Everything else on this site is about deciding quickly. This page is the counterweight. A coin is the right instrument for a choice whose cost of deliberation exceeds its cost of error, and the wrong one for everything else. Telling the two apart is most of the skill.
First, sort the decision
Before choosing a method, establish which kind of decision you have. Three questions, in order:
- Is it reversible? If walking it back costs little, speed is worth more than accuracy.
- Does it affect anyone else materially? Money, health, time, safety. If yes, you owe them a reason, and chance cannot supply one.
- Is it repeated? A choice you make weekly deserves a rule, not a decision.
Reversible, private, one-off, and genuinely close: flip a coin. Anything else, use one of the following.
1. The weighted matrix
For decisions with several options and several criteria that are not equally important. Choosing a flat, a supplier, a laptop.
List your criteria and assign each a weight out of ten reflecting how much it actually matters. This step is the whole exercise, and it should be done before you look at the options, or the weights will bend to fit the answer you want. Score each option out of ten on each criterion, multiply by the weights, and total.
The output is not the answer. The output is a number you can argue with, and the argument is where the value is. When the total contradicts your instinct, you have learnt something: either a weight is wrong, or a criterion is missing, or your instinct was wrong. All three are worth knowing.
2. Regret minimisation
For decisions about direction rather than options: whether to leave a job, move city, start something.
Project forward to an age at which you are old enough to be honest with yourself, and ask which choice you would regret not having made. Jeff Bezos used this to decide whether to leave a secure position to start a bookshop, and reported that framed that way the decision became immediately obvious.
It works because it strips out the two forces that dominate short-term thinking: fear of immediate loss, and the pull of the status quo. Neither survives the projection. What remains tends to be what you actually value.
Its weakness is that it systematically favours action, because omissions loom larger in hindsight than mistakes do. Use it for decisions about whether to try something, not for decisions about how much risk to accept.
3. 10/10/10
For decisions distorted by present emotion: an argument, a resignation, a purchase you are talking yourself into.
Suzy Welch's version: how will I feel about this in ten minutes, in ten months, and in ten years? The three horizons expose the shape of the feeling. Something that matters enormously in ten minutes and not at all in ten months is an emotional reaction, not a decision. Something that barely registers now but changes the ten-year answer is the opposite, and is usually being under-weighted.
It takes about a minute and is unusually good at defusing the decision you are about to make while angry.
4. The pre-mortem
For plans you are already committed to and want to stress-test.
Gary Klein's technique: assume it is a year later and the decision has failed badly. Then write the story of how. Not a list of risks, but a narrative, written in the past tense, explaining what went wrong.
The reframing matters. Asked to list risks, people are cautious and vague, because raising doubts about an agreed plan feels disloyal. Told the failure has already happened, the same people become specific and blunt, and the specific version is the useful one. Klein's studies found the imagined-certainty framing produced substantially more concrete failure modes than a conventional risk review.
When to use chance, precisely
Randomness is legitimate in three cases, and they are narrow.
- The options are genuinely equivalent. You have done the work and cannot separate them. That is a finding: the difference is below your measurement threshold, and further analysis produces delay rather than accuracy.
- Fairness is the point. Who goes first, who gets the last seat. What matters is not the outcome but that no party selected it. This is why coin tosses are used in sport and, occasionally, in law.
- You need the reaction, not the answer. Flip, and notice whether you are relieved or dismayed. This is a diagnostic rather than a decision, and it is covered in full in the overthinking guide.
Outside those three, a random answer is an abdication dressed as decisiveness. The tell is simple: if you would be uncomfortable telling an affected party how the decision was made, the method was wrong.
Matching the method to the stake
The frameworks above take between one minute and an afternoon. That is the right budget for decisions that are hard to reverse or that other people depend on. For everything else, which is most of your day, the correct amount of deliberation is close to none, and the tools on this site exist to make that fast.
Spending an hour on a lunch order and ten minutes on a job offer is the common failure, and it is a failure of sorting rather than of thinking. Sort first, then choose the method.