beginning of the Sunday “Parade of Champions” as the most logical choice. And what would that mean? One, two, or three days of free publicity as the media, who love nothing better than a shoot-out, would hype this for all it’s worth, knowing that will only result in more readers and viewers for their advertisers. It would probably be the best thing to happen to skating since Rudy became King of the Trailer Park.

But, of course, as long as we use accountants’ solutions those reasonable shots at fame and glory will continue to be denied us. Counter-productive, wouldn’t you say?