Rularoo

Rularoo starts from a famous psychology experiment. In Peter Wason's 2-4-6 task, people shown 2, 4, 6 and asked for the rule mostly guess something like add two each time, and every test they run to confirm it passes, because the real rule is broader. Rularoo turns that into a daily game. You see a sequence that fits a secret rule, and you may test any sequence you like; each probe comes back FITS or FAILS.

When you think you have it, you build the rule from tokens: what it is about (every term, every gap, the sum, the first term), how it behaves (is greater than, is even, is prime, is a perfect square) and what it is compared to. Your rule is judged on behaviour, not wording, so any rule that accepts and rejects exactly the same sequences counts. You have three commits and three hints per puzzle.

Each day has three puzzles. Easy uses three terms from 1 to 30 with a par of two to five probes; Medium uses three terms from -20 to 99, with negatives, sums and gaps; Hard has four terms and rules that combine two conditions, with par up to twelve. Par is what a perfect player would need, choosing at each step the probe that splits the remaining possibilities most evenly, so finishing under it beats the solver.

Rularoo is made by Digify Platforms and came out in September 2026, in English and Spanish. Today's puzzles and the archive are free to play in the browser, and the apps are free. Rularoo Pro adds unlimited practice, a Coach that compares your probes with the shortest line, streak insurance for one missed day a month and the complete archive; it never buys extra commits or hints.