How to make components relevant for men?

[jinabee]

Ok2, no need argue.
Can someone post the vid if Hanyu's 2017 Worlds LP with each element marked? Someone did one to show difficult transitions and SS, etc as proof why they think Hanyu was underscored during 2017 Worlds LP.
They also did one for Hanyu's 2018 Olympics LP(Seimei). Forgot the titles though.

Then do one for Nathan's 2018 Olympics and Worlds. Heck, do one for Shoma too, since there are complaints 88 is too high PCS for 3 falls.

The H&L video you mean is this one

If there was one for OWG Seimei, it's not on YT and has probably been deleted elsewhere because of copyright

As for expecting the fans who made the program analysis of H&L to do other skaters' programs too....the makes of that video were accused of bias, trying to claim he was grossly underscored, being ungrateful that he won and set a WR and exaggerating things like GOE bullets hit etc. you think they won't be accused of being biased against other skaters doing analysis for them too? These fans were also mocked for spending their time doing this - even though it was meant as much as a tool for newer fans entering the sport to start to understand how the judging works as much as it was a celebration of a wonderful performance. Why would they invest their time in doing the same for other skaters when they were mocked for doing it in the first place?

Some comparisons of transitions and one-foot skating were made - here and here- and the response here was that they were biased to make certain skaters look good and other skaters look bad.

So...If Nathan fans, Shoma fans, Boyang fand and Mikhail fans think they can be about as neutral as possible want to put their time and effort into making analysis videos for their favourite skaters, they should! But there's not any point in the fans who made these videos doing it since they just get accused of attacking other skaters or whatever.

Anyway. to make PCS relevant, the judging needs to change. Judges need to allow for the disparity between components so you can see skaters with strong PE and IN rewarded without raising SS if they don't have it, or in the case of some ladies reward their TR but knock down their CO, PE and IN when it suffers. Marks between judges with a greater gap than 0.5 should be reviewed (so if judges are giving 8.5, 8.25 for SS but one or two judges are giving 9, 9.25 etc. tbh goe values between judges that vary too wildly should also be reviewed), judges blatantly favouring skaters from certain countries should be investigated (not just the chinese judge from the men's fs at olympics but the USA judge too) and the componants should be scored by the paremeters actually lay out in the technical handbook not whatever arbitrary ass rules judges are currently following which seems to be a mix of country, reputation, technical content and whether or not the judges like you or the ISU wants you to lead.

Shoma, this season in particular, was mostly overscored in TR. Since his free was disappointly barren. His strengths lie mostly in PE, IN and he has pretty solid SS. At Worlds, everything took a little hit and maybe it still should've been lower but I didn't find it that terrible - while he did fall, he was quick to get up and get back into his program and worked to retain as much of the performance as possible. I think Nathan was overscored in every componant but his SS score was particularly laughable. And it was pointless because he would have won without the rediculous PCS scoring. He would've won wothout the marginal overscoring in the SP too. He would've won with his UR called too. That his Worlds performance outscored his already overscored Olymics FS despite not being as clean OR as well performed? Utterly pathetic on the judges' part in all honesty. Esspecially since he clearly would have won, with a comfortable margin, without the score inflation.

PCS is NOT just 'artistry'; things like SS, TR are technical, CO is fairly technical too since it involves ice-coverage and quality use of space and movement to match musical phrasing. And tbh what is involved in PE and IN isn't all that vague or completely subjective. Judging PCS better will result in better quality programs. Scrapping PCS will lead to programs being just entirely skating to pick up speed before jumping, doing the jump, and skating to another part of the rink to jump again.

 

[eaglehelang]

The H&L video you mean is this one

If there was one for OWG Seimei, it's not on YT and has probably been deleted elsewhere because of copyright

As for expecting the fans who made the program analysis of H&L to do other skaters' programs too....the makes of that video were accused of bias, trying to claim he was grossly underscored, being ungrateful that he won and set a WR and exaggerating things like GOE bullets hit etc. you think they won't be accused of being biased against other skaters doing analysis for them too? These fans were also mocked for spending their time doing this - even though it was meant as much as a tool for newer fans entering the sport to start to understand how the judging works as much as it was a celebration of a wonderful performance. Why would they invest their time in doing the same for other skaters when they were mocked for doing it in the first place?

