Is figure skating becoming acrobatics on ice?

[rain]

No judge can ever give a number like 7.23 in IJS. Their only PCS options are 0.25 decimal increments -- either 7.0 or 7.25, for example. And for an element GOE, their only options are -5 -4 -3 -2 -1 0 1 2 3 4 5.

The finer decimal places that you see in the right column of the protocols are a result of averaging different judges scores for each component, and of the percentage of base value that determines how many (tenths or hundredths of) points each GOE score is worth for that element, which is then averaged across judges.

I know. My point stands. Parsing scores into quarter points and +5 to -5 is absurd, and it is almost never done properly, IMO, or course. That doesn't mean I disagree with all results under this system; I don't. I just think that there's a lot more intuitive ranking (and sometimes prejudging and shenanigans) going on than sticklers would like to believe. I don't blame the judges (except for the shenanigans), since I think it is actually impossible to watch a skating performance and run through every bullet point for every element AND every bullet point for every PCS score.

No one ever claimed that 5.5 meant anything specific under 6.0. It was always a placeholder. If anything, it might mean "world class but not quite medalworthy" in a very general sense -- which might turn out to be medalworthy at this particular event depending on the rest of the field.

With IJS scores, it's not really necessary to leave room for later skaters. If an earlier skater really does do many things just about as well as can humanly be done, there's no reason not to give them 4s and 5s for most of their GOEs and 9.75 and 10.0 for most of their components. If a later skater also does a similar number of things equally well or better, they can also earn the same kinds of GOEs and components, including the exact same marks for some elements and some components. It will be the one or two areas where the judge does see a difference in the later skater's favor that could make the difference, along with the base values that the judge has no control over.

Only straight 10.0s and straight +5s across the board would leave no room for a judge to reflect any difference. And even then the base values would determine the winner.

This is one of the strengths of IJS over 6.0. Neither system is perfect, and there are a lot of things I like about IJS. This is one of them. Also, one of the big weaknesses of ordinals was that, barring some really strange and unusual events, you had to be in the top 3 in the short program in order to even have a chance at winning. With simply adding scores together, that is no longer true, and we can have a champion who finished 7th in the short win the event. This is far more interesting, and IMO, a far cleaner gage of what's put out on the ice than what happened with ordinals. Especially since I've seen skaters 1-8 virtually tied after the short in terms of scoring. That means it's anybody's game in the long.

Had you and the other judges in this contest spent years evaluating similar contests with similar sections and similar marks? Over time, I'd expect you would develop your own sense of what numbers were appropriate for what level of quality, and in comparing notes you and the other judges would come to more agreement about just what each score represents.

It took some time for 6.0-trained judges to come to that understanding with the new IJS scoring when it first came in, but after a decade and a half I think there is closer to a consensus.

I've been doing this for years, trust me, I know what I'm looking at in my own field. Relative judging is way, way more natural a way to judge anything for us human beings.

 

[Mathman]

The numbers themselves were only meaningful in terms of each individual judge keeping track of how to rank each of the skaters against each other.

Evidently this sentence must be repeated many times.

Even back in the day when the 6.0 ordinal system was in force, people didn't really get it. Back then there was continual handwringing about "flip-flops." Skater A is ahead of skater B. Then skater C skates and now skater B is ahead of skater A. This seemed like some sort of paradox, or at least an undesirable feature of the scoring system.

But all it really meant was, don't award the ordinal until all the skaters have skated. Saying that "so far" skater A has the first place ordinal and skater B has the second place ordinal, with C yet to skate -- that is meaningless.

Actually, this was an undesirable feature, but not for the reason that fans usually griped about. It is undesirable from a sporting PR point of view. Audiences like to see running scores and do not want to be completely in the dark about who is winning or losing midway through the contest. Flip-flops cannot occur under IJS, and this is psychologically satisfying to sports fans -- plus you get to keep track of your favorite's "slay count"even if she is first to skate. :yes:

 

[Mathman]

Parsing scores into quarter points and +5 to -5 is absurd, and it is almost never done properly, IMO, or course....

Relative judging is way, way more natural a way to judge anything for us human beings.

This is confirmed by a huge number of controlled studies of the capacity of the human brain to make distinctions. By and large people can discriminate among about 7 different gradations, but not more.

