How it works

Everything the game is made of

The core loop: every card holds a complete equation. Two cards pair only when the whole equation is identical — 2 × 3 = 6 ↔ 2 × 3 = 6. The child never solves a blank. All cards share one fixed visual design; the only variable is the numbers, so nothing but memory of the equation can identify a pair.

The card — all states

Equation colour tokens

Fixed across the whole game. Never colour-code by times table — colour marks the role of a symbol, not the fact.

Type, spacing, radii

Fredoka 600 — display / equations

Nunito 800 — labels, stats, buttons

Nunito 600 — body & helper copy

Equation size by grid: 2-col 34px · 3-col 24px · 4-col 17px. Never below 16px.

Spacing scale 4 / 8 / 12 / 16 / 22 / 28. Screen padding 20–22px.

Radii: card 20 · tile 18–22 · sheet 32 · pill 999. Min tap target 48px.

On the web the board scales with its column, and the equation is sized as a fixed share of a card's width — so it grows on a laptop and still clears the 16px floor on a small phone.

Main interaction flow

Tap rules. The first tap flips a card and locks it face-up. The second tap flips the second card and locks input for the resolve window — 1.3s on a match (the celebration plays, then the pair fades out over 500ms) and 1.4s on a miss (a ±6° wobble, then both flip back). Taps during the lock are ignored, never queued — so a child mashing the board can't accidentally burn an attempt. Back exits to the level map; facts already practised are kept.

Adaptive engine

Each fact carries a weight. The deck builder samples its review slots by that weight, so missed facts return sooner and mastered facts drop back to a periodic refresh.

weight = max(0.25, 1 + misses × 2 − hits × 0.4)

A worked example — six facts, sorted by weight:

A lesson always seeds its own new facts first, then hands floor(pairs / 3) slots to the sampler for review. A brand-new player has nothing to review, so the first lesson deals exactly the deck it was designed with.

Micro-interactions

The table path

Every level is one times table (×2 to ×10, nine facts), walked as three lesson boards of 6 cards — three new facts each — and then a test. Since every card looks identical, only the equation distinguishes a pair. The test is three questions drawn by the adaptive sampler — so it asks about whatever has been going wrong — and two right passes it. Passing is the only thing that opens the next table, and it is where that table's trophy comes from. Speed Memory sits beside the path rather than on it: four rounds with the preview window shrinking from 2.5s to 1.0s, playable from Home at any time.

Stars. One star count, shown on the map, on Home and in the quiz header. A lesson pays three stars for pairs solved in at most one extra attempt, two within the pair count, one for anything else — and only on improvement, so replaying a lesson can't farm them. A quiz run pays on the same scale: three for all right, two from 75%, one from 50%.

Score is a different number, and it is only the Quack Quiz's: 100 a correct answer, up to 100 more for speed, multiplied by the streak. It buys nothing and unlocks nothing — it is there to be beaten. Stars are progress; score is the run you just had.

Everything the game is made of

See it in motion