Fair random choice guides

Set equal chances, freeze the roster, and record a fair random choice between several people.

A three-option tie needs three equal outcomes; a normal coin has only two. Use a die with three equal result groups, or repeat a coin rule that explicitly discards one outcome.

An online randomizer is fair for an equal-choice task when it gives every valid outcome the same probability, preserves the declared inputs, and does not silently turn a rejected result into a different rule. A pretty spinning wheel alone is not evidence of equal odds.

A browser random number generator is usually pseudorandom, not a physical source of unpredictable events. For a casual picker it can still produce useful independent-looking results; for security-sensitive values, use the browser Web Crypto API, whose getRandomValues method is designed to provide cryptographically strong random values.

Math.random() can be wrong for a random picker when the result protects something valuable, when an implementation maps its floating output unevenly into choices, or when observers need a stronger stated source. It is often adequate for a harmless animation, but MDN says it is not cryptographically secure.

Repeated results do not by themselves show that a random picker is broken. Independent draws allow repeats: with 4 equally likely options, the chance that the second draw matches the first is 1/4, or 25%, and a run of three identical results has probability 1/16, or 6.25%.

Use chance when the remaining options are acceptable, the stakes are low, affected people accept the rule, and a random result resolves a genuine tie. Do not use chance to avoid evaluating safety, consent, rights, expertise, or consequences that differ materially between options.

Use a decision wheel when the options are already equally acceptable and you need an impartial tie-break. Use a decision matrix when options differ on criteria that can be named and compared. A wheel distributes chance; a matrix exposes trade-offs before any tie is broken.

A tie-break feels fair when participants know the eligible options, the rule, the timing, and what happens after a result before the result is generated. Publish those four facts first, make one visible draw, and preserve the result instead of rerunning for a more comfortable outcome.

A yes-or-no wheel can help expose hesitation or settle a harmless binary tie, but it should not make an important decision for you. Use it only after both outcomes are safe and acceptable; if either answer would affect health, rights, money, safety, consent, or another person’s future, use information and responsible judgment instead.

A yes-or-no wheel is a fair 50/50 randomizer only when it has two equal-probability outcomes and its final selection is made by an unbiased random process. A spinning animation shows motion, not evidence that Yes and No each have probability 1/2.

A fair coin flip and a correctly configured yes-or-no wheel have the same one-trial odds: each named outcome has probability 1/2, or 50%. Their experience, verification method, and possible implementation failures differ, so equal mathematics does not mean they are interchangeable in every setting.

Five heads in a row has probability (1/2)^5 = 1/32, or 3.125%, when five coin flips are independent and each flip has a 50% chance of heads. After four heads, the fifth fair flip is still 50% heads; the earlier four results do not make tails due.

For n independent fair coin flips, the chance of exactly k heads is C(n,k) × (1/2)^n, where C(n,k) counts the different positions for k heads. For exactly 3 heads in 8 flips, the probability is C(8,3)/2^8 = 56/256 = 21.875%.

A best-of-three coin-toss contest needs two wins and a best-of-five needs three wins. If each toss is fair and independent, each side still has a 50% chance to win the series, but a longer series makes the final winner less dependent on a single toss and uses more tosses on average.

A fair coin-toss series for a tie-break names the two sides, uses an odd length or a first-to-wins target, fixes the stopping rule before toss 1, and records every toss. In a fair best-of-three series, the first side to two wins takes the tie-break and both sides start with probability 50%.

The sums from two fair six-sided dice are not equally likely because different sums have different numbers of ordered face pairs. Sum 7 has 6 pairs out of 36, so its probability is 1/6 or 16.67%; sum 2 has only (1,1), so its probability is 1/36 or 2.78%.

Rolling two six-sided dice produces 36 ordered outcomes, because the first die has 6 possible faces and the second die has 6 possible faces: 6 × 6 = 36. There are only 11 possible sums from 2 through 12, but sums are summaries of the 36 face pairs, not equally likely outcomes.

To calculate probability for custom dice, state every die’s faces and rule, count the equally likely ordered face combinations, then divide favourable combinations by all combinations. Two fair d8 dice have 8 × 8 = 64 ordered results; a total of 9 occurs in 8 of them, so its probability is 8/64 = 12.5%.

With a d20 success chance p on one roll, advantage—roll two d20s and keep the higher—has success chance 1 − (1 − p)²; disadvantage—keep the lower—has chance p². If a roll needs 15 or higher, p = 6/20 = 30%, advantage is 51%, and disadvantage is 9%.