The COMPLETE Guide to Fixing the Scoring System and Improving Ice Skating | Page 2 | Golden Skate

The COMPLETE Guide to Fixing the Scoring System and Improving Ice Skating

Blades of Passion

Skating is Art, if you let it be
Record Breaker
Joined
Sep 14, 2008
Country
France
Could you try to calculate the TES of some famous performances basing on your rules (I'm not talking about the changes in the structure of the progams, obviously) and see how they change?

Yuzuru Hanyu's TES in the LP at 2014 Olympics would be scored as approximately 77.33 rather than 87.66 (including fall deductions). Assuming good judging his PCS would obviously be significantly lower than the 90.98 he received as well. We'd be looking at about 156 for that performance rather than 178.

For a comparison between performances, at 2014 Worlds Hanyu's TES would increase to approximately 91.31 - a difference of 13.98 points for a clean performance as compared to a performance with two falls. In this particular instance it's not a huge differential between what happened in the current system (his TES increased by 12.27), although that's because under the current system Hanyu's 3S in sequence at the Olympics received no credit at all.

The major difference is that under the current system Hanyu's 4S fall at the Olympics was worth 6.5 points. Under this system it would be worth 3.4 -- a very big difference in loss of points for falling on that element. His 3F fall under this system would also only be worth .9 points, rather than 2.4 points.
 

Meoima

Match Penalty
Joined
Feb 13, 2014
I still don't understand ISU's reason for rewarding a fall with so many points. It's like they are encouraging the skaters jump as much as they can. :scowl:

What about stumbling and put your hand on the ground or 2 foot? How many deduction it would be?
 

HanDomi

Record Breaker
Joined
Feb 27, 2014
Yuzuru Hanyu's TES in the LP at 2014 Olympics would be scored as approximately 77.33 rather than 87.66 (including fall deductions). Assuming good judging his PCS would obviously be significantly lower than the 90.98 he received as well. We'd be looking at about 156 for that performance rather than 178.

For a comparison between performances, at 2014 Worlds Hanyu's TES would increase to approximately 91.31 - a difference of 13.98 points for a clean performance as compared to a performance with two falls. In this particular instance it's not a huge differential between what happened in the current system (his TES increased by 12.27), although that's because under the current system Hanyu's 3S in sequence at the Olympics received no credit at all.

The major difference is that under the current system Hanyu's 4S fall at the Olympics was worth 6.5 points. Under this system it would be worth 3.4 -- a very big difference in loss of points for falling on that element. His 3F fall under this system would also only be worth .9 points, rather than 2.4 points.


Can you show GOEs that you counted for jumps at Worlds ?
 

Alain

Match Penalty
Joined
Apr 28, 2014
If it were valid in Sochi, the Olympic champions would be Volosozhar/Trankov, Sotnikova, Ten.
 

Alba

Record Breaker
Joined
Feb 26, 2014
The problem is, PCS guidelines is so vague and subjective. Personally I think PCS, must be like TES, is judged live and by the quality of the performance at that very moment. Skaters have their on and off days so their PCS couldn't stay the same for every competition or just raise gradually despite the quality. A skater might get 70 PCS at skate america because, then 90 PCS in GPF, then 80 PCS at worlds. Personally I think it's possible and more reasonable than constant 90s PCS in every competition.

What do you mean PCS must be like TES?
 

Meoima

Match Penalty
Joined
Feb 13, 2014
What do you mean PCS must be like TES?
But up and down PCS sounds more reasonable to me, rather than 90s PCS constant for top skaters all the time. Skaters have their day on and off. If even on their day off, their PCS are still over the top, it's unfair isn't it?
Of course top skater should have more consistency than the rest, but that doesn't mean they don't have their day off.
 

Alba

Record Breaker
Joined
Feb 26, 2014
But up and down PCS sounds more reasonable to me, rather than 90s PCS constant for top skaters all the time. Skaters have their day on and off. If even on their day off, their PCS are still over the top, it's unfair isn't it?
Of course top skater should have more consistency than the rest, but that doesn't mean they don't have their day off.

