Written, yes. Matter of opinion? Well, here is the complete list for levels for step sequences.
<skip>
The rule is, if you do two of these four things you get a level 2, three for a level 3, and all four for a level 4 -- although #1 and #3 are not an either-or like 2 and 4.
1. Does your step sequence have no variety to speak of (level 1), "simple variety" (counts twoward level 2), "variety" (counts toward level 3) or "complexity" (counts toward level 4)?
#1 is pretty much either-or. There are specific clarifications for what consitutes "simple variety" (at least 6 turns and 4 steps, none of the types can be counted more than twice); "variety" (at least 8 turns and 4 steps, none of the types can be counted more than twice); and "complexity (at least 5 different types of turns and at least 3 different types of steps all executed at least once in both directions).
http://www.usfigureskating.org/content/Singles Difficulty Levels.pdf
The definitions have changed a bit from year to year, but within any given season it's pretty much an either-or question whether a step sequence fulfills one of those definitions or not.
Types of "turns" are currently defined as three turns, twizzles, brackets, loops, counters, rockers, which pretty much covers all possible turns on one foot
Types of "steps" listed are toe steps, chasses, mohawks, choctaws, change of edge, cross rolls -- other steps are possible, but if they're not listed in the documentation there might be some subjectiveness involved in whether to count them.
Because fewer "steps" are required and there are more possibilities available, if a step sequence fails to qualify for this feature it's a lot more likely to be for not enough types of turns than not enough types of steps.
It is hard to identify and count all the turns and steps in real time. I understand that often (usually?) one member of the tech panel is assigned to do just that and another member keeps track of the other features.
Occasionally they might count wrong or identify a turn wrong, but that would be a mistake, not an application of subjective opinion. They might identify a turn or step incorrectly. The skater may perform a turn or step in such a way that it's ambiguous what edge s/he was on and therefore what the turn should be called. In those cases the panel would review the sequence on video replay except in very low-tech events where it isn't available.
As long as the edges are executed cleanly as intended, it's just a matter of identifying and counting, which is about as objective as you can get.
3. Moderate (for levels 2 and 3) to full (level 4) use of upper body movement. (This is the one that is currently the most abused, with flailing arms and bobbing up and down seeming to count for "upper body movement").
The difference between "modest" and "full" upper body movement has an either-or definition -- modest requires visible use of two out of three of arms, head, and torso; full requires all three.
http://www.usfigureskating.org/content/First Aid Singles 2008-09.pdf
There might be some borderline cases where one or two of those body parts are used so slightly that it's debatable whether they were used enough to count toward the feature or not. That's where some subjective judgment of the tech panel would come in. The reason why skaters make such big movements of their arms and torso is to make sure that that there's no question they used them.
4. Immediate changes of direction by rockers, counters, etc.
There probably could be some room for subjective interpretation as to whether a change of direction was "immediate" or not -- e.g., is 1 second between turns quick enough, or does it need to be 1/2 a second?
Right about the tech panel, but it is hard to tell what does or does not influence the judges. They might be more likely to give out higher PCSs, for instance, if the skater is awarded a lot of level fours by the tech panel.
The judges don't know what levels were awarded by the tech panel when they're giving marks. Back in the first year of the new system they did get that information, but because it was found to influence them the process would changed so that they wouldn't.
The folks who write the ISU rules know what they want. They want Alexei Yagudin's step sequence in Winter. So they write the rules accordingly.
If that's what they were trying to achieve, they failed dismally in the attempt to write the rules accordingly.
Under the current rules, both of Yagudin's step sequences in Winter would be level 1, because they don't meet the definition for even simple variety, and that first feature is now mandatory for any higher level step sequence. Even if the skater does features 2, 3, and 4 well beyond the minimum required to earn those features, it will still be level 1 without the necessary variety of steps and turns.
That wasn't true when the level rules were first written. It's possible that the Winter step sequences would qualify for level 2 or 3 under the 2004 rules (there were only 3 levels at that time). But even then the original version of the steps and turns feature was listed first and seemed to be the most important.
If we believe that the folks who write the ISU rules know what they want, then based on the rules they wrote I'd have to guess that the most important thing they want in step sequences is brackets, rockers, and counters.
) In fact, I would hope that the judges would tell me, thank you very much but don't do that any more.
