PCS per judge - Data Analysis AWG Mens FS

[Miller]

Wonder if the final 3 'performance' categories could somehow be compressed into 2 e.g. something like Choreography and Composition, and Interpretation and Execution. Thing is get the impression that the judges tend think of an overall PCS base figure and then go up and down for the individual components, hence why they're so similar, plus they've got 5 things to think of, amongst everything else.

However occasionally you get a program like this https://www.youtube.com/watch?v=CnPMx3s9BhY where the British Eurosport guys describe it as the best Carmen they've ever seen, but she got 7.50/7.21/7.32 for PE/CH/IN compared with 7.04/6.93 for SS/TR. If the last 3 were combined into 2, with 30% weighting each, you might get the judges thinking SS, TR and 'Performance' (rule of 3 things in a list) and then upping and downing the last one accordingly. Is this feasible from what people know about the individual components, because if so you might get more PCS scores/individual components that better reflect the performance, IMO.

 

[Mathman]

Wonder if the final 3 'performance' categories could somehow be compressed into 2 e.g. something like Choreography and Composition, and Interpretation and Execution.

I like this idea. To me, the component "Musical Interpretation" is the hardest to objectify and resolve into bullet points. I would be in favor of incorporating it into a broader category.

This still leave Choreography sticking out like a sore thumb. We do not want to give extra points to a skaters just because he can afford an expensive choreographer and another skater cannot. So what we should be looking at is whether the choreography is well-matched with the skaters strengths and how well the skater executes the intent and details of the "choreography-on-paper." Just as "Musical Interpretation does not focus on how great the composer was. (Everyone would skate to Haydn if that were the case. )

[The current scoring gives] the impression that the judges tend think of an overall PCS base figure and then go up and down for the individual components, hence why they're so similar, plus they've got 5 things to think of, amongst everything else....

I think this is correct. The judges do indeed give the impression that they are saying to themselves, "This is on overall 7.5 program. Maybe the Musical Interpretation was a little stronger than the Transitions.." I don't know if the judges actually do think this way, but the scores make it seem as if they do.

However, to me that is not really an adverse criticism of the judges' mindset. On the contrary, that seems like a very sensible -- possibly a psychologically inevitable -- way to go about it.

 

[Miller]

I like this idea. To me, the component "Musical Interpretation" is the hardest to objectify and resolve into bullet points. I would be in favor of incorporating it into a broader category.

This still leave Choreography sticking out like a sore thumb. We do not want to give extra points to a skaters just because he can afford an expensive choreographer and another skater cannot. So what we should be looking at is whether the choreography is well-matched with the skaters strengths and how well the skater executes the intent and details of the "choreography-on-paper." Just as "Musical Interpretation does not focus on how great the composer was. (Everyone would skate to Haydn if that were the case. )



I think this is correct. The judges do indeed give the impression that they are saying to themselves, "This is on overall 7.5 program. Maybe the Musical Interpretation was a little stronger than the Transitions.." I don't know if the judges actually do think this way, but the scores make it seem as if they do.

However, to me that is not really an adverse criticism of the judges' mindset. On the contrary, that seems like a very sensible -- possibly a psychologically inevitable -- way to go about it.

Yes, agree with you on this, and your choreography comments, especially re paying. Perhaps 20% for Choreography/Composition, plus 40% for Interpretation and Execution is the way to go, which in some ways gets us back to where we started, but we've compressed Performance/Execution and Interpretation into one category where the judges can perhaps concentrate on it a bit more, and maybe reward outstanding interpretative programs a bit more, rather than them falling under the plus or minus mindset.

 

[gladiolusc]

I think, at the end of the day, the categories have to be such that, when the skaters/fans look at the protocol to find ways to improve, it should be obvious where they can focus on to gain points for that particular category. For example, the answer to 'how can I improve my SS score' should be in the ballpark of 'do more stroking exercises and improve edges' etc. and not 'do more difficult jumps or improve my spins'. And later, when the edges have improved, the scores should reflect in the SS score. Perhaps an easier way to judge an elite skater would be 'how else can this skater improve' and score the categories accordingly by deducting from the maximum points. 10 does not necessarily equal absolute perfection, but the fact that it's the maximum score in this system means that there is inherently a 'perfect' level of what that category could be in the judges mind So if the impression that the judges get from a skate is that 'he/she could have better transitions and should work on that', then they should deduct from the transitions score instead of trying to quantify it by starting from 0.