Some comparisons of transitions and one-foot skating were made - here and here- and the response here was that they were biased to make certain skaters look good and other skaters look bad.

So...If Nathan fans, Shoma fans, Boyang fand and Mikhail fans think they can be about as neutral as possible want to put their time and effort into making analysis videos for their favourite skaters, they should! But there's not any point in the fans who made these videos doing it since they just get accused of attacking other skaters or whatever.
.

My apologies, I didnt elaborate since I was typing from my phone. What I meant was Hanyu fans did a video with details to justify Hanyu's score/high PCS(Since there were complaints the 2017 worlds LP was overscored <facepalm>).

I saw the 2018 Olympics Seimei one when it first came out 1 or 2 weeks ago. Now not sure if it still exists, but I reckon they will re-upload later if it's deleted for copyright

So Nathan fans especially can do one too to justify his 91 pts PCS at Worlds. Detail it element by element. Same goes for Shoma, Boyang, Michail fans. I'm not suggesting the same people who did Hanyu's vid do for the rest. There's the comparision videos for that. Even then, the guy who did for one foot skating, transitions, jumps analysis gets criticized .

 

[Old Cat Lady]

The problem is that well choreographed programs are HARDER and RISKIER so it is a competitive disadvantage and a waste of training time to work on transitions, line, positions, and musical nuance over a jump when there is no competitive reward for doing so. Are we okay with never seeing another Ina Bauer, spiral, spread eagle, hydroblade, cantilever, footwork beyond the requirements? Do we want ugly hunched spins and club footed landings? Why spend time learning quality edging? Why even have music playing?

Look at N. Chen. He didn’t skate his program “cleanish” until after he dropped 40 seconds of choreographic content.

Maybe P. Chan’s problem wasn’t that he had nerves of jello but that it was too hard for him to do the jumps combined with the difficulty of his programs. He had at least some better skates when his programs were a little emptier.

Strategically, Hanyu is a moron. His biggest weakness is his stamina. Maybe if he wasn’t doing spread eagles, spirals, footwork, etc. he wouln’t be out of gas at the end of his program and make mistakes on the jumps. The only time it makes sense to do transitory elements is right before a jump, and even that is minimally rewarded when considering some skaters seem to get those GOE points without them. With the judges using components like ordinals, it makes more sense to cut a minute of content off his program - he’d still get the highest component scores, save energy for his elements, then win off of GOE.

But how sad for the sport if we don't have morons like Hanyu, Fernandez, and Chan.

 

[lyndichee]

The problem is that well choreographed programs are HARDER and RISKIER so it is a competitive disadvantage and a waste of training time to work on transitions, line, positions, and musical nuance over a jump when there is no competitive reward for doing so. Are we okay with never seeing another Ina Bauer, spiral, spread eagle, hydroblade, cantilever, footwork beyond the requirements? Do we want ugly hunched spins and club footed landings? Why spend time learning quality edging? Why even have music playing?

Look at N. Chen. He didn’t skate his program “cleanish” until after he dropped 40 seconds of choreographic content.

Maybe P. Chan’s problem wasn’t that he had nerves of jello but that it was too hard for him to do the jumps combined with the difficulty of his programs. He had at least some better skates when his programs were a little emptier.

Strategically, Hanyu is a moron. His biggest weakness is his stamina. Maybe if he wasn’t doing spread eagles, spirals, footwork, etc. he wouln’t be out of gas at the end of his program and make mistakes on the jumps. The only time it makes sense to do transitory elements is right before a jump, and even that is minimally rewarded when considering some skaters seem to get those GOE points without them. With the judges using components like ordinals, it makes more sense to cut a minute of content off his program - he’d still get the highest component scores then win off of GOE.