For instance, suppose you show a test subject a "standard" measure of the length of a stick: This stick is an example of the shortest. This stick is an example of the next shortest. On so on. Then you take away the models and present the subject with a random stick taken from the pack. Into what category does it fall? If there are more than 7 categories the subject will not be able to decide most of the time.

I always thought it was cool that 6.0 figure skating had 7 categories: the 0s, the 1s, ets. up to 6.0. So it really was possible for a human judge consistently to get the first part right (this performance is in the 4s). Also when GOE was given from -3 to +3 -- yes! Seven categories, just what we humans can handle.

But comparing one by one, that's another story. If I give you 100 sticks and say, "line them up from shortest to longest," that's no problem even if the difference in length between any two is only a few millimeters.

GKelly's take above is great way to look at it. If you give a 5.7 that means (1) this performance falls into the 5s category, and (2) within the 5s category you have the digit after the decimal point to compare one skater to another.

Similarly under IJS, the judge can be pretty confident that this skater's Interpretation score should be in the 8s. Then you can use the .25 or the .50 to show that of the skaters in the 8s, one is a little better than the other.

(On another thread GKelly also made this interesting observation. Suppose you have two skaters who are both definitely in the 8s, but the difference between them is tiny. You can give one of them 8.5, 8.5 and 8.5 in the three "artistic" components, and give the other 8.5, 8.5 and 8.75." This allows the judge to say, these two skaters are in a virtual dead heat, but if I had to choose I would go ever so slightly for skater #2.)

 

[gkelly]

Evidently this sentence must be repeated many times.

Even back in the day when the 6.0 ordinal system was in force, people didn't really get it. Back then there was continual handwringing about "flip-flops." Skater A is ahead of skater B. Then skater C skates and now skater B is ahead of skater A. This seemed like some sort of paradox, or at least an undesirable feature of the scoring system.

But all it really meant was, don't award the ordinal until all the skaters have skated. Saying that "so far" skater A has the first place ordinal and skater B has the second place ordinal, with C yet to skate -- that is meaningless.

Actually it's a little more complicated than that.

I'm not going to get into the weeds explaining how 6.0 scoring worked.

I'll just say that the term "ordinals" referred to the rankings on each individual judge's card.

When referring to who was ranked ahead of whom in the current competition phase, or whole competition, according to the panel as a whole, one would refer to standings or placements or rankings, not "ordinals."

Any individual judge's ordinals themselves did not flipflop, but how the accounting algorithms combined the ordinals from all the judges could indeed lead to flips in the panel's consensus results for the current competition phase.

And in later phases (e.g., freeskate) of the competition the "factored placements" from the earlier phase (figures and/or short program) combined with the current phase could also lead to flips in overall standings for the combined event even if the results of each individual phase remained stable.

It's just that the word "ordinals" had a very specific meaning in skating scoring and it only confuses an already confusing topic further to misuse the word when referring to the interim (or final) standings/results/placements of the competition phase or combined event.

 

[Tavi...]

I haven't seen a rigorous study saying that tightrope walking is more difficult than sidewalk walking and I don't think I need one to make a conclusion.

For quad, you need speed, fast rotation and height of the jump. Basic physics that tell you that all this requires more power (do not mistake with energy* or work) in comparison with e.g. the step sequence due to time. There are activities that can cost relatively big amount of energy, but dispersed in time, while there are others, that require power. You need to release proper amount of energy in exact time to achieve height and rotation. And you can't correct it during the process.

The other thing of course is the danger. Again, physics. Of course sometimes you fall even from a stsq, but again, usually you fall with relatively low speed and "from the ground". When you fall from a jump, you fall from height, you fall fast and the kinetic energy from the fall is much higher. To avoid this, you need coordination, you have to train it practically all the time of your career. Every jump means you have to lift all your weight to the air, to "defy gravity" , but not because of the magic broomstick, but through your own body. In walking the powers are balanced, but in jump you have to multiple the strength with which you normally execute your steps.

Empirically, the frequency of failure with a jump is much higher than with stsq or with a spin. More coordination (=work of the whole organism, brain, muscles, eyes) is needed, learning is longer. Every average skater's spins are sometimes evaluated with lvl 4, while quad is just for the elite who takes all the risks and effort to achieve it.

