Yeah, unlike almost every other possible figure in the book, there are no clearly defined criteria for exactly how to make deductions on a loop. Everyone has to adapt their own method, being consistent with everyone as possible. I can attest that being a line judge is hard work, not only to see the flight in minute detail, but to translate what you see in that figure (often flown at 3000' AGL) into a fair score. Yes, there's significant judging variation, but the system and software is designed to mitigate that. Just like pilots, judges are human and make mistakes. But IMO at all the contests I've participated in, the pilot who flies the best and/or makes the fewest technical (non-subjective) mistakes is the one who ends up placing first in their category. It generally all works out fairly.