But how sad for the sport if we don't have morons like Hanyu, Fernandez, and Chan.

Morons? Am I missing something huge here? I think strategically Hanyu and Fernandez have been great. Hanyu isn't a moron if his strategy made him the first man since Dick Button to defend their Olympic title and break world records. Fernandez has won tons of major titles and Chan did well for a quite a bit but I don't think his pursuit for quality skating really cost him. I think age and injuries caught up to him. Just look at his 2016 4CC FS, he basically maxed out everything he had in that FS. He just doesn't have the technical content to compete and also this season he was lacking a lot of motivation which I think affected his skating.

 

[Old Cat Lady]

Morons? .

Context. I don't actually think they're morons. Just using hyperbole to emphasize my point - that component scores don't properly reward the non-jump elements.

 

[Baron Vladimir]

^^^^^
Yeah. Well, ISU wants to reduce a base value of the jumps a little bit and percentage which + GOE adds to quads. So good executed triple will value more than bad quad and TES score will generally be a little lower if you just rotate quads (without additional feature or quality)...

 

[epicdreamer]

Context. I don't actually think they're morons. Just using hyperbole to emphasize my point - that component scores don't properly reward the non-jump elements.

Haha didn't the Italian B.ESP guys say the same thing in one of their podcasts? How Yuzuru doesn't get compensated proportionally for all the "extra" stuff he does, so he seems like a fool for even bothering when reputation for multiple quads seem to just save you anyway.

 

[moriel]

To make components relevant to any discipline, they must be scored accurately. Not because of reputation, not because of consistency, not because of fed.
They must actually reflect what they are meant to, instead of just being 5 scores somewhat proportional to TES.
They must reflect what was actually done, not just "ok, this skater gets 75 for a terrible performance and 76 for an awesome performance".
They must reflect what was actually done, not just "this guy here is from a big fed, so he gets a +10 over that other guy who is from a random small fed".

That would be a good first step.
Right now, PCs are a joke.

 

[Old Cat Lady]

For example: How to determine P/E score
Create Scaled points for hitting levels or all of their intended jumps, scaled points for GOE related success, etc. The point is to tie it directly to blade to ice performance. :think:

The more I think about it, the more I think this idea is fantastic!

With the upcoming GOE going from -5 to +5, it allows for much more variance in scoring, which is what is important. One of the big problems with the current system is that judges will only award a fraction of a point for drastically different quality of component. If each of the 5 components were marked on a GOE, it would allow for more of a swing on one component mark than we get with all the component marks put together currently.

drop the high and the low, any judge that is greater than 2 points off the average goes under review. Too many times under review, that judge gets suspended, with each additional suspension increasing.

 

[Metis]

I’m willing to wager that most rational, reasonable people would stipulate to the following points:
1. PCS and GOEs are fundamentally broken systems at this point. We may not all agree on the why and where to go from here, but it’s clear that a 9.5 means “whatever the judges want on a given day” and has no absolute or relative value at this point.
2. There’s an inherently subjective element to figure skating that we will never be able to quantify. That’s okay. What’s not okay is arbitrary judging, PCS inflation, reputation PCS, the tech panel enforcing the rules only on certain days and often only with regard to certain skaters, PCS being used to create podium order, and the many other complaints that have been named.
3. Not even ISU seems interested in clarifying what a “7” actually means, and at this point, PCS has functionally maxed out for some skaters (deservedly or undeservedly).
4. Two of the five PCS categories can actually be rather objectively measured and quantified (SS and TR), while three of them need, minimally, more guidance. (What does CO mean? Is opening with three toe jumps thoughtful programme composition? Should elements not only be evenly distributed but also evenly represented? I don’t know!) Why are they grouped together? I have no clue.
5. It may be impossible for the human brain to evaluate performances in isolation (which means PCS will always be a comparative judgment, even if judges do their best to fight that impulse). If that’s the case, PCS is functionally a flawed implementation of ranked choice voting, making it basically more opaque ordinals. (Ordinals, while stupid, had the benefit of being gloriously, transparently stupid to the point of amusement, and you never wondered why the American judge was lowballing Katerina Witt.)
6. There’s a correlation between TES (not BV, but TES) and PCS, which could be solved by changing PCS factouring. It’s not a neat and tidy ratio, however, and it doesn’t account for all of the variance. (Not does shared nationality of a skater and a judge, which was already studied.) As much as I’d like to just blame TES, skating order still matters (it shouldn’t), though it’s inarguably part of the problem.
7. ISU has about zero interest in dealing with any of this, let alone detecting statistical anomalies and actually punishing judges, giving tech panels proper equipment, etc.