So yes, jumps are basically harder. than any other element, no matter that so many people don't like them, because spins are somehow more artistic

As for the GOE, there are pretty detailed rules what deserves which grade, both positive and negative, height, entry, exit, UR, fall. On the other hand, the rules about componets are pretty vague in comparison and while you can disagree with the GOEs given by the judges, I have often bigger problem with components, when they are often high even with tehnically weak programs (with errors).

-
*For instance, the last bar of chocolate you've just eaten contains more energy than the same amount of TNT. BUT, the energy from the chocolate is released very slowly in your digestive system, while the energy from explosive comes out in spit of a second.

Thanks for your response. I appreciate the zealous advocacy in support of your position, but it’s not really evidence.

Seriously, I think it’s interesting that you’re comparing non-jump elements to walking on the sidewalk, especially combined with your personal definition of “difficult” elements as those requiring “power” and exposing athletes to “danger.” I also think you’re relying on a lot of unproven assumptions - eg, how do you know it takes longer to learn skill x than skill y? Since I need to get to work, all I have time to say is say is that FS isn’t weightlifting. I’d also ask why - if hard equals power/danger - pairs lifts, which require the combined power of two skaters and are probably more dangerous than quads, are worth fewer points. As a recreational skater, I’d also say that regardless of whether you’re elite or not, it’s pretty easy to fall when doing things like crossovers and turns; depending on how you fall, the falls can be quite painful and result in injuries such as concussion and broken bones, too.

@mathman, I think the problem with anecdotal evidence is that not every skater in the world says jump x is harder than jump y. We’ve all heard of skaters for whom a 4T is easier than a 3A. Added to which, for someone like Jason - for whom a quad is clearly harder than a 3A - we don’t know why (all in his head? Body type? Fact that he was initially trained in a delayed rotation technique? All of the above?) I think a first step to analyzing what elements are harder than others is to define what difficulty means. @flanker has one definition, but I doubt it’s universal. Once you have a definition, I’m sure that someone could come up with ways to test for it.

Sorry, gotta go, and this is probably a bit OT for the thread anyway.

 

[Mathman]

Actually it's a little more complicated than that....

I can't seem to break myself of the habit of regartding numbers like "first, "second," "third" as "ordinals."

Anyway, the whole point was, if you announce partial results before all the (ranked) votes for all the candidates are counted, then you might get a surprise at the end.

One thing I never understood, though, is why people were concerned about "flip-flops" due to factored placements. This just means that the skaters who performed the best in the short program messed up in the long, and vice versa. This can happen under any scoring system.

 

[rain]

Thanks for your response. I appreciate the zealous advocacy in support of your position, but it’s not really evidence.

Seriously, I think it’s interesting that you’re comparing non-jump elements to walking on the sidewalk, especially combined with your personal definition of “difficult” elements as those requiring “power” and exposing athletes to “danger.” I also think you’re relying on a lot of unproven assumptions - eg, how do you know it takes longer to learn skill x than skill y? Since I need to get to work, all I have time to say is say is that FS isn’t weightlifting. I’d also ask why - if hard equals power/danger - pairs lifts, which require the combined power of two skaters and are probably more dangerous than quads, are worth fewer points. As a recreational skater, I’d also say that regardless of whether you’re elite or not, it’s pretty easy to fall when doing things like crossovers and turns; depending on how you fall, the falls can be quite painful and result in injuries such as concussion and broken bones, too.

@mathman, I think the problem with anecdotal evidence is that not every skater in the world says jump x is harder than jump y. We’ve all heard of skaters for whom a 4T is easier than a 3A. Added to which, for someone like Jason - for whom a quad is clearly harder than a 3A - we don’t know why (all in his head? Body type? Fact that he was initially trained in a delayed rotation technique? All of the above?) I think a first step to analyzing what elements are harder than others is to define what difficulty means. @flanker has one definition, but I doubt it’s universal. Once you have a definition, I’m sure that someone could come up with ways to test for it.

Sorry, gotta go, and this is probably a bit OT for the thread anyway.

Just to add, if exceptional skating skills were so easy to learn, we'd have a whole field full of Patrick Chans. It's telling that we don't.

 

[gkelly]

I can't seem to break myself of the habit of calling numbers like "first, "second," "third" as "ordinals."

