I don't know if you're falling for any bias but it's alas true that judges don't really judge only what happens on the ice. (Costume is a small part of the Composition score
Costume is not mentioned anywhere in the Composition criteria.
Or anywhere else in the scoring rules aside from the deduction (given if a majority of the judges+referee hit the deduction button) for a costume that doesn't cover 50% of the upper body, not counting illusion fabric.
There's a separate deduction if part of the costume comes off and falls on the ice.
The Composition criteria are:
"The intentional/developed and/or original arrangement of the repertoire of all types of movements into a meaningful whole according to the principles of proportion, unity, space, pattern and musical structure.
*Multidimensional movements and use of space
*Connections between and within the elements
*Choreography reflecting musical phrase and form
*Pattern and ice coverage
*Unity"
Nothing about costumes.
A costume might serve as a signal to the judges about what the intended meaning is behind the meaningful whole, so a costume that points judges' thinking in the right direction is likely to result in somewhat higher Composition scores than a costume that completely confuses them and seems to have nothing to do with the music or the choreography.
But a completely neutral costume can support most programs just as well as an overtly literal translation of the music source.
And sometimes a skater might show up at a competition but their costume was lost in transit or suffered some mishap that rendered it unwearable, so the skater had to compete in practice clothes or the other program's costume or an unrelated costume borrowed from another competitor. There's no penalty for that.
but it seems visible that judges rarely separate the Components, even now that there are two less, they give a global Presentation score and share it rather evenly in the three Components scores boxes.)
I would like to see wider ranges for skaters whose skills are less well balanced among the separate components.
(Many skaters' skills are pretty similar among the three components, so in those cases I do expect the numbers to be close to each other.)
Can we seperate one from another? We talk about how things *should* be weighted or scored because the way they are or have been for the past years.
Imagine if a complete newbie or someone who just has a passing interest in figure skating is watching. How do you explain that a skater that *should* be getting 6s is getting 9s , while a skater who should be getting 9s is getting 7s?
How do you determine what score the skater "should" be getting?
If I'm watching at home on video and disagree with the judges, my first thought is that something came across live that I'm missing on video.
Even if I'm watching in the arena and that I would have scored a given skater higher on one component or lower on another, I attribute the difference of opinion to different priorities of what to reward most among that component's criteria. Judges on the panel may disagree with each other as well, in which case at least one of them might agree with me.
If they all agree with each other and I'm the outlier, I figure I'm the one who's missing something.
We might in general believe that judges should be less afraid to use wider ranges and to give 8s and 7s rather than 9s for performances full of quads and not much else. But even if you and I both agree we'd like to see that happen, we probably don't agree on
exactly what score to give a given component for a given performance. It's not an exact science.
On GOEs there are sometimes right and wrong answers, but even there it's more a matter not of right and wrong but rather better and worse. For PCS, it's all better and worse. There is no single correct answer.