And then there’s the fact that use of a trimmed mean has an exploit a truck can drive through and creates an inescapable metagame. tl;dr of my own tl;dr: if you’re a judge and want a skater to score well, the most optimal play is to try to be the highest mark in the average, thus preserving all the next-highest marks for the average. You are more useful as the discarded high than you are as part of the actual average. There is absolutely zero incentive to score “honestly,” for any given definition of the term.

(You don’t need to understand game theory to understand this, though I referenced it in explaining why “aim to be thrown out of the average” is in fact the optimal choice. If you have to guess what everyone else is going to score and you can accurately assume the range is going to be roughly 9.0-9.50, you want to be 9.75. Why? If there’s another 9.75, at worst, one of the two is tossed. If there isn’t, every 9.50 is now in the average. If you had put down 9.50 and there was a 9.75, you just lowered it. If there wasn’t, you didn’t do any harm, but mostly you were simply lucky. Rather than rely on luck and roll the dice, it’s to your advantage to be the highest mark, as now all scores equal to or below yours make the average. When multiple judges on a panel are doing this — and I suspect some are — there’s a point at which collapse is inevitable, as the metagamers will create an upward shift in PCS as they fall into the average, and eventually, everyone maxes out and/or achieves an unreasonably high PCS “floor” that leaves no room for growth or inflation when necessary for podium ordering. It’s a great system!)

As far as fixing it? I have ideas. Most fall into “lulz will never happen” territory, but here’s one: turn GOE levels into a progressive scale, rather than the current à la carte system. I will note in advance that the scale I took thirty seconds to make could be seen as punishing “small” jumpers, and I also did not include the current GOE bullet for matching the element to the music; in the case of the latter, I feel that’s better handled by PCS (specifically IE), whereas GOEs should be about execution, and as for the former... I spent thirty seconds ordering the current GOE bullets in what I felt was the most sensible way, but, yes, that’s obviously an issue (at least for Satoko). But on principle, I would prefer to see GOEs be progressive rather than awarded based on sheer number of bullets, as it means judges can’t abuse their discretion and reward poor technique with high GOEs. I also believe GOEs should actually reward mastery of a given element, with consideration for musical phrasing rewarded in CO and IE, not TES. PCS and TES should be clearly delineated.

As for PCS? Stop using a trimmed mean. Is a generic average better? No, not really, if you’re concerned with inflation and arbitrary values. In order for numbers to mean anything, ISU will have to issue actual guidance and flesh out the categories; even a basic rubric is a start. I’ve played around with alternative ways of calculating PCS values, such as the median or averaged median — and there are pros and cons. It’s also just not going to happen, as moving to some form of taking the median value reduces the power of each judge, weakens the ease of bloc voting, etc. A median also is more likely to represent a narrower range of values than a trimmed mean (not always, but I’ve messed around with last year’s scores from WC and Uno’s components are a headache), which is a real weakness. On the other hand, a median is much better at rewarding honesty and showing the consensus view. Which is why it’s pointless to pretend the trimmed mean is going anywhere — ISU wants judges to be able to mess with PCS enough to be able to use it as an ordering system.

