How it's calculated
flat: curved = raw + points; linear: curved = raw + (target − class mean); bell: curved = M + S × (raw − mean) ÷ sd
raw = original score, points = flat bonus, target/M = desired mean, S = desired spread, mean/sd = class mean and standard deviation. Curved scores are clamped to 0–100%.
Common questions
What does a square root curve do?
It adds the square root of the missing points, so a 64 becomes 72, which helps lower scores more than high ones.
Which method keeps the ranking of students?
All five preserve the order of scores, since each one applies a consistent rule, so a higher raw score always stays higher after the curve.