Not cEV. I don’t know enough about the other options to say, but you’re in a bubble situation, so cEV is not going to be a good guide. I’d probably just do ICM, but I don’t have HRC in front of me to remind me what the other options are.

The Riceman said
Another scenario: 18man Turbo. 4 places paid, 5 players left. Presumably I am using ICM here, and presumably the software will take the exact money bubble in to consideration right?  

Yes. That’s what ICM is. What it won’t take into account is how future hands will play out. ICM is not a perfect measure of your stack’s value because it doesn’t look at things like where the blinds are and what opportunities will be available to you in future hands. So ICM somewhat undervalues big stacks because it doesn’t take into account their ability to accumulate chips through pressuring other stacks. HRC has an option called Future Game Simulation or something like that enables it to simulate a few hands into the future and more accurately value stacks. It’s processor-intensive, though, can’t do more than a few hands at most.

There’s no bright line. ICM is always a slightly better calculation, it’s just a question of whether it’s worth the extra effort. Setting up ICM for more than a couple of tables is a lot of work. Mostly you want to use it for situations where ICM is a major factor, mostly the bubble or approaching the final table.