12 May 2026 · 6 min read

The Psychology Behind Random Choice Tools

Flipping a coin to decide something looks like the most low-effort piece of decision-making available. It isn't. The Danish writer Piet Hein put the trick more precisely than anyone else has:

Whenever you're called on to make up your mind,
and you're hampered by not having any,
the best way to solve the dilemma, you'll find,
is simply by spinning a penny.
No, not so that chance shall decide the affair
while you're passively standing there moping;
but the moment the penny is up in the air,
you suddenly know what you're hoping.

That's the whole point. The coin isn't the decision. The coin is the diagnostic.

This is what every random choice tool actually does, whether the user knows it or not. It externalises a question you've been carrying around internally and forces a reaction. The result isn't the answer. The way you respond to the result is the answer.

Why this works

Most of the choices that paralyse us aren't choices between things we don't have feelings about. They're choices where we have feelings we haven't admitted to yet, often because the feelings sit slightly against what we think we "should" want.

You spend twenty minutes pretending to consider whether to go to the gym. The hesitation isn't lack of information. You have all the information. The hesitation is that some part of you wants to skip it and some part of you thinks you shouldn't. So you stay frozen, performing deliberation to yourself, hoping for a tiebreaker.

A random tool gives you the tiebreaker without making it real yet. You hit the button, it says "yes, go to the gym," and one of two things happens. Either you feel relief, in which case the tool didn't really decide anything; you already wanted to go and were just looking for permission. Or you feel a small slump of disappointment, in which case you've just learned that you actually wanted the other option. Either way you now have information you didn't have thirty seconds ago.

Daniel Kahneman has written extensively about how our two systems of thinking interact. System 1 is fast, intuitive, emotional. System 2 is slow, deliberate, analytical. Decision paralysis is usually System 2 spinning on a question that System 1 already has an opinion on. The randomiser doesn't override System 2. It collapses its options so System 1 can be heard.

The regret problem

A second reason random tools work is regret aversion, which behavioural economists have spent decades documenting.

Humans are not symmetric about loss and gain. Losing £10 hurts about twice as much as finding £10 feels good. This is the famous Tversky and Kahneman finding from their 1979 paper on prospect theory, and it shows up everywhere once you start looking. The same asymmetry applies to choices. The pain of "I should have ordered the other thing" is bigger than the pleasure of "I'm glad I ordered this." Which is why we delay choosing, even when both options are perfectly fine. We're trying to dodge a regret we haven't experienced yet.

A randomiser removes this. If a coin tells you to order ramen, you can't really blame yourself for it. There's nobody to regret to. Some of the cognitive load that would otherwise have gone into avoiding regret simply doesn't need to be spent.

This is also why random tools work best for choices between options of roughly equal merit. If one option is genuinely worse, the tool doesn't help, because the regret of being told to take the worse one outweighs the comfort of having something else decide. People are sometimes surprised that the Yes or No generator feels useful but somehow doesn't help with the question "should I quit my job." That's the reason. The randomiser only frees you from regret when regret was the bottleneck. For big asymmetric decisions, the bottleneck is real, and you have to face it yourself.

When other people are involved

The third mechanism is one you've probably felt in a group. Four people, choosing where to eat, all saying "I don't mind, whatever you want." Nobody minds, but nobody is willing to commit either, because committing means owning the choice when it turns out to be wrong.

This is sometimes called diffusion of responsibility, a phenomenon Bibb Latané and John Darley first described in the late 1960s in the context of group inaction during emergencies. The dynamic in low-stakes choices is the same, just less dramatic. The group can't decide because nobody wants to be the one who decided. A random tool absorbs that responsibility. The coin chose, not you. Everyone can blame the coin, which is fine, because the coin doesn't have feelings.

This is the most common single use case for the custom picker. Four restaurants on a list, one tap, done. Nobody at the table feels obliged to defend the choice afterwards, because nobody made it.

Input randomness versus output randomness

There's a useful distinction in the creativity literature between input randomness and output randomness. They're not the same thing.

Output randomness is when you use a random tool to settle a choice you've already framed: pizza or ramen, gym or sofa, this film or that one. The tool just picks. You react. This is what most people use a randomiser for.

Input randomness is using a random tool to break out of a framing that isn't serving you. Brian Eno's Oblique Strategies cards, which he developed with the painter Peter Schmidt in 1975, are the classic example. You're stuck on a song. You draw a card that says "Honour thy error as a hidden intention." The card doesn't decide anything. It re-opens the question by introducing a constraint you didn't choose, which forces you to look at the problem from an angle you hadn't.

Random tools work better as output mechanisms than input mechanisms for most everyday use. But it's worth knowing the difference, because occasionally you'll find yourself stuck not because you can't choose between A and B but because you can't see beyond A and B. In that case what you need isn't a randomiser. It's something different. Talking to someone, going for a walk, doing literally any other task for ten minutes.

What the tools don't do

A random generator is not magical, and we should be clear about what it isn't. It will not:

Tell you what you actually want. It only reveals what you wanted by listening to your reaction.
Help with high-asymmetry decisions. If one choice is much better than the other, regret aversion isn't the bottleneck. Fear is. A coin flip won't dissolve it.
Replace thinking. The point of using a randomiser on small choices is to save thinking for choices that deserve it.

Used well, a random tool is a slightly faster version of the "imagine yourself a year from now, having taken option A; how do you feel?" exercise. It compresses that mental simulation into half a second of gut response. For the dozens of small-stakes choices that fill an ordinary day, that's enough.

If you want to try the test on something concrete right now, the Yes or No generator is the cleanest version. Think of a low-stakes question. Tap the button. Don't read the answer yet. Notice what you're hoping for before you look. That hope, not the result, is your answer.