You are a Man of Math and therefore probably use that word in other contexts in a more general sense a lot more than most skating fans or skating officials.

Anyway, the whole point was, if you announce partial results before all the (ranked) votes for all the candidates are counted, then you might get a surprise at the end.

Absolutely.
And of course it didn't help that announcements of interim standings during freeskates often give only the scores for the freeskate but the "currently in X place" statement for the competition as a whole, without specifying where the skater is currently ranked in the freeskate. Which can be confusing to audiences, under any system, when a skater with a significant lead from the short skates poorly, places behind better performances in the freeskate, but still leads overall.

One thing I never understood, though, is why people were concerned about "flip-flops" due to factored placements. This just means that the skaters who performed the best in the short program messed up in the long, and vice versa. This can happen under any scoring system.

It's certainly possible under any system that skaters who perform best in the short and mess up in the long can end up ahead of skaters who beat them in the long. Which, as I said in the previous paragraph, can be confusing if the verbal announcements and obvious visual graphics do not communicate the freeskate-only placements to the audience.

But flipflops between skaters who had already finished skating earlier and had been ranked relative to each other in a different order can only happen with factored placements or something similar.

Flipflops can't happen if absolute scores from each competition phase carry over and are added together.

With IJS scoring for figure skating, if skater A has a higher total of SP+FS scores than skater B, there is no way that B can move ahead of A thanks to whatever happens with skater C. B might move ahead of A overall immediately after B finishes the freeskate, and then B will remain ahead of A for the rest of the competition no matter what happens with the other skaters or how many skaters place between them or ahead of both. Or B might be stuck behind A because A had a larger point lead in the SP than B has in the FS -- and that lead will never change based on what anyone else does later.

(With factored placements, A remaining ahead of B after B skates but B later moving into the lead because of other skaters getting between them in the freeskate did happen fairly frequently, the simplest and often predictable type of flipflop.)

With speedskating and other racing sports that take the time from yesterday's race and add it to the time from today's race, those total scores are also static and the relative placements of A and B will never change 10 or 60 minutes later because of what C or Q do.

If a racing sport converted the actual times from yesterday's and today's races into placements and then used only the placements rather than actual times to determine overall results for the combined event, that would allow for flipflops in the relative positions of earlier competitors.

Are there any timed sports that combine results in that way?

 

[el henry]

Tbh, it's little bit disrespectful to Ted, even though it's said with honey. In fact, you said that Ted is just a "goody uncle" and his words and opinions are worthless. I recommend that you watch all of his interviews. But if you don't have time here he talks about his principles of commenting. "Always be honest."


....

Well, I'm going to go off topic here, and say it is not at all disrespectful to Ted. Any fool can criticize, carp and complain, just ask me Sarcasm and such? Pffft, no particular skill there.

What Ted does: finding the best in every single skater, explaining it, describing it, respecting every single skater who takes the ice, whether a world champion or a first timer from a small fed, that is *hard* work. And I in turn respect Ted beyond all measure for that commentary.:thumbsup:

That said, Ted's interviews are fluff, whether they are of World Champions or first timers, and I take them for what they are. Hard hitting interviews of substance is not what I'll get from Ted, and that's OK, I can get that elsewhere

 

[drivingmissdaisy]

But flipflops between skaters who had already finished skating earlier and had been ranked relative to each other in a different order can only happen with factored placements or something similar.

This was sort of the weird problem, not the final placements changing based on the SP rankings.

Say you have these LP ordinals with one skater to go:
1 1 1 1 2 2 2 2 4
2 2 2 2 1 1 1 1 5
Skater 1 is winning the LP, because you have to look at the final judge's placement to break the tie.
Then the last skater competes and judges 5, 6, 7, and 8 place that skater in between the leaders:
1 1 1 1 3 3 3 3 4
2 2 2 2 1 1 1 1 5
Now skater 2 wins the LP because of the multiple 2 ordinals, of which skater 1 has none (although it doesn't change that more judges had skater 1 ahead)

 

[dorispulaski]

It is a fascinating discussion, but as a person who is primarily a dance and pairs fan, who has a few favs amonst the group of single skaters, I don't have much to say about jumps.

