2018 Internationaux de France Pairs Short Program

[DanseMacabre]

Dark figure skating gods, give me the protocols.

 

[Metis]

I don't think any of the skates today was anything worth standing up and waving in the k&c for, but I'm probably just bitter and cranky from all the skating today.

Rather than having a Cup of Epic Calamity, we’re having a bit of a Cup of Disaster at every event, aren’t we?

 

[DanseMacabre]

PRotocols: http://www.isuresults.com/results/season1819/gpfra2018/gpfra2018_Pairs_SP_Scores.pdf

B/K got a lvl 1 on the death spiral. And GOE from -2 to +2 on it. Hmm. J/C also got a lvl 1 on it. K/O got a B. Damn.

 

[Danny T]

... what happened :shocked:

 

[DanseMacabre]

... what happened :shocked:

The curse of the French GP.

Many skaters were sacrificed so that Jason Brown's short program could live.

 

[NoNameFace]

I think I just saw NoNameFace!!!

Each time I saw myself I got embarassed, but overall I enjoyed competition and practices today so much!

 

[Baron Vladimir]

USA and France judge really lowballed Boikova/Kozlovskii. All the lowest scores for this team (many of which system didn't take into the final results thankfully) came from them

 

[gold12345]

I don't understand all these elements with GOE ranges of -2 to +2. Why is there this confusion. If an element is average looking to begin with and not done error free, there really should be more negative GOE and not +2s.

There is even a -1 to +4, which just seems silly because James/Cipres certainly didn't do a -1 twist.

 

[Baron Vladimir]

I don't understand all these elements with GOE ranges of -2 to +2. Why is there this confusion. If an element is average looking to begin with and not done error free, there really should be more negative GOE and not +2s.

Not really. If an element has more positives than negatives it will end up with positive GOE. And in the +5/-5 system one bullet point worth one GOE so its easier to stay in plus comparing to the +3/-3 system (where for +1 you needed to hit 2 bullets). It is not easy to get +4/+5 tho and for big mistakes you can go in minus even more. The problem for the judges is that is hard to see everything/count in real time, so 2 judge's marks/scores per element/component haven't been taken (the system is expecting these mistakes) but average of 7 marks (from 9 judges).

 

[gold12345]

Not really. If an element has more positives than negatives it will end up with positive GOE. And in the +5/-5 system one bullet point worth one GOE so its easier to stay in plus comparing to the +3/-3 system (where for +1 you needed to hit 2 bullets). It is not easy to get +4/+5 tho and for big mistakes you can go in minus even more. The problem for the judges is that is hard to see everything/count in real time, so 2 judge's scores per element/component don't count (the system is expecting these mistakes) but average of 7 marks (from 9 judges).

I understand that it's easier to get positive GOE in the +5 system. And I don't really expect judges to be adding up bullet points in their heads for every element in real-time. I'm assuming the GOE awarded are more comparative and not just strictly following the bullets.

Regardless, it doesn't explain why ranges like -2 to +2 for an element happen so often. In some cases, yes, there is a grey area where an element is so-so and you aren't quite sure what it should score. But in some cases it's obvious that an element is flawed and/or weak and there shouldn't be a huge range in GOE.

 

[WeakAnkles]

The curse of the French GP.

Many skaters were sacrificed so that Jason Brown's short program could live.

And the problem with that is what?

:confused2:

 

[Baron Vladimir]

I understand that it's easier to get positive GOE in the +5 system. And I don't really expect judges to be adding up bullet points in their heads for every element in real-time. I'm assuming the GOE awarded are more comparative and not just strictly following the bullets.

Regardless, it doesn't explain why ranges like -2 to +2 for an element happen so often. In some cases, yes, there is a grey area where an element is so-so and you aren't quite sure what it should score. But in some cases it's obvious that an element is flawed and/or weak and there shouldn't be a huge range in GOE.

Its totally the same if you get from the judges three times -2,three times 2 and three times 0, comparing to nine 0s. Individual judges opinion is not matter, but average...And i dont understand what you mean to say with GOEs are comparative. Judges are not giving points by comparing the skaters, if you meant that. You can get the same score for a totally different reasons. (I mean you can get +5 and -5, and 9,5+ only in one way, but for +2 for an example or for 8 in components, there are a lot of different ways to get there)... But generally (if i understand you well) because now one bullet worth one GOE the difference between judges scoring/counting will be biger, cause the GOE scale they are using now is more sensible as a product of that - for the one (same) score/number they were giving in the +3/-3 scale, now there are two possible scores/numbers to give

 

[Metis]

I don't understand all these elements with GOE ranges of -2 to +2. Why is there this confusion. If an element is average looking to begin with and not done error free, there really should be more negative GOE and not +2s.

There is even a -1 to +4, which just seems silly because James/Cipres certainly didn't do a -1 twist.

