The board is a model, not a poll
A trivia game that scores you on how popular your answer is has an obvious implementation problem: popularity is data, and data implies collecting it from players. There is another way to do it, and this game uses it — the board behind each question is prepared in advance rather than counted from the people playing right now. It models what most people would say about that category.
Two consequences follow, and they pull in opposite directions from what you might expect. First, the board is stable: the same answer scores the same all day, so you can compare your run with a friend's honestly. Second, the board cannot reward a genuinely surprising answer that nobody predicted, because if your valid answer is absent from the list it is refused rather than scored.
- Stable by design: the same answer carries the same band for the whole day, so scores are comparable.
- Pre-estimated: the board models what most people would say, it does not measure what today's players said.
- Bounded: a valid answer outside the accepted list is refused, not scored at an unknown band.
- Mode-dependent: the same five base scores are reordered per mode, so Crowd and Rare pay opposite ends of one board.
- Never mode-free: a Daily total always needs its mode attached to it before it can be compared with anything else.
That second consequence is the real cost of the design, and it deserves to be named rather than explained away. A live poll would occasionally pay out for an answer nobody had seen before, and a prepared board never will. What it buys in exchange is comparability: two runs from opposite sides of the world are scored against the same yardstick, and the crowdict board behind each question is built on exactly that assumption.
What the model leaves out
The interesting engineering is not the scoring table, it is deciding which answers belong in the list at all. Accepted spellings, alternative names, plurals, singulars and a forgiving rule for typos are all part of that list, and they are data rather than code — which means a new alias is a content change, not a release.
In practice the free browser version runs with both scoring modes, and its rules are written out on the site rather than left to be discovered by trial and error.
Top comments (0)