I really agree with this. I think this would be lovely and better for the field at the end of the day. It reminds me of what Lambiel said about Deniss in his interview at Inside Skating. Essentially their answer to "How do I improve my components score" under the current way skating is judged have to be, unfortunately, "Do more difficult jumps".

I enjoyed his way of expressing himself, he has something special. Of course, we work very hard on the quad, and we want him to get a very high technical score. And if the technical score gets higher, for sure the components will follow.

 

[Geek On Ice]

All of your posts are soooo interesting! :luv17:
I regret that I don't have enough time to reply to each point that I agree and join the discussion in depth...

I'm sure many others talked about this, but I'm starting to think that the "new" scoring system may be getting "old".

It could be due to the rapid growth of brilliant skaters with many/different quads but also the longevity of skaters may play a role. As skaters tend to stay active longer and keep improving, if skaters reach almost at the max PCS, they don't have more room to grow, compared to new comers whose PCS scores can keep rising. Also I haven't gathered exact data, but it seems that young skaters tend to receive higher PCS faster than before? Which is good, if skaters are evaluated by their skills instead of seniority, but ATST, it could be problematic if the increasing speed/curve is different, especially when we have the maximum in PCS, no? (ie. I wonder if Patrick/Javier have same speed/curve in PCS increase in course of their career, compared to 'younger' skaters like Nathan/Shoma. I know, each skaters has different speed in their growth, but as an average as a whole 'era'... Am I making sense??)

Even when a skater makes an absolultly wonderful perfomance, by displaying exceptional skills in PCS components, if PCS difference is only about 0.5 each, then that is only 5.0 difference in FS... Less than one 3Lo. This may be the reason we see huge gap from top clusters with multiple quads and the clusters that follows them. (Especially, "IF" PCS tend to be pulled higher for higher TES/GOE, as Lambiel and others point out.)

I am NOT taking TES lightly, because I have a huge respect for those raising the bar technically. Also top skaters usually excell in both TES/PCS terms. However, I can't help to worry that "variety" or "unique characteristics" of skaters could be lost if we don't improve the system. Just my honest opinion...

Another thing... I agree that "skating skills" should BE the base for the most of evaluation, because after all, this is figure "skating". However, ATST, I wonder if "non-skating-skill-related-factors" is valued correctly. I'm not saying it should be valued equally as "skating-skill-related-factors", but the ratio currently now seems too small maybe? (* "non-skating-skill-related-factors" - I don't know if this wording is a good one... I mean those skills not directly connected to edge work etc. Yes I guess it is subjective & hard to quantify or list detailed points... but something like Jason's fantastic sensitivity to music in 4CC or Jullian's powerful projection in AWG, for example...?)

And I think even for skaters, it is hard to tell what is exactly evaluated in current system PCS anyway, as Daisuke once wrote in his article.

 

[Mathman]

Another thing... I agree that "skating skills" should BE the base for the most of evaluation, because after all, this is figure "skating". However, ATST, I wonder if "non-skating-skill-related-factors" is valued correctly.

Well, the old 6.0 ordinal system divided the scores into tech vs. performance, 50-50. When the CoP came along this ratio was changed to tech (elements plus SS plus TR) = 70% and Performance/artistry = 30%. (This is a sport after all, so technical skills should be more important than performance, and also easier to evaluate with some degree of objectivity.)

Obviously skating skills are the most important because if you can't skate, then you can jump, carry out the choreography, or anything else. A 0 in skating skills automatically means a 0 for the whole program. But the component SS is of course more narrowly defined than "being able to skate well." To me, that is why which it would be better to find a new name for that component. How about BS ("blade skills"?)