But there are two or three points I would like to make: whatever ranking system is used, the judges are never, at least in international competitions, seeing a skater for the first time when they are judging them. The Judges, now and forever, go to the practices. (And there are youtubes). By the the time the first competitor takes the ice, a judge has a score range in mind for this skater. People have complained about judges going to practices for years, but I think in general, it is a good thing.

Second: knowing what thing is harder than another thing of a similar sort. It is hard to do. IJS does this badly, particularly in dance, and particularly for dance lifts. The hardest lift I have ever seen done is probably Meryl & Charlie's opening lift in Scherazade. Why do I think that?

It is six years later and I have never seen anyone else even try it. It took Meryl and Charlie two years to develop it. So it is an impressive skill that is too hard to learn and put in a program.

Yet it got essentially the same scores as Tessa and Scott's lifts, all of which I have seen other teams perform

It is common to hear a commentator say Chock & Bates have the best lifts in a competition-Sinead Kerr said last year that if others were getting +/-5, Maddy and Evan should be getting +/-10. But they got what other top teams got for lifts on average. This was because once you get to level 4, there is no way to score a lift as harder than that other than with GOE, which is really about how you did it, not about how hard it is.

In pairs, the best throw 3A I ever saw was the first one done in the Olympics, by Inoue & Baldwin. It got as low as -1 from one judge. And it got no bonus for being hard, because mechanically, a throw Axel is hard, because the partners tend to get in each other's way. Consequently, the risk for doing a throw 3A is too great for the amount of points you might get for one.

I suspect 6.0 would deal with that sort of thing better than having pre-set points for things no one has yet done

I have no problems with acrobatics, since I like pairs.

But I also like a stirring step sequence with deep edges, and surprising changes of direction and flair. and a spin with speed changes and arm and leg movements that hit the beat and express the music. However, I do not expect IJS to pander to my tastes. I just wish that when a top skater does a thing badly that the judges do not throw reputation points at him in GOE.

This year there have been overscored shaky lifts that pluck my last nerve, and truly ugly 3(or4) Euler 3 sequences given +3.

If the judges actually used the system they have, I would be better pleased with them.

 

[Mathman]

Say you have these LP ordinals with one skater to go:
1 1 1 1 2 2 2 2 4
2 2 2 2 1 1 1 1 5
Skater 1 is winning the LP, because you have to look at the final judge's placement to break the tie.
Then the last skater competes and judges 5, 6, 7, and 8 place that skater in between the leaders:
1 1 1 1 3 3 3 3 4
2 2 2 2 1 1 1 1 5
Now skater 2 wins the LP because of the multiple 2 ordinals, of which skater 1 has none (although it doesn't change that more judges had skater 1 ahead)

A perfect example of violation of principle of independence of irrelevant alternatives. (The placement of the last skater should be irrelevant to whether skater A or skater B was better than the other. But the outcome (Skater B beat skater A) was not "independent" of this "irrelevant" datum.

All ranked voting systems with more than three candidates are subject to this problem. Mathematically, the solution is simple: don't announce partial results until the last skater is done. (Not good theater, though. )

 

[Orlov]

Well, I'm going to go off topic here, and say it is not at all disrespectful to Ted. Any fool can criticize, carp and complain, just ask me Sarcasm and such? Pffft, no particular skill there.

What Ted does: finding the best in every single skater, explaining it, describing it, respecting every single skater who takes the ice, whether a world champion or a first timer from a small fed, that is *hard* work. And I in turn respect Ted beyond all measure for that commentary.:thumbsup:

That said, Ted's interviews are fluff, whether they are of World Champions or first timers, and I take them for what they are. Hard hitting interviews of substance is not what I'll get from Ted, and that's OK, I can get that elsewhere

That's just Ted, unlike you, gave a reasonable opinion. You get off with the general phrases "about research". Earlier, was quite rightly pointed out that no one did any "research" and about the triples, too. I told you obvious thing - that no quad can ever compare to THIS in terms of its effect on the body. This has been practiced for many decades an we don't see any disturbing news from pair skating. And I don't see that you care. But when it comes to your favorite "revolutions in the air" at single skating - ooh, this is completely different ("after all, there are my favorite Andrew T and Jason Brown, and they are not so good with revolutions in the air, so research, and only research!")


