Coin Flip

You can't decide between two equally good options and you've been deliberating for forty minutes. Click once. Heads or tails. The decision is made — and if you find yourself disappointed in the result, that disappointment tells you which option you actually wanted.

The flip uses cryptographically random output, so it's genuinely 50/50. No skewed weighting, no patterns over time. Visually animated because watching a coin spin briefly feels more like a real flip than seeing the answer pop up with no ceremony.

For decisions that actually matter, randomness is a poor proxy for thinking. For minor stuff — restaurant choice, who goes first, settling an unproductive argument — it works fine and saves time.

Underneath the animation, the flip uses `crypto.getRandomValues` to pull a random byte and tests whether the first bit is 0 or 1. That bit is genuinely uniformly random — the probability of either outcome is exactly 0.5, with no detectable bias across millions of flips. Real physical coins have measurable bias — an often-cited Stanford study found that flipped coins land on whichever face was up when flipped about 51% of the time, due to subtle physics of the tossing motion. Software flips don't have that bias.

A worked example for probability illustration: flip 100 times. You'd expect 50 heads on average. The actual result might be 47 or 53 — that's normal random variation. If you got exactly 50 heads every time, that would suggest something was wrong with the randomness. The standard deviation of 100 fair flips is about 5, so most batches of 100 will land between 45 and 55 heads. Getting 30 or fewer heads in 100 flips would happen by chance roughly once in a million tries; if you saw it from this generator, the randomness would be suspect.

The psychological purpose of the animation matters even when the result is computed before the animation runs. Watching a coin spin briefly creates a moment of suspense that mirrors the real-world experience. Without the animation, the result feels arbitrary — instant, anticlimactic, unconvincing. With it, the brain registers 'a choice has been made by an external process,' which is the actual point of using randomness for decisions you keep flip-flopping on. The coin doesn't decide for you; the coin lets you observe your own gut reaction to the result, which often tells you what you actually wanted.

Where coin flips are useful: trivial decisions where either option is genuinely fine, breaking ties in casual settings, settling unproductive arguments by removing the disagreement from human politics into mechanical chance, randomizing turn order. Where they're not: anything that matters financially, anything where the outcome's distribution of consequences is asymmetric (a 'small loss vs huge gain' bet shouldn't be a coin flip — it should be analyzed). Randomness is a poor proxy for decision-making in any case where the cost of being wrong on one side outweighs the cost of deciding more carefully.