Account

What new tool do you want?

Math & Education

Grade Curve Calculator

Curve a set of test scores — raise the top score to 100, add a flat bonus, or scale to a target average — and see each student's new score.

How to use

  1. 1Paste the class's raw scores.
  2. 2Choose a curving method and see the new scores and class average.

How it's calculated

Top-to-100: add (100 − highest) to everyone. Flat: add a fixed number. Target average: add (target − current average). Square-root curve: 10 × √score. Scores are capped at 100.

Frequently asked questions

Which curve is fairest?

Adding points preserves the gaps between students; the square-root curve helps low scores most. Choose based on your grading policy.

Is this a true bell curve?

No — assigning grades by a fixed distribution isn't common anymore. These methods adjust scores rather than force a distribution.

Can scores go above 100?

They're capped at 100 here.

The four methods help different students by different amounts

Raise-top-to-100 and flat-bonus both add the same fixed number of points to every score, so the gaps between students never change — someone who scored 20 points below the top still scores 20 points below the top after curving. The square-root curve works differently: it rewards low scores the most. A 64 becomes 10 × √64 = 80 (a 16-point gain), while an 81 becomes 10 × √81 = 90 (a 9-point gain) — the lower the original score, the bigger the boost, which is why the example pair's average rises from 72.5 to 85, not a flat amount.

What "curving to a target average" actually solves for

This method finds the single flat bonus that, if added to every score, would move the class average to wherever you specify, then applies it uniformly: bonus = target average − current average. It's really the same operation as the flat-bonus method, just with the bonus computed for you backward from a goal instead of chosen directly.

This isn't a bell curve in the statistical sense

"Curving on a bell curve" historically meant forcing grades into a fixed distribution — a set percentage of As, Bs, and so on regardless of actual performance. None of these four methods do that; they all adjust raw scores upward using a rule, while every student keeps their relative rank (except under the square-root method, which can compress the gap between a strong and weak score). Scores are capped at 100 either way, so a curve can't push anyone above a perfect score.

The square-root method assumes a percentage, and guards against a bad one

10 × √score only makes sense as a curve when the raw score is already a percentage out of 100 — feeding it a raw points total like '34 out of 40' would treat 34 as if it were 34%, producing a nonsensical curved score of 10 × √34 ≈ 58.3, lower than the original. The formula also clamps anything below zero to zero before taking the square root (Math.max(0, x)), purely as a safeguard against a stray negative number breaking the calculation — square roots of negative numbers aren't real, so without that clamp a single bad entry would produce 'NaN' instead of a usable curved score for that row.

Nothing is allowed past one hundred, and the square-root curve hides the points added

Every curved score is capped at 100, including a flat bonus that would otherwise push the top student over. Raising the top score to 100, which is the method that opens, adds whatever gap remains between the highest raw score and 100, and that same addition is applied to everyone. The flat bonus adds the amount I typed. Scaling to a target average adds the difference between that target and the current average, which can be negative if the class is already above the target. The square-root curve ignores that amount, replaces each score with ten times the square root, and then still caps at 100. On that method the points-added line is omitted, because there is no single addition to report.

Related tools

More: Math & Education