I figure 5 = average is in the context of international-level competition. I.e., 5 is clearly senior level but nothing special beyond that. For younger skaters or those at lower skill levels, 5.00 could be an excellent score.
This document also relates the numbers to the percentage of the program time that the skater demonstrates each of the criteria.
I don't find it as useful to think of in those terms. One skater may be pretty good at performance or at interpretation for 100% of the program with no real breaks and no real highlights. Another might be excellent for 80% of the program and only average for the other 20%. How do you translate that difference into actual scores?
In theory, no. In practice, if skater A is a stronger all-around skater than B, then even when A has a bad day and B has a good day, A's score will and should still be higher. That's the system working correctly.Its pretty much on all relative and comparing a skater A against skater B. Meaning skater A will always score higher PCS mark regardless of how he/she performs in actual competition, and skater B will be always marked less.
If there is not much difference between them to begin with, then when A has a bad day and B a good one, B should score higher, at least in the areas where B does well and A does poorly.
Reputation effects might come in and prevent that from happening, which would be the system not working correctly.
The trick is to recognize the difference between occasions when A is still better and occasions where B deserved to score higher. How much do A's struggles on this occasion cancel out superior basic skills, superior program construction, etc.? Often it's a judgment call and different judges, different fans, would disagree with good reasons on both sides. Under 6.0 in those situations, we'd probably see mixed ordinals.