Would You Rather Generator

Returns a randomly selected 'would you rather' dilemma from a curated set of paired choices. Categories include trivial preferences, ethical dilemmas, hypothetical scenarios, and physical or social trade-offs.

The game format dates to mid-20th-century American party games and serves as a low-stakes tool for surfacing values, biases, and conversation. Output is intended as a conversation prompt, not a serious decision aid.

The format's structural design forces commitment. By presenting two and only two options, both with non-trivial costs, the dilemma defeats the polite non-engagement that open-ended questions invite. Asked "what's your favorite food," a respondent might say "I like everything." Asked "would you rather only eat pizza or only eat sushi for a year," they must choose. The forced choice surfaces preferences that would otherwise stay hidden behind politeness.

The corpus design follows a few principles. Questions where 80%+ of respondents pick the same option are filtered out as boring. Questions that require specialized knowledge (e.g., "would you rather have full read access to the FAA database or full write access") are categorized as niche. Questions that map cleanly onto political or moral identity are flagged because they tend to harden positions rather than open conversation. The bulk of the corpus aims for genuine ambivalence — choices where thoughtful people disagree.

A worked example output: "Would you rather always know when someone is lying, or always be able to get away with lying?" This question splits respondents fairly evenly. The "always know" choice appeals to truth-seekers but produces social isolation (you can't unknow). The "get away with" choice grants a power most consciences would refuse. Discussing the choice tends to reveal the respondent's relationship with social trust and personal ethics in a way that no direct question would.

Aggregate response patterns across thousands of players reveal demographic clustering. Younger respondents pick novelty over comfort more often than older respondents. Certain ethical dilemmas (the trolley problem variants, "save five strangers vs. save one family member") cluster strongly by political identity and education. Comfort-vs-experience trade-offs ("a year living in a tropical paradise vs. a year traveling difficult places") split by personality dimensions like openness and neuroticism. The patterns make the format useful for cultural research as well as parlor games.

Limitations: the format works for opening conversation, not for resolving real decisions. Real life rarely offers clean dichotomies; the trade-offs that matter typically involve probabilities, partial outcomes, and hidden information. The dilemmas in the corpus are simplified for engagement, not realism. Treating "would you rather" answers as predictive of behavior in actual analogous situations is a known fallacy in social psychology — stated preferences and revealed preferences diverge consistently.