Two options, both bad, and you have to pick one. Everyone votes, the split appears, and whoever ended up in the minority takes whatever your group decided the forfeit is. A dilemma only works when neither side is obviously better, so each card here was written to divide a table rather than to have a right answer. Thirty-one dilemmas across three heat levels, with the top tier behind an 18+ confirmation.
A tie has no minority on purpose. Forcing a loser out of an even split is arbitrary, and everyone can tell.
The vote is tallied in this browser from the names you entered; nothing is sent to a server.
No dilemma asks anyone to drink. The forfeit, if there is one, is whatever your group agreed before the first card.
The whole game rests on the vote being close. «Would you rather be rich or ill» is not a dilemma, it is a question with an answer, and the room finishes it in three seconds.
The reliable way to build one is to make the two options bad along different axes — time against money, comfort against status, knowing against not knowing. Then people cannot compare them directly, and the choice reveals which axis they weight more heavily.
That is also why a good dilemma keeps working after the vote. The argument is never really about the two options; it is about what the person values, which is a more interesting thing to hear at a table.
A nine-to-one split usually means the card is broken rather than that one person is strange. Note it and move on — arguing with the outlier is the least interesting version of this game.
Close splits are worth stopping on. Ask two people from opposite sides to explain in one sentence each; that is where the evening actually happens, and it costs less than a minute.
Over a whole deck, the pattern is the payoff. A group that consistently picks certainty over upside learns something about itself that no single card shows.
Mild dilemmas trade in preferences — food, weather, planning style, commute against pay. Nothing anybody says can be held against them later.
Hot dilemmas trade in information and social cost: reading the group chat where you were discussed, knowing what everyone earns, succeeding at the price of ten friendships. These are still hypothetical, but the answers say more about a person than the mild tier does.
The top tier is relationships, and it sits behind an age confirmation. The dilemmas there are the ones most likely to be about a real situation somebody at the table is currently in, which is exactly why a group should open that tier deliberately.
The common rule is that the minority drinks, which has a nice property: it punishes being unusual rather than being wrong, so nobody is defending a mistake.
It has an ugly failure mode too. One person with genuinely different taste ends up in the minority repeatedly, and by the tenth card the game is targeting them. Watch for it and switch to «everyone toasts» for a few rounds if you see it forming.
Playing with no forfeit at all costs nothing. The vote and the argument are the game; the drink is a house rule bolted onto it, and the deck is written so that removing it changes nothing.
The English dilemmas are written for this deck rather than translated. Several Korean cards turn on specifics — a wedding cash gift, a group chat culture, a commute norm — that need a footnote elsewhere, and a dilemma with a footnote is not a dilemma.
The card counts per heat level match the Korean deck exactly, so a full round is the same length and reaches the same intensity in either language. A test checks that rather than leaving it to memory.
Both decks avoid dilemmas that only work if you know somebody's financial or family situation. That is not squeamishness — those cards produce a vote that everyone at the table can already predict, which is the one thing a dilemma must not do.
Both options have to be unattractive in different ways. If one side is clearly better the vote lands nine to one and there is nothing to argue about. The best dilemmas split a table close to evenly and then produce five minutes of people defending a choice they are not sure about.
In the usual house rule, whoever voted with the minority. This page tallies the split and names them; it never instructs anyone to drink, and plenty of groups play with no forfeit at all.
Nothing. A tie is reported as having no minority, because picking a loser out of an even split is a coin toss dressed up as a result.
You can, but the vote stops being interesting — two people either agree or produce a permanent tie. Four or more is where the split starts to say something about the group.