Which makes me wonder why we ISU doesn’t actually just implement an ordering system. Have the judges rank the skaters 1-6 at the end of each flight. Personally, I’d love it if this never were released publicly and only I got to see the data, as then the judges would have more reason to be honest and I could actually test whether or not their scores matched their rankings. Since that’s not going to happen... ISU could actually still keep it private and use it as another self-checking mechanism that’s factored into their judging corridor (which needs significant revision in PCS). Or make it worth some extra 20% in PCS (there’s the 1.2 factouring). But there’s a lot to be learned from voting systems, actually, and if PCS is being used as a comparative measurement and ranking system... add a comparative measurement and ranking system, for crying out loud. That way, there’s perhaps less incentive to use PCS for that purpose, and no reason to “save marks” when you’re ranking the entire flight after its members have skated. This will also never happen.

Something that almost certainly is happening that is not going to fix anything because maths is hard: reduced BV on quads.

 

[Baron Vladimir]

I’m willing to wager that most rational, reasonable people...

And then there’s the fact that use of a trimmed mean has an exploit a truck can drive through and creates an inescapable metagame. tl;dr of my own tl;dr: if you’re a judge and want a skater to score well, the most optimal play is to try to be the highest mark in the average, thus preserving all the next-highest marks for the average. You are more useful as the discarded high than you are as part of the actual average. There is absolutely zero incentive to score “honestly,” for any given definition of the term.

(You don’t need to understand game theory to understand this, though I referenced it in explaining why “aim to be thrown out of the average” is in fact the optimal choice. If you have to guess what everyone else is going to score and you can accurately assume the range is going to be roughly 9.0-9.50, you want to be 9.75. Why? If there’s another 9.75, at worst, one of the two is tossed. If there isn’t, every 9.50 is now in the average. If you had put down 9.50 and there was a 9.75, you just lowered it. If there wasn’t, you didn’t do any harm, but mostly you were simply lucky. Rather than rely on luck and roll the dice, it’s to your advantage to be the highest mark, as now all scores equal to or below yours make the average. When multiple judges on a panel are doing this — and I suspect some are — there’s a point at which collapse is inevitable, as the metagamers will create an upward shift in PCS as they fall into the average, and eventually, everyone maxes out and/or achieves an unreasonably high PCS “floor” that leaves no room for growth or inflation when necessary for podium ordering. It’s a great system!)

As far as fixing it? I have ideas. Most fall into “lulz will never happen” territory, but here’s one: turn GOE levels into a progressive scale, rather than the current à la carte system. I will note in advance that the scale I took thirty seconds to make could be seen as punishing “small” jumpers, and I also did not include the current GOE bullet for matching the element to the music; in the case of the latter, I feel that’s better handled by PCS (specifically IE), whereas GOEs should be about execution, and as for the former... I spent thirty seconds ordering the current GOE bullets in what I felt was the most sensible way, but, yes, that’s obviously an issue (at least for Satoko). But on principle, I would prefer to see GOEs be progressive rather than awarded based on sheer number of bullets, as it means judges can’t abuse their discretion and reward poor technique with high GOEs. I also believe GOEs should actually reward mastery of a given element, with consideration for musical phrasing rewarded in CO and IE, not TES. PCS and TES should be clearly delineated.

As for PCS? Stop using a trimmed mean. Is a generic average better? No, not really, if you’re concerned with inflation and arbitrary values. In order for numbers to mean anything, ISU will have to issue actual guidance and flesh out the categories; even a basic rubric is a start. I’ve played around with alternative ways of calculating PCS values, such as the median or averaged median — and there are pros and cons. It’s also just not going to happen, as moving to some form of taking the median value reduces the power of each judge, weakens the ease of bloc voting, etc. A median also is more likely to represent a narrower range of values than a trimmed mean (not always, but I’ve messed around with last year’s scores from WC and Uno’s components are a headache), which is a real weakness. On the other hand, a median is much better at rewarding honesty and showing the consensus view. Which is why it’s pointless to pretend the trimmed mean is going anywhere — ISU wants judges to be able to mess with PCS enough to be able to use it as an ordering system.
.