Because the system doesn’t flag marks that are outside the corridor (or even just “REALLY aberrant, press YES to wipe hard drive and issue missile alert to Hawaii” levels of statistical deviance, legit data entry errors make it through and actually effect the average. I can’t remember which discipline and whether it was long or short, but, yeah, there was an obvious input error for one judge that (-2 to -4 split and he went with +4, when not missing the average ordinarily or otherwise overmarking the skater). I think that’s what’s going on with that weird -1 from J9, as their GOEs and PCS marks are not out of whack.

Outside of forcing final GOE to be a specific negative value (which is not often done — pages 15 and 16), any negative GOE bullets can be offset by positive features (or positive features, when combined with negative features, can end in the negatives). Which you probably know, but if not, cool. As for the -2 to +2 splits... yeah, those make my head hurt. I would like to see ISU implement some basic checks to keep protocols from being a mess:
1. If a judge attempts to input a GOE value that is 2 full points off the next closest GOE value, have Clippy appear and cheerfully ask, “Is that your final answer?” Maybe force a replay of the element in question. If the judge wants to submit a GOE well outside established values regardless, they can, but because it’s been automatically flagged and they were forced to tell the system “Yes, I Really Mean To Do This,” no one can claim happy/slippery fingers later on. Accountability!
2. If the GOE spans from a negative integer to a positive one or the GOE value range is greater than three full points, all judges are alerted that there’s a lack of consensus in the panel and the replay of the element is played. You can change or reenter the GOE value you originally selected after the replay. (This mechanism would not trigger if the range of values is over the allowable limit due to cases of +1,+2,+3 and a data entry error of -2.) So this would mean a 1,0,+1 is acceptable variance; -2,-1,0,+1 is not. 0,+1,+2 and a whacky +5 would also trigger a check. You could make the parameters for what causes the system to ask for a review much more complex and nuanced, but I chose 3 full marks because +4 versus +2 is something reasonable people can debate in good faith, but +5 versus +1 is ... bad optics, if nothing else.

You could also cause the system to just flag the aberrant judge(s) and alert them they’re out of consensus, but at that point, they can guess whether they’re too high or too low and manipulate accordingly. And while judges don’t have to agree, and frankly shouldn’t, on GOE, I think it’s not really credible to return protocols that show a judging panel that can decide whether an element was below average (-2,-2,-2,-1) or above average (+1,+1,+1,+2,+3). But I admit it would be very hard to implement real-time safeguards that don’t have the unintended effect of forcing more “corridor judging,” especially on PCS — but if you want scores to remain within a certain amount of acceptable deviation, you can’t just have a judges’ meeting after the event, as literal typos are just as valid to the average as any other input.

 

[gold12345]

1. If a judge attempts to input a GOE value that is 2 full points off the next closest GOE value, have Clippy appear and cheerfully ask, “Is that your final answer?” Maybe force a replay of the element in question. If the judge wants to submit a GOE well outside established values regardless, they can, but because it’s been automatically flagged and they were forced to tell the system “Yes, I Really Mean To Do This,” no one can claim happy/slippery fingers later on. Accountability!

Yes this would be nice.

But I admit it would be very hard to implement real-time safeguards that don’t have the unintended effect of forcing more “corridor judging,” especially on PCS — but if you want scores to remain within a certain amount of acceptable deviation, you can’t just have a judges’ meeting after the event, as literal typos are just as valid to the average as any other input.

Exactly. A certain amount is deviation is fine, but there is a problem if half the panel thinks an element was bad and the other half thinks it was good. Unlike PCS, things like element quality are not THAT subjective.

I've even seen an element with a fall marked as all -5s and one random +5 accidentally thrown in. There should be some way for the system to flag the typo so that an error like that doesn't embarrassingly show up on an official scoresheet.

 

[Ichatdelune]

First reaction to the results: :shocked: Trying to find videos of other teams, but whee for B/K No issues with the SBS jumps this time, and the spins were pretty synchronized as well! Death spiral could have been better (they didn't have a problem with it last time, wonder why this happened), but being ahead of J/C is still a feat. On J/C, was not expecting them to struggle like this. Major props to Vanessa James on saving that throw flip, but overall the jumps were not up to their usual standards. I have a feeling that the pairs did not do overall well, why did Youtube ban Brau Avita

Update: Watched the other teams, well that was not great. What was with the death spirals? Only R/K got a lv 3, the rest got even lower. Did SC use up the good pair skates? Anyway, my 200 won (approx. 0.176 cents) on the other teams:

- R/K did pretty well, if they don't bomb the free they are probably going to get a medal. Not really wowed though.

- K/O did better than at their last GP (can't remember which), but overall not impressed here. Last lift was nice.

- R/W are not good at twists, are they? I was actually scared while watching her get down, her leg and his face seemed so close. And a B in their death spiral. Ouch.

- L/M also got a B in their death spiral, ouch again. And also struggled with the twist. They have a nice program but so far they haven't done it to its potential.

- H/S, what a pity. Starting with a very shaky twist, and then falling that hard on the 3T... They did it lovely at SA (I think), was not expecting this much of a disaster.

 

[moonvine]

There are so few pairs. It was over in like 10 minutes! And all you have to do to podium is finish in the top half. That just seems wrong.