Yes of course but I don't understand what do you mean by PCS should be like TES. In the end they are up and down, depending on the performance. Usually (not always, see Kostner) when TES is good PCS are as well and vice versa.
 

Meoima

Match Penalty
Joined
Feb 13, 2014
Yes of course but I don't understand what do you mean by PCS should be like TES. In the end they are up and down, depending on the performance. Usually (not always, see Kostner) when TES is good PCS are as well and vice versa.
Ah, I mean there should be more details guidelines about PCS just like TES, because at the moment, it's so vague and subjective. PCSs raise up then fall down without any reasonable explaination at all.
 

skatedreamer

Medalist
Joined
Feb 18, 2014
Country
United-States
I still don't understand ISU's reason for rewarding a fall with so many points. It's like they are encouraging the skaters jump as much as they can. :scowl:

What about stumbling and put your hand on the ground or 2 foot? How many deduction it would be?

For a non-skater, especially the most casual fans who don't pay much attention to CoP (if any), the apparent lack of penalties for falls is the most frustrating thing. To them, a fall = fail, pure and simple.

I do understand and appreciate the idea of trying to reward a skater for attempting and at least partially completing a difficult jump, but still get annoyed that falls aren't penalized more heavily. At some point, IMO it gets to be kind of like some kids' events where everyone who competes gets a ribbon or something just for showing up. As in, what's the value of a prize for coming in 10th out of 15? Just trying to represent the viewpoint of the "average" fan. :biggrin:

On another matter, this somewhat-above-average fan :laugh: would like to nominate BoP for ISU President.
 

Alba

Record Breaker
Joined
Feb 26, 2014
Ah, I mean there should be more details guidelines about PCS just like TES, because at the moment, it's so vague and subjective. PCSs raise up then fall down without any reasonable explaination at all.

Ah ok. I don't think they can be like TES though. In the end there are some subjective matters involved like intepretation and also choreo, choice of music etc.
 

Meoima

Match Penalty
Joined
Feb 13, 2014
For a non-skater, especially the most casual fans who don't pay much attention to CoP (if any), the apparent lack of penalties for falls is the most frustrating thing. To them, a fall = fail, pure and simple.

I do understand and appreciate the idea of trying to reward a skater for attempting and at least partially completing a difficult jump, but still get annoyed that falls aren't penalized more heavily. At some point, IMO it gets to be kind of like some kids' events where everyone who competes gets a ribbon or something just for showing up. As in, what's the value of a prize for coming in 10th out of 15? Just trying to represent the viewpoint of the "average" fan. :biggrin:

On another matter, this somewhat-above-average fan :laugh: would like to nominate BoP for ISU President.
The funny thing is, the new edge calls rules are somewhat too harsh comparing to the deduction for the falls. People can have wrong edge but stay on their feet, to me it looks much better than take off on the right edge but fall on your butt. :scowl:
 

gkelly

Record Breaker
Joined
Jul 26, 2003
Falls

The penalties for falls and other obvious errors could be larger and/or more explicit. BoP has suggested some possibilities, and we've discussed other potential options in previous threads.

Suppose the result of a fall on an element is 0 points for the whole element, regardless of how difficult the element was or how much of the element was completed. Suppose even an otherwise great level 4 spin or step sequence, or quad-triple-triple combination, with a fall at the very end would earn 0 points.

Falls on elements would then end up being more costly the more difficult the element.

So a skater who tries something really hard and falls in the attempt would be taking a risk and probably lose placements when the risk doesn't pay off.

A skater who tries something really hard (such as a 4-3-3 combo) and completes the element successfully, or close to it, without falling, would be heavily rewarded for that risk. But he might fall elsewhere in the program, maybe a couple times, and lose fewer points than the skater who fell on the big trick.

