Your ROI goes up for every player in the tournament who is less good than you, and it goes down for every player who is better than you. This is why playing HU with a tough player is a bad idea, but you might play the same tournament if one or more weak players registered as well. The value of your skill advantage (or disadvantage) is a function of the blind and payout structures.

Your ROI is also a function of the rake, and you can basically treat travel expenses as extra rake, so you're right that it usually doesn't make sense to travel for small buy-in events. I usually only travel for tournaments when I expect to get other benefits from the trip as well, such as getting to see friends or travel in an interesting place. I think American pros who fly to Europe to sit in a casino for a week and then fly home are making a pretty big mistake.

Variance is a function of field size. I don't have an exact formula to give you, but you'll be bankrolled for a 100-person $200 MTT well before you are bankrolled for the Sunday Million.

If you're trying to make optimal game selection decisions, you need to take all of this stuff into account. Among other things, you can reduce the swinginess of your bankroll by playing a mix of small- and large-field events as opposed to just focusing on tournaments with huge first prizes and correspondingly huge fields, even if your ROI is higher in the latter.