The best questions force a real trade-off
"Would you rather always be 10 minutes late or always be 20 minutes early" works because both options genuinely cost something. "Would you rather have a million dollars or lose a finger" does not really work as a trade-off, since one option is obviously better for almost everyone.
Matching questions to your group
Silly, absurd questions work well as icebreakers with new people or with kids. Deeper, values-based questions land better with close friends, or later in a conversation once people feel comfortable sharing more of themselves.
Frequently Asked Questions
How many players does the game need?
It works with as few as two people and scales well to large groups — a bigger group actually adds to the fun, since you get to compare a wider range of answers.
Should there be a "right" answer?
No. The point is exploring reasoning and preference, not correctness. A question with one obviously correct answer for nearly everyone is not a well-designed one.