Or you could call TES "Listed elements" and expand TR to "connecting moves, moves in the field, and non-listed jumps." Anyway, that's sort of why I think that it worth considering to have "Listed and Required Elements (jumps, spins, footwork -- 50%) and "Program" (50%). With the program components further divided into "Blade skills, connecting moves and unlisted elements"(25%) and "performance/artistic expression" (25%). Something for everyone.

 

[karne]

Anyway, I think we've strayed from the point, which was the men's PCS at AWG, and IMO, the absolute disgrace that is Ten's slop-fest getting more PCS than Brendan's FS. Hmph.

 

[gkelly]

I wonder if Patrick/Javier have same speed/curve in PCS increase in course of their career, compared to 'younger' skaters like Nathan/Shoma. I know, each skaters has different speed in their growth, but as an average as a whole 'era'... Am I making sense??)

It would take more analysis -- and still only be an estimate -- to separate out the effect of individual skater improvements vs. systematic changes in the way judges have been awarding the scores. I do think that judges have become more willing to give top-level PCS in the last 3-4 years than they were a decade ago. And also that the overall standard of what the skaters collectively are putting out there in terms of performance has increased just as much as the jump content, as well as these two specific skaters improving their skills.

But for what it's worth, here are Chan's PCS from his Worlds debut in 2008, where he finished 9th:

SP: 7.07 6.68 6.82 6.79 6.79
FS: 6.96 6.46 6.79 6.89 6.71

One year later he won silver with scores of
SP: 7.55 7.15 7.50 7.35 7.40
FS: 7.70 7.45 7.60 7.70 7.60

Fernandez first competed at Worlds in 2007, placing the 35th in the SP; in 2008 he placed 30th in the SP. Both years his PCS averaged in the low 4s and he did not qualify for the freeskate.

His first year skating the whole competition at Worlds was 2009, finishing 19th. His PCS were

SP: 5.85 5.30 5.70 5.75 5.75
FS: 5.70 5.45 5.70 5.50 5.50

So Fernandez had a longer way to go to reach the medal stand and it took him more years to get to that level.

 

[4everchan]

^^

well i think that every year, we start at a higher level... we don't see many 4s anymore in senior ranks do we? Even in top 30s.... and we don't see many silver medallists below 8s... and especially not a guy like Patrick who is a near ideal PCS skater.

So, yeah, it seems to me that the curve is already started each year higher.... No offense to the young guns of the last two years, but they get around 7.75 to 8.75 right now, for PCS that are to my eye at least, lower than a Patrick of 2009....

 

[4everchan]

Well, the old 6.0 ordinal system divided the scores into tech vs. performance, 50-50. When the CoP came along this ratio was changed to tech (elements plus SS plus TR) = 70% and Performance/artistry = 30%. (This is a sport after all, so technical skills should be more important than performance, and also easier to evaluate with some degree of objectivity.)

Obviously skating skills are the most important because if you can't skate, then you can jump, carry out the choreography, or anything else. A 0 in skating skills automatically means a 0 for the whole program. But the component SS is of course more narrowly defined than "being able to skate well." To me, that is why which it would be better to find a new name for that component. How about BS ("blade skills"?)

Or you could call TES "Listed elements" and expand TR to "connecting moves, moves in the field, and non-listed jumps." Anyway, that's sort of why I think that it worth considering to have "Listed and Required Elements (jumps, spins, footwork) and "Program" (50%). With the program components further divided into "Blade skills, connecting moves and unlisted elements"(25%) and "performance/artistic expression" (25%). Something for everyone.

Just wanted to point out that you cannot assign percentage to the TES versus PCS because PCS has an ultimate maximum and was always conceived like that, but TES doesn't have it.... Nobody has reached the TES maximums but certainly many have been in the high 90s in PCS.

Also, what bothers me most in that PCS is not a stand alone score... it looks to me that judges normalize it based on TES, to make sure that there is not a 15-20 point gap ...so that way a guy who is known for his great skating will not "win on PCS" or a guy who does many great quads will not "lose on PCS entirely" so we don't see 55 in TES with something like a 35 in PCS.... that 35 is raised up to a 43 for instance.... yest just about 7-8 years ago, a 43 wasn't even possible to get for the world champion

In other words, as the TES has been raising, the normalization to the PCS score has been adjusted.... but one thing was forgotten, the TES is pretty much limitless (only a couple guys get close to its max in the SP and there is still room to grow in the FP) while the PCS is capped.

