When Should You Use Chance to Make a Decision?
A random decision is a decision rule that selects among pre-approved options by chance rather than by a ranking criterion.
A worked example with numbers
Two friends have already checked that three cafés are open, equally affordable, accessible, and suitable for dietary needs. They cannot agree on a preference. A three-way draw is reasonable because each outcome meets the non-random constraints. By contrast, selecting a repair contractor, a medical treatment, or an employee at random would ignore differences that need evidence and judgment. Chance is a tie-breaker after evaluation, not a substitute for evaluation.
How to set the rule before the result
Use a four-question screen. Are all options safe and permissible? Has each affected person had a fair chance to raise a constraint? Would you accept any outcome before seeing it? Can you state a single outcome rule without exceptions? If all four answers are yes, use an equal decision wheel, a coin flip, or a list picker that matches the number of options. If the answers differ because one option has higher value, use criteria rather than pretending the difference does not exist.
Common mistakes that change the odds or the process
The biggest mistake is asking chance to decide a question whose wording hides a real choice. “Should I quit my job?” bundles finances, notice, health, skills, and alternatives into a yes/no output. Split that into information questions and a decision process. Another mistake is repeatedly drawing because the answer feels unwelcome; the discomfort may be useful evidence about your preference, but the second click does not make the first rule wiser. A third mistake is imposing a draw on someone who did not consent to the eligible list or the consequence.
Where this method stops being appropriate
A random method fits a classroom speaking order, which film to watch after content restrictions, a game starting player, or a fair rotation among already-qualified volunteers. It does not fit legal entitlement, emergency response, medical triage, education eligibility, lending, hiring, discipline, safety assignments, or gambling. In those settings, the relevant fairness question is usually not equal probability; it is whether a lawful and reasoned criterion was applied and documented.
How the random source fits into the rule
The tool should match the object. A coin gives two outcomes of 1/2 each; a six-face die can make three equal groups of two faces; a list shuffle makes an order without repeats. The browser randomness described by MDN’s Crypto.getRandomValues reference can make an informal choice hard to predict, but unpredictability alone does not make a decision ethical or defensible. Record the rule and the result when other people need to understand why the tie was resolved that way.
Use randomness after judgment, not instead of it
Chance works best at the final narrow point of a decision. First remove options that violate a constraint. Then compare options whose effects differ. Only then ask whether the survivors are tied enough that equal allocation is honest. This sequence prevents a wheel from hiding a safety issue behind a colourful result. It also makes disagreement productive: if someone objects before the draw, identify the missing condition rather than arguing about the next spin.
Consider reversibility and consent as well as size of consequence. Picking a board game can be reversed next round. Assigning an unwanted night shift, even if low paid or informal, may need availability and rotation rules. A random rule is legitimate only when people know it applies to them and can understand its consequence. If a draw reveals that participants never accepted one option, treat that as information about the process, not as a reason to insist that probability has settled the matter.