Magic 8-Ball

The classic shake-and-flip toy turned into a browser widget. Each shake picks uniformly from a 20-answer pool (the same 10 positive, 5 neutral, 5 negative answers the Mattel toy uses) via crypto.getRandomValues. Swap in your own answer list for an in-joke version, or for decision-making with a controlled vocabulary. The last 20 questions and answers stay in history so you can argue about which way the ball was leaning today.

The pool is deliberately lopsided. Ten of the twenty answers are affirmative and only five are negative, so a fair draw still says yes half the time and no a quarter of the time. That is not a quirk of this implementation; it is how the toy has always been weighted, and it is a large part of why people remember it as encouraging. Replacing the list with your own options makes every entry equally likely, which is a different game.

Randomness comes from the browser's cryptographic generator rather than Math.random, so a run of five identical answers is genuine chance rather than a stuck sequence. Streaks like that are much more common than intuition allows: with twenty options, any specific answer repeating twice in a row happens once in twenty shakes. The history list exists partly so you can see that for yourself.

Each shake is independent. Asking the same question again does not refine anything, and the toy has no memory of what it said a moment ago, which is the difference between this and the tie-breaking tools that draw without replacement.

The genuinely useful version of this is the custom list. Replacing the twenty stock phrases with a real set of options turns it into a uniform picker with a fixed vocabulary, which is a reasonable way to break a deadlock when every choice is acceptable and nobody wants to be the one who decided.