The Random Decision-Maker Quiz

You walked into my office around seven with that look — the look of a person holding a free evening and no idea what to do with it. Happens to the best of them. The fridge is a suspect, the couch has an alibi, and your group chat went quiet an hour ago.

Here is the deal: answer seven questions about your mood, honestly, and I hand you one clean verdict on how tonight goes down. No second-guessing, no appeals. You came here because deciding is hard. Lucky for you, deciding is my whole racket.

Result	Patterns	Share of patterns
Verdict: Cook Something New	3,850	23.50%
Verdict: Call an Old Friend	1,147	7.00%
Verdict: Start the Series	4,911	29.97%
Verdict: Go Outside	2,688	16.41%
Verdict: Do Absolutely Nothing	3,788	23.12%

Athens ran its government on a stone lottery

Classical Athens filled most public offices by lot rather than by vote, on the reasoning that elections favour the rich and the well-known. Candidates were sorted using a kleroterion: a slab of stone cut with columns of narrow slots, into which identity tokens were slid, with a tube down one side holding a jumbled mix of black and white balls. Releasing the balls one at a time accepted or rejected an entire row of candidates at once.

Fragments of these slabs survive from the Agora, which is unusually hard evidence for something so procedural. Juries of several hundred were assembled this way on the morning they sat, to make bribery impractical.

Knucklebones came first

Long before cubic dice, people cast astragali, the ankle bones of sheep and goats, which come to rest on four distinguishable faces. They are not fair; the faces differ in area and in odds, and ancient scoring conventions reflected that. Drawing lots to settle a question turns up across the ancient world in law, in land division and in religion, usually on the understanding that the outcome was answerable to something.

Venice pushed the principle further. For centuries its doge was chosen through ten alternating rounds of lottery and ballot, arranged so that no faction could plan far enough ahead to rig the result.

Nobody can produce randomness by hand

Ask people to write out a random-looking string of heads and tails and the result is recognisably manufactured. They switch too often, dodge long runs and grow uncomfortable after three or four repeats, even though runs of five or six are ordinary in a genuine sequence. It is one of the more dependable findings about human intuition and chance.

Which is why serious randomness gets outsourced. In 1955 the RAND Corporation published a book containing nothing but a million random digits from an electronic noise source. The raw output carried a slight bias and had to be corrected before printing.

What the machine in your pocket is doing

Most software randomness is not random. A pseudorandom generator runs a fixed formula forward from a seed, emitting a stream that passes statistical tests while staying perfectly repeatable for anyone who knows where it began. For choosing between noodles and a walk that is plenty. For anything involving money or keys, machines fall back on physical noise such as timing jitter, because being predictable is the entire problem.

This page sits closer to the coin than the cryptography, and deliberately so. The useful part of a random verdict is often not the verdict; it is noticing whether you feel relieved or disappointed by it.