Ted is constructive, you are not. He has an argument, you don't. Unlike you, Ted is a man from the world of figure skating and a former coach. All you have to do is say "well, this is Barton, he always talks like sugar syrup". Okey

 

[Mathman]

THIS

Thank you for posting that. This is a beautiful sport.

As for the rest, though -- we are here to talk about skating, not each other.

 

[Orlov]

As for the rest, though -- we are here to talk about skating, not each other.

You're right. In my defense, I can say that I was hurt by this comment about Ted. All I heard that el henry said "Ted Barton is just a sweet talker, his opinion doesn't matter. He just says nice things to these Russians in an interview." After watching this half-hour interview, I learned a lot about him. This is a sincere and passionate man. He is no doubt sincere. And this is the opinion of a notable person from the world of FS, and not a forum chatterer like all of us. And he brought some arguments, so show respect, damn it, counter-argue them! Not just "well, he's sweet-mouthed"

 

[drivingmissdaisy]

A perfect example of violation of principle of independence of irrelevant alternatives. (The placement of the last skater should be irrelevant to whether skater A or skater B was better than the other. But the outcome (Skater B beat skater A) was not "independent" of this "irrelevant" datum.

All ranked voting systems with more than three candidates are subject to this problem. Mathematically, the solution is simple: don't announce partial results until the last skater is done. (Not good theater, though. )

Something like that could have happened in SLC in the ladies LP. The ordinals for the top 4 were all over the place; even Hughes had a few 4th place rankings. (https://www.sports-reference.com/olympics/winter/2002/FSK/womens-singles-free-skating.html)

 

[Mathman]

Something like that could have happened in SLC in the ladies LP. The ordinals for the top 4 were all over the place; even Hughes had a few 4th place rankings. (https://www.sports-reference.com/olympics/winter/2002/FSK/womens-singles-free-skating.html)

:yes: This in fact was a double whammy for someone trying to understand figure skating judging.

When the dust settled, Sarah had 5 first place ordinals. She was the winner of the LP under any system, no matter what. However, because of factored placements, Michelle did not have to beat Sarah to win overall, she only had to finish second ahead of Irina (assuming Sasha Cohen didn't jump up and win the LP).

Removing Sarah and counting 1st and 2nd place ordinals, Irina had 5 and Michelle had 4, so Irina got 2nd overall and Michelle dropped to third (in the LP and overall. Michelle had been in first place overall, ahead of Sarah, because of factored placements, before Irina skated. Irina stole some 2nd place ordinals from Michelle, just as in your example above, and this flip-flopped Sarah and Michelle at the top, in terms of overall placement, although not in the LP.)

The kicker was that one of the judges that favored Irina over Michelle for second was the U.S. judge, Joe Inman. If he had changed his ranking to Sarah, Michelle, Irina, instead of Sarah, Irina, Michelle, then Michelle would have been the Olympc gold medalist.

But no! That's with the majority of ordinals system. This system was still in use in the U.S., but internationally the ISU had switched to the OBO (one-by-one) system. By OBO, Irina beat Michelle on 6 judges' cards, to 3 for Michelle. So Inman was off the hook -- even if he had changed his vote it wouldn't have changed the final result after all.

None of this cool stuff can happen now. It's just the old blah, whoever gets the most points wins. What's the fun of that?

 

[denise3lz]

IIRC, I saw a video clip including Gillis Grafström's wife calling "acrobatic skating" and it was probably Boitano's era.

 

[[email protected]]

This happened in American football for instance.

American football guys are dressed like astronauts. I wonder about concussion concerns in rugby where they do not wear helmets or boxing where concussion is the name of the game.

 

[el henry]

American football guys are dressed like astronauts. I wonder about concussion concerns in rugby where they do not wear helmets or boxing where concussion is the name of the game.

They are only dressed like astronauts in recent times, when, as @Mathman stated, folks started looking at 50 year old former football players with CTE and thinking, hmmmm, maybe winning a Super Bowl isn’t worth your health? Now, thank God, the concussion protocol is very serious, is enforced. And star players are taken out of a game if they don’t pass the metrics, with no hoo hah about who’s a tough guy.

I know nothing about rugby. I believe boxing has gotten more serious (for example, helmets) but any sport could always get more serious:thumbsup:

Sadly I fear figure skating will never be popular enough for this type of long term follow-up.