How scoring works

Every game uses the same scoring curve so difficulty feels comparable across senses. Each round is worth 0–10 points; five rounds make a maximum of 50.

score = 10 × 0.5^(error ÷ half)

The curve

We measure your error in each game's natural perceptual unit, then convert it to points with score = 10 × 0.5^(error ÷ half). At zero error you get 10. At the game's "half-score error" you get exactly 5. At twice that, 2.5 — and so on, never dropping below 0.

Perceptual units

Errors are measured where perception actually lives: log2 of the duration or count ratio (Weber's law), ERB-rate for pitch, CIEDE2000 ΔE for colour, and shortest circular distance for angle. This is why a small mistake on a long duration costs the same as the same proportional mistake on a short one.

What's a good score?

Calibrated so a first-timer lands around 29/50 and an expert reaches the high 30s to low 40s. A clean 45+ is genuinely rare.

Fairness

Targets are generated deterministically from a seed. The daily challenge uses the same seed for everyone on a given day, and challenge links replay an exact set of targets — so the leaderboard compares like with like.