Whatever the fall rules, sometimes skaters who fall a few times will come out ahead of skaters who fall once or not at all. Maybe because the one who fell a couple times completed significantly more difficulty in all the other elements. Maybe because all her successful elements were of high quality and another skater didn't fall but also only barely squeaked out all the elements with barely acceptable quality or subtle errors.

If casual fans are just counting falls without understanding relative difficulty and relative quality of the successful elements, they're going to be confused. In addition to adequate penalties for falls, it's important to make it clear what's being rewarded and what's being penalized -- in the scoring system itself, which the protocols do, and in the TV commentary, which the ISU has little control over but is usually the only source of knowledge for casual fans.

PCS

These are all the scores for global aspects of the program as a whole, rather than for individual elements.

How could the rules/guidelines for scoring these aspects and reporting those scores be improved to reflect differences that knowledgeable audiences see from one performance of a program to another, or between different aspects of the same performance?

I think we all want PCS scoring to be better, but it's a lot easier to say "It should be better" than to make concrete suggestions to help it get there, including defining exactly what "better" would consist of.

So who has some specific suggestions? Change the rules? Change the judges' training? Change the way the scores are reported so audiences have more insight into where the numbers come from?
 
Joined
Jun 21, 2003
Falls

Whatever the fall rules, sometimes skaters who fall a few times will come out ahead of skaters who fall once or not at all. Maybe because the one who fell a couple times completed significantly more difficulty in all the other elements. Maybe because all her successful elements were of high quality and another skater didn't fall but also only barely squeaked out all the elements with barely acceptable quality or subtle errors…

I think that the CoP has posed for itself a conundrum with no solution. IMHO no point system, however excellent, however rich in detail, can resolve questions of this sort in a way that is both consistent and appropriate to every combination of circumstances.

In ordinal systems it was the responsibility of the judge to decide the proper balance between difficulty and successful execution and to judge the extent to which a fall marred the effect of the program as a whole. Judges might disagree, but I never felt that such disagreement compromised the judging system or even that it was a negative feature of the system.

As for the audience, whatever the rules say some viewers will think, "He fell, how can he win anything?" and others will say, "Oh, too bad about the fall, but it was a great performance even so."
 

Blades of Passion

Skating is Art, if you let it be
Record Breaker
Joined
Sep 14, 2008
Country
France
If it were valid in Sochi, the Olympic champions would be Volosozhar/Trankov, Sotnikova, Ten.

Well, if this system was in place, programs would be constructed differently, performed differently, and rewarded differently. This system would have separate PCS judges, better training of judges, and judges being selected by a universal standard of peer-review. Yu-Na or Kostner most likely would win (assuming everyone else besides Sotnikova still made a lot of mistakes) because the system would allow them more artistic freedom and they would really be able to show that strength...and get rewarded for it!

If everyone skated exactly the same, Hanyu would still be the winner because everyone made a lot of mistakes over the course of the competition but he dominated in the SP and still did enough in the LP. Sotnikova would not win because she gave up more points in the SP with her easier jump combination, she would not have received a free pass on her flawed 3Lutz+3Toe in the long, her footwork sequences would have scored lower, and her PCS would not have been quite as high. This guide doesn't cover improvements for Pairs and Dance programs (although the judging improvements would be universal), but given better scoring systems there, V/T would still win - there's really no debate about that - and the battle for Dance Gold probably would have been closer.
 

Blades of Passion

Skating is Art, if you let it be
Record Breaker
Joined
Sep 14, 2008
Country
France
PCS

How could the rules/guidelines for scoring these aspects and reporting those scores be improved to reflect differences that knowledgeable audiences see from one performance of a program to another, or between different aspects of the same performance?

I think we all want PCS scoring to be better, but it's a lot easier to say "It should be better" than to make concrete suggestions to help it get there, including defining exactly what "better" would consist of.

