If the 80% rule is right for practice, then I think it should also apply to your competition results. In other words, if you aren't Q'ing 80% of the time, then your team isn't ready to compete. Time to go back home and drill your skills -- become a real team. If done correctly, it shouldn't take very long to see big improvements.
The 80% rule states that when D is completing an exercise correctly 80% of the time, it's time to move on to the next exercise.
The only way to tell this for sure is to keep accurate training records. You put 10 treats in your pocket, reward at the end of each correct performance until 10 treats are gone. If that took 10 tries, you are operating at 100%, if 12 tries, 80%, if 15 tries, 66%, etc. At 80%, D knows it well enough to move on, at least according to some literature I've read and Elaine Coupe's Agility Seminar I took last year. I don't know whose idea this was originally.
I don't understand why some competitors object to repetition of a course or course segment. How will you reward correct choices and discourage wrong choices or innacurate performances if you don't stop when they happen, go back, repeat, and reward only when they are right? Maybe not 10 times at this level, but at least 3 correct performances in a row.
I've heard it said dogs can memorize a course within 3 turns and do it right from then on, so their correctness on repeated tries doesn't truly test the team's cohesiveness. So what? How else but by repetition can a team learn to work well together, feel each other's field, read each other's body language, develop muscle memory?
I'm sticking with my guns -- repetition, repetition, repetition. -- until you've covered every possible combination of moves multiple, multiple, multiple times. Only at that point does Q'ing become more than just a happy accident.
Then again, what do I know.