Even if one judge use highest mark of average for one PCS category, he/she doesnt know if it is enough to award that skater with better placement in the end. First, because many judges can use the same logic so every skater who are medal contender gets the same push by some judge. So, at the end that judges mark is not making a difference. Also that judge doesnt know for how much he/she should up some number to neutralize the difference in tech score because he/she cant control tech pannel calls. The point is - even if judge wants to put someone in first place he/she cant do that in current system because his/her numbers are just a small part of other numbers involved and other games playing around. While in Ordinal judging same judge can manipulate the final placement just with one number. I mean, you understand statistic at least for that to know that manipulation where more numbers and more variables are included is much harder to be done. And that ordinal system as matematical concept has less positives.

 

[Metis]

Even if one judge use highest mark of average for one PCS category, he/she doesnt know if it is enough to award that skater with better placement in the end. First, because many judges can use the same logic so every skater who are medal contender gets the same push by some judge. So, at the end that judges mark is not making a difference. Also that judge doesnt know for how much he/she should up some number to neutralize the difference in tech score because he/she cant control tech pannel calls. While in Ordinal judging that same judge can manipulate the final placement just with one number. I mean, you understand statistic at least for that to understand that manipulation where more numbers and more variables are included is much harder to be done. And that ordinal system as matematical concept has less positives.

1. I said the use of a trimmed mean creates an incentive to try to score in such a way as to be the high in the average, not that any attempts at doing so grant a 100% success rate.

2. Using a rational actor model, scoring is an iterated game. I already wrote way too much on this. The system is not actually all that volatile.

3. Of course a judge’s marks make a difference. The whole point of trying to guess high is to uncap the high end of the scoring distribution for the average. Look at protocols. Your inflation argument actually works against you — it’s arguably why there’s less volatility than you’d expect than if everyone (or at least some members of the panel) weren’t already scoring “upwards.” Chen’s largest gap in components is one point, and that only occurs once, in SS; for all other components, the variance is within 0.75 to 0.50. Uno’s component marks are all within a 0.75 or 0.50 range. And so on.

I’m not going to fight you on this. The system flat-out incentivizes going high. This was part of the Sochi saga and was noted by Buzzfeed in their coverage of this year’s OWG. It is not a matter of “knowing” what the rest of the panel will do but making an educated guess, which is honestly not difficult to do after hours of judging and gaining insight into where the panel is averaging. Not to mention the fact that judges have scoring histories and it’s likely possible to build a probabilistic model of how a given panel will score skaters based on their own prior data.

Base inflation + half a point or so should throw you out of the average without triggering a 10 or major red flags. Or you could YOLO and go full 10s. But the basic maths here is incontrovertible: if you want to raise someone’s PCS, try to save all the next highest marks for the average by being the highest. I’m not pointing out anything especially new, for the record — this has always been an issue.

 

[Sam-Skwantch]

One good thing about the top PCS marks going so high is it eliminates some opportunity for the judges to save their favorites who make boo boos on their jumps and spins. Some skaters you could almost feel the PCS going up with each fall.

I stand by what I have been saying for years. The only way to remove power from dirty judges is to reduce the value of PCS. At this point there is almost nothing logically quantifiable about the scores we see regardless of the endless debating. We need to see performance based PCS scores (via hitting levels, number of landed jumps, etc) so that the skater has actual control of their fate with what they do on the ice making the difference. The way a host selects the judges and the type of panel the ISU assigns shouldn’t affect the outcome. The skaters training and athletic efforts...yeah....that should be what wins the day :yes:

Maybe some rich artistic purists will get together and start a performance based pro circuit that reduces the value of technical elements and celebrates acting and gliding. I say go for it. Maybe in Europe or Japan it would work but the US...I’m not so confident. What do I know though? I’m just a rando on the Internet who may or may not have experience in such matters

 

[sanfan]

I agree that PCS and GOE are too biased and not functioning well, but reduce/remove them will not solve the problem.
Some people say if you like artistic performance, just go to ice shows.
I would say why don't we just have a jumping on ice competition. No need for music, choreography, or costume.
Why can't we have both?