So who has some specific suggestions? Change the rules? Change the judges' training? Change the way the scores are reported so audiences have more insight into where the numbers come from?

I suggest all three of those things in the guide. Plus, judges should be selected from an "academy", wherein the quality of their judging is itself reviewed by the skating world and judges are selected for competitions based upon their individual merit, not just because they were appointed by a country and their name was drawn out of a hat.
 

gkelly

Record Breaker
Joined
Jul 26, 2003
Here's what you have to say about PCS in your "Guide":

*The program component judges are necessary for this aspect of programs to be judged accurately; the program needs to be looked at on its own without being influenced and distracted by also scoring technical elements at the same time. Judges need to start separating the 5 program components properly; it would be perfectly fine for a skater to receive a '9' on skating skills but a '6' for interpretation (or vice versa), if that's what happened in the performance. Judges need to examine choreography not for how many transitions are in the program, but based upon the actual intellectual concept, visual appeal, and cohesiveness of the movement itself. There will no longer be a "judging corridor" that judges must worry about. All program component judges must have a firm mathematical understanding of what their scores mean (ie - you can't just say "this skater had better transitions" and score them .25 higher in that component. There must be a clear differential between the scores given to each skater for each performance, if that is what the performances merit).

*All judges must have extensive historical training. A trained judge will have watched hundreds and hundreds, if not thousands, of high-level performances across a wide spectrum of decades. This basis is absolutely necessary for having an objective opinion of what constitutes good figure skating in the many different forms it can take.

That's a fine overview. Now let's discuss how judges should separate the 5 component scores properly -- what does "properly" mean in this context?

How can we tell from the outside, looking only at their final numbers, whether someone is using the proper process?

Let's say I'm a judge who wants to be trained to do a better job of identifying and evaluating all the different component criteria and translating those evaluations into numbers. Do folks here want to talk about that process in more depth?

Shall we make a separate thread to get specific about PCS, to separate that discussion from all the nitty gritty technical details?
 

skatedreamer

Medalist
Joined
Feb 18, 2014
Country
United-States
I suggest all three of those things in the guide. Plus, judges should be selected from an "academy", wherein the quality of their judging is itself reviewed by the skating world and judges are selected for competitions based upon their individual merit, not just because they were appointed by a country and their name was drawn out of a hat.

2 questions:

(a) Who would evaluate the judges and what would be the qualifications for "judging the judges?"

(b) In a scenario where judges are chosen based on merit, what would be the response to a federation that might complain that its judges are never chosen or not chosen frequently enough?

Just playing devil's advocate here -- I like the ideas but can imagine allegations of bias from countries who might feel their judges aren't given "equal time."
 

CanadianSkaterGuy

Record Breaker
Joined
Jan 25, 2013
As per a few posts above, a fall shouldn't completely negate an element. Skaters will be far less likely to attempt greater difficulty. I mean, a fully landed quad with a fall shouldn't receive less points than a level B scratch spin. That makes no sense.

The system BoP, while rather thorough, is unfortunately just logistically a nightmare even if it's more accurate. I also think it doesn't reward quality elements as much as it should. You'll get a lot of skaters avoiding levels because the GOE really isn't worth it going from a level 3 to a level 4 spin.

Also given the reduction penalty is 0.5, and given the scores, a fall on a 4S (after -0.5 and -2.3) would be worth 8.0 points, which is just 0.5 lower than the base value of a clean 3A -- a 4S with -3's as in a severe error without a fall gets the BV of a 3A. I'm all for rewarding quads and pushing for difficulty, but that doesn't seem right. The current system drops GOE a full 3 points.

A fall on a 3A would be worth 6.2 points, which is 0.5 greater than the value of a normal 3Z with +0 GOE... and worth about the same as if the 3Z got close to perfect grade of execution. :unsure: I think a +2 worthy triple lutz should score higher than a fall on a 3A, shouldn't it? Especially when you claim this system is to penalize errors more?
 
Top