 

[gkelly]

well i think that every year, we start at a higher level... we don't see many 4s anymore in senior ranks do we?

Not at Worlds. Not only because of judges marking everyone on a higher scale, but also because the skaters who are unable to earn at least 5s by today's standards also tend to be unable to meet the minimum technical scores for Worlds.

At 2016 Europeans, most of the guys who failed to make the top 24 cut for the freeskate had PCS in the 4s. This year, it was only a handful of them.

At this year's Four Continents, the bottom 4 guys had PCS in the 4s. There were only 26 entries, so two guys with 4s did make the cut.

At Senior B events, there might be plenty.

Also, what bothers me most in that PCS is not a stand alone score... it looks to me that judges normalize it based on TES, to make sure that there is not a 15-20 point gap

I don't know that they think of it those terms, because they have no reason to memorize the scale of values including the values of underrotated < quads and do the math in their heads. But they might on general principles believe that the difficulty of executing the difficult elements in a program that is presented reasonably well if not brilliantly should be rewarded enough to keep those skaters in contention against competitors with stronger presentation and lesser tech content, and boost the PCS accordingly.

 

[eppen]

I wonder if Patrick/Javier have same speed/curve in PCS increase in course of their career, compared to 'younger' skaters like Nathan/Shoma.[/B] I know, each skaters has different speed in their growth, but as an average as a whole 'era'... Am I making sense??

There have been some charts online showing in diagrams the development of both scores by the old guard, but unfortunately I have not kept saved the links.

If I remember correctly, they showed a relatively similar development for Hanyu and Chan - a fairly steady overall development from junior to senior in PCS. But fairly large variation is possible also, as noted by uours truly a couple of weeks ago in another PCS related thread:

"Out of curiosity, I took a look at Hanyu's track record from 2011-2 season onwards (that was his second in the seniors). He was the young upstart, not maybe yet particularly accomplished, but nevertheless a huge promise at the heels of Chan, Takahashi and the like. His PCS in SP varied from 36.07 to 38.31 and in FS 74.36-83.00. The next season his SP PCS rose to lower 40s. FS fairly similar to the previous year, 78.78-85.16. 2013-4 the GP series in the same figures until the final when he crosses the 90 point line - btw, the difference in his FS PCS score more than 10 points between TEB and GPF (so it IS possible to raise your PCS 10 points in 3 weeks after all?). Apart from the fall 2014 when he had the accident, has not really been scored lower than 90 (I must admit I did not check all the scores) and the top was hit of course at GP final 2015 with 49,14 in SP and 98,56 in FS.

Chan's career is long... In his first senior year 2007-8, he scored 33.15 and 77.20 in PCS. Won the 4CC in 2009 with 38+ and 80+ scores. By the time he was already World Champion in 2011-2012 he received PCS in the 40s for SP and in the upper 80s and lower 90s in the FS. TEB 2013 was the world record at the time with 46,18 in SP and 96,50 in FS. Since his return 40s and 90s consistently."

As for Fernández, his score history is interesting, because before he went to work with Brian Orser, his usual PCS values were below 7 and at best lower 7s. In 2010-1 season his PCS range of scores went down as low as 4.75. His top PCS scores in the spring of 2011 (under Morozov) were 33.39 at Worlds for SP and 70.50 at Europeans for FS. In a recent interview Brian said that Javi came to work in Toronto late July 2011 and had his first GP at Skate Canada late October. His PCS scores at SC were 39.11 in SO and 81.14 in FS - rise of 6 and 11 points! In 3 months... That is was a real change is perhaps indicated by the scores remaining at that level for the rest of the season. During 2012-3 season there was another (smaller) jump upwards to lower 40s and upper 80s. From then on a more steady rise. In Javi's case, it was a case of paying more attention to the components and probably general good quality in skating and working on them - and very good results were achieved very quickly indeed.