Also, the technique of figure skating is not just about rotation of jumps, while BV only captures the rotations.
As people pointed out earlier, PCS and GOE are still mainly about the technique.
We just haven't find out a better way to quantify them. It doesn't mean we should give up.

 

[Baron Vladimir]

1. I said the use of a trimmed mean creates an incentive to try to score in such a way as to be the high in the average, not that any attempts at doing so grant a 100% success rate.

2. Using a rational actor model, scoring is an iterated game. I already wrote way too much on this. The system is not actually all that volatile.

3. Of course a judge’s marks make a difference. The whole point of trying to guess high is to uncap the high end of the scoring distribution for the average. Look at protocols. Your inflation argument actually works against you — it’s arguably why there’s less volatility than you’d expect than if everyone (or at least some members of the panel) weren’t already scoring “upwards.” Chen’s largest gap in components is one point, and that only occurs once, in SS; for all other components, the variance is within 0.75 to 0.50. Uno’s component marks are all within a 0.75 or 0.50 range. And so on.

I’m not going to fight you on this. The system flat-out incentivizes going high. This was part of the Sochi saga and was noted by Buzzfeed in their coverage of this year’s OWG. It is not a matter of “knowing” what the rest of the panel will do but making an educated guess, which is honestly not difficult to do after hours of judging and gaining insight into where the panel is averaging. Not to mention the fact that judges have scoring histories and it’s likely possible to build a probabilistic model of how a given panel will score skaters based on their own prior data.

Base inflation + half a point or so should throw you out of the average without triggering a 10 or major red flags. Or you could YOLO and go full 10s. But the basic maths here is incontrovertible: if you want to raise someone’s PCS, try to save all the next highest marks for the average by being the highest. I’m not pointing out anything especially new, for the record — this has always been an issue.

It's a issue itself, i agree, but my point was - even one judge is using that kind of educating guess, it just doesn't prove that his educated guess can change the final placement of some particular skater. So, even if system allows that kind of judges behaviour, it doesn't mean how judge's behaviour can 'broke the system' that easily (in terms that can manipulate its final products) cause system control itself by other variables involved. Individual judges can't see the score of other judges (who put a lot of numbers in, and any of those numbers can change the game in every moment), so one judge doesn't know how he/she should play the 'game' to change the game other judges are already playing for other skaters. So if you look at one judge - yes there is a problem. But other 8 individual judges plus more from tech panel control his game, like he/she controls theirs. So, regarding the outcome, his individual behaviour doesn't get as a problem any more. Also, we had only 2 systems of scoring for now. Of course that current is not perfect, but it still control judge's individual manipulations much better than the one existed before. And if the current system is bad, you must asking the question - comparing to which system it is bad? The some other possible system who could exist? I mean, dont get me wrong, your analysis is great and it can show something about some individual judge"s behaviour. But only that. We just cant jump to any big conclusions (aka system is bad and can be manipulated) based on that.

 

[Ares]

One good thing about the top PCS marks going so high is it eliminates some opportunity for the judges to save their favorites who make boo boos on their jumps and spins. Some skaters you could almost feel the PCS going up with each fall.

