I don’t think there’s one rule that will work for everyone. It’s the job of the organizer and the TD to do what’s best for each specific situation, in most cases after consulting with the players.
Sometime it may make more sense to have some players face opponents more than once than to have them face players with large ratings differentials (which is often the case when pairing high score groups against much lower score groups.)
Keep in mind that the ratings formulas DO NOT CARE why players were paired against specific opponents, all they care about is trying to use those results to produce accurate ratings. The limitations on match rules exist because those tend to work against the goal of producing accurate ratings, and because we have a long history of attempts to manipulate the ratings system via matches.
When computers were slower a lot of people used an effort level of 5, but I was fairly fast with entry and had a decent printer so I wasn’t bothered by the extra time used with an effort level of 50. As computers get faster and faster, increasing the effort level doesn’t really add much for time. I don’t know how much of a difference there is between 50 and much higher numbers.
I don’t remember the specific effort level in that non-rated blitz tournament but I was expecting that it would eventually reach the point where rematches simply could not be avoided because I hadn’t tried to make it a round-robin directed Swiss. We had a fixed finishing time rather than a fixed number of rounds and the rounds were going quicker than anticipated.
I have run/assisted in 4 rd swisses with 6 players. You do the best you can. I’ve also been in a situation where the only legal pairing was an expert vs. a 900. I vetoed it.
Wasn’t there an FIDE Grand Prix…18 player 9 round swiss? What was done?
Unlike 6 players-4 rounds, with 18/9 there almost certainly will always be a legal pairing. Of course, you’ll be down to “pick a victim” pairings by around round seven, but if everyone is a world class GM, it’s hard to say that downfloating 2.5 points is going to be a mismatch.