I would suspect that analysing more data, gradual development to higher PCS scores might be the norm for the top men at least last two Olympic cycles. Patrick probably achieved the high PCS levels ever so slightly quicker than Hanyu. Shoma started his senior career last year with lower 40s and upper 80s PCS scores, Chen's PCS so far in the same bracket. Of the youth squad only Boyang lags behind in this respect (for the record, IMO the difference is clear between Shoma/Nathan and Boyang).

E

 

[eppen]

Disregard the potential screaming but these are actually excellent resources
SP GPF 2016
http://chibura.tumblr.com/post/153879348150/original-post-in-russian-by-yulena
FS GPF 2016
http://chibura.tumblr.com/post/154115680890/transitions-in-mens-free-programs-at-gpf-2016

And she did this for last season too, so
SP at Boston
http://chibura.tumblr.com/post/143787559605/full-translation-of-the-whole-analysis-of-top-men
FS at Boston
http://chibura.tumblr.com/post/144143945205/full-translation-of-the-russian-article-analyses

We'll see the final one only post Helsinki but still, it's a good way to look at both.

Kudos for the effort for Yulena, but unfortunately she is not doing it exactly the same for both years. It is not possible to realiably compare the transitions situation pre and and post quad madness... (Plus I just wonder how she counted the transitions in Fernández's SP last year and this year - the program is exactly the same when it comes to transitions. Only the StSq was changed, but even that entry was kept from the old one.) Plus having Yin included would have been necessary. Overall, however, no apparent great changes from previous years apart from a slight increase in numbers for each guy - more crossovers for everyone, though. But I just wonder could this difference also result from the fact that Yulena counted the transitions differently?!?

In the ladies threads, the busy programs (by particularly the Russian skaters) have been also described as "cluttered" and the quantity is regarded as a negative aspect even. How about guys then? Is Hanyu's SP then approaching that ("just one big transition")? What is the ideal for transitions? Last season, Hanyu received an TR score of 9.63 at GPF and Fernández 9.61 at Worlds (both FS), both have received even a couple of 10s for TR...

E

 

[4everchan]

^^ sometimes less is more.... sometimes not... to me it depends how the program is built of course but also on the music. Doing transitions for the sake of it, just doesn't work. We want to see pure gliding on the ice as well. Also, very few skaters are able to keep their speed during transitions. So that's why I never quite appreciated these analysis. They are very interesting but to me, they do not automatically mean higher TR score for most transitions. I am not here judging any of the programs. Also, when I went to Nationals, there was a lady right below me who was practicing scoring. She had done the courses and had a bunch of judging sheets. You have to keep in mind that PCS where divided in five large boxes on the right, while on the left, there was a table to evaluate each one of the elements for GOE etc. So, when she would see very striking transitions, she would mark them down... she would use symbols for number etc... but there is no way as a judge that she was counting every single transition, every turn.... The mark for this was an "overall impression". There was just absolutely no time unless she would have been looking only at transitions to mark them all while judging GOE and other PCS components.... that's why I take such analysis with a grain of salt. Interesting but won't reflect directly in the judging.

 

[gkelly]

The current criteria for judging transitions are
Continuity of movements from one element ot another
Variety
Difficulty
Quality

"Quantity" per se is not a criterion, though it's hard to have much variety if there are very few transitional moves.

Just listing all the transitions can help keep track of the difficulty and variety, and where transitional moves connect directly to elements or elements connect directly to each other. (It could also be useful to keep track of examples of extreme lack of continuity/telegraphing.)

You'd need to rely on general some additional notation or else just overall impression to keep track of the quality.

I do think that listing and analyzing transitions can be useful for fans or trial judges as part of one's own process of learning to keep track of all the actual transitions in a program and the transitions criteria, and also for newer fans or trial judges to become aware of things to look for they hadn't considered previously.

But a list alone won't tell the whole story, and someone else's assessment of the continuity or quality won't substitute for your own perceptions. Or your own experience and muscle memory regarding difficulty, if you have been a skater yourself.