I stand by what I have been saying for years. The only way to remove power from dirty judges is to reduce the value of PCS. At this point there is almost nothing logically quantifiable about the scores we see regardless of the endless debating. We need to see performance based PCS scores so that the skater has actual control of their fate with what they do on the ice making the difference. The way a host selects the judges and the type of panel the ISU assigns shouldn’t affect the outcome. The skaters training and athletic efforts...yeah....that should be what wins the day :yes:

Maybe some rich artistic purists will get together and start a performance based pro circuit that reduces the value of technical elements and celebrates acting and gliding. I say go for it. Maybe in Europe or Japan it would work but the US...I’m not so confident. What do I know though? I’m just a rando on the Internet who may or may not have experience in such matters

Your suggestions could make sense if not for the fact that TES-marking is very arbitrary too and blatantly unfair too when you look into detail even if it means just few points less or more here or there it could still make a significant difference in placements. I mean there's one Russian lady who flutzes every time yet gets no calls while the other skaters do get such calls in Pyeong Chang, also has barely over 2 rotations in her triple jumps yet astronomical GOE and full BV and overscored spins too. Or Aliev who received full BV for underrotated and blade-pick 4Lz and almost the same points as Boyang Jin for his impeccable quadruple lutz combination in Pyeong Chang, but somehow got rightful call in Milano. Or Uno who is treated in very generous way as far as his jumps are concerned. There's lots of bias, some skaters are untouchable.

 

[Old Cat Lady]

Unfortunately, no matter what system you use there is always opportunity for manipulation so the actual judging process should be based purely on rating the value of each aspect of the skating and allowing for proper reward. That's why it's so important to have a system of checks and consequences for the judges. I'm not sure how other sports do it or why the ISU doesn't have an independent set of judges, but if they're not going to train a set of judges that are independent of federations (but still from different countries), the ISU should at least dictate which judges the federations are allowed to send. Right now, judges are incentivized to work in the federation's interest rather than skating's interest. The ISU needs to have a standardized system of review, an outlined set of consequences for infractions (or for incompetency), and follow it for every incident no matter what. No judge should be allowed to judge international competitions unless they pass an ISU judging test and if they are found to not adhere to ISU standards, they need to be suspended.

What's interesting is that the disparity in the relevance between components vs TES was actually more pronounced in the ladies. Zagitova got the 2nd LOWEST components score in the top 5 at the Olympics yet still won in spite of not having anything overwhelmingly more difficult. she didn't have to do anything all that special technically to gain her advantage - no triple axel or quad, good but not exceptional quality in everything she did, didn't bring any new element to the sport - in spite of only being ranked 4th in components, she won by gaming the tes score better than everyone else.

 

[Baron Vladimir]

^^^
topic: To be fair, she had something technicaly special, 100% backloaded programe. Also, the first one who ever did it. So, she indeed had something for which they can always say it make a difference that night, no matter some like it or not...
And i agree, the first thing ISU can do is not to let judges from countries with skaters competing (and who are medal contenders) in Oly and Worlds judging the competition... So no conflict of interest there...

 

[Old Cat Lady]

^^^
To be fair, she had something technicaly special, 100% backloaded programe.

That's what I was referring to when I wrote "gaming the system".

The point I was making was that the way the components are judged make them so irrelevant that even the 4th place scorer can win without extraordinary difficulty. I think she deserved her win and she did have the most difficult program (made more difficult with the back loading), but her technical wasn't extremely above the field the way it was with Nathan at worlds. Medvedeva's technical base value was only 3.68 points lower, had comparable or higher GOE's, and was the clear first with the component mark, but the second mark matters so little that she couldn't overcome such a marginal difference in the TES.

The ISU sent the message that a single mediocre triple toe is worth more than all the transitions, line, carriage, skating skills, performance, composition combined

 

[Florian]

I would not make major changes regarding the system, especially not regarding the TE. I think two separate panels would already help, in combination with relatively concrete guidelines and monitoring of the judges. The system has produced undeservedly inflated and incomparable PC scores and this needs to stop, but all in all it worked quite well. I mean at least it hasn't produced any undeserved world champion in the past years as far as I can see.

I also think that it was necessary to increase the technical level in the men's competition and that this should not be reverted by rewarding quads any less. One could at best think about reducing the jumping elements in the free skate from 8 to 7.
Chen's technically outstanding and unique victory should not be used to question everything. Not trying to defend the judges, but exaggerating his PC after his performance in the context of the last group should not be overestimated.