Grade Curve Calculator
Five ways to curve a test, with before-and-after statistics, a histogram and a CSV.
Class scores
Curve
Every score is multiplied by maximum ÷ highest score, so the top score becomes the maximum.
curved = √(score × maximum) — 10 × √score on a test out of 100. Low scores gain most; 0 and the maximum stay.
Before and after
| Statistic | Before | After |
|---|
Distribution
Score as a % of the maximum; each pair of bars covers ten points (60 = 60–69%).
Show the distribution as a table
| % of maximum | Before | After |
|---|
Letter grades
Every student
About the Grade Curve Calculator
Paste a class’s scores — one per line, with names if you like — and curve them by one of five methods: add points to everyone, scale to the highest score, rescale to a target mean and standard deviation (the z-score “bell curve”), the square-root curve (10 × √score) or assign letter grades by rank (the top 10% get an A, and so on). The calculator shows the class before and after — mean, median, standard deviation, lowest and highest — a histogram of both, letter grades before and after, and every student’s new score.
Options keep a curve fair to every student: cap scores at the maximum, never lower anyone’s score, and round to whole marks or decimals. Download the result as a CSV for your gradebook. Everything is worked out in your browser; the class list is never uploaded.
How to use it
- Paste or type the scores, one per line — a name or ID may come first (“Asha, 72” or a two-column copy from a spreadsheet) — or open a CSV or text file. Set the test’s maximum score.
- Pick the curve: points to add, scale to the highest, the target average and standard deviation, the square root, or the share of the class for each letter.
- Choose the options: cap at the maximum, never lower a score, rounding, and the letter scale used to show grades before and after.
- Compare the statistics and the histogram, check the student table, then Download CSV or Copy summary.
Examples
Add 8 points
Average 68.8 → 76.8; the top score 92 becomes 100; letters A–F go from 1/1/3/3/2 to 2/2/3/2/1
Same class, curved = score × 100 ÷ 92
45 → 48.9, 83 → 90.2, 92 → 100
Same class (mean 68.8, SD 13.48)
45 → 57.3, 83 → 85.5, 92 → 92.2: 75 + 10 × (score − 68.8) ÷ 13.48
49, 64, 81 out of 100
70, 80, 90: 10 × √score
Same class, A 10% · B 20% · C 40% · D 20% · F 10%
A: 92 · B: 83, 78 · C: 75 down to 64 · D: 60, 52 · F: 45
Common uses
- Lifting a test that turned out harder than intended, without changing the order of the class.
- Putting several sections’ results on the same average and spread before combining them.
- Checking what letter grades a curve would give before announcing it.
The five curves
- Add points (flat): curved = score + k. Everyone gains the same; the spread stays the same. A common choice is the gap between the top score and the maximum.
- Scale to the highest: curved = score × maximum ÷ highest score. The top score becomes the maximum and everyone gains in proportion to their score (a 46 gains half as much as a 92).
- Target mean and SD: curved = target mean + target SD × (score − class mean) ÷ class SD. Each score keeps its z-score — its distance from the average in standard deviations — on a scale with the average and spread you choose.
- Square root: curved = √(score × maximum), which is 10 × √score out of 100. Low scores gain most (36 → 60, 64 → 80), 0 and the maximum stay where they are, and the order of the class never changes.
- Letter bands by rank: the scores are ranked, the top share gets the first letter, the next share the second, and so on. Tied scores always share a letter, so a band can hold more students than its share. Scores are not changed.
Keeping a curve fair
Never lower a score keeps every student’s original score whenever a curve would reduce it — a target-mean curve lowers scores when the class is already above the target, and a narrower target SD pulls the top scores down. Cap at the maximum stops a flat or z-score curve from giving more than 100%. Both are on by default; each result says how many scores they touched. Because capping and keeping change some scores, the curved average and SD can end a little away from a target.
Rounding is applied last (half up). Letters before and after use the A–F scale with 90/80/70/60 cut-offs or the plus/minus scale (A 93, A− 90, B+ 87 …); the histogram groups percentages of the maximum in tens.
The statistics
The standard deviation is the population SD of the class (divided by the number of students), so a curve to a target SD of 10 gives exactly 10 on that measure. A spreadsheet’s STDEV.S divides by one less and shows a slightly larger figure; STDEV.P matches. The median is the middle score, or the average of the two middle scores.
Limitations
- One test and one maximum at a time, up to 5,000 scores. Scores must be 0 or more.
- A curve changes grades: follow your school’s or department’s grading policy, and tell students how the test was curved.
- Letter bands by rank grade students against each other, not against a standard; in small classes a single tie can move a band noticeably.
Privacy
Everything happens in your browser. What you enter or open here is not uploaded or stored by MySmartCoPilot.
Frequently asked questions
How do I curve grades?
Pick a method and apply it to every score: add the same points to all, scale so the top score becomes 100%, rescale to a target average and spread, or use the square-root curve. Paste the scores here, choose the method, and the calculator does it for the whole class with the statistics before and after.
What is the 10√x curve?
It replaces each percentage x with 10 × √x: 49 becomes 70, 64 becomes 80, 81 becomes 90, while 0 and 100 stay. It helps low scores most and never changes the order of the class.
How does bell-curve grading work?
Each score is turned into a z-score — how many standard deviations it is from the class average — and placed on a scale with the average and spread you choose: curved = target mean + target SD × z. With a target of 75 and 10, a student one SD above the class average gets 85.
What is the difference between a flat curve and scaling to the highest score?
A flat curve adds the same points to everyone. Scaling multiplies every score by maximum ÷ highest, so higher scores gain more points: in a class whose top score is 80, a 40 gains 10 and a 60 gains 15.
Why did some scores not change?
Because never lower a score kept them, or cap at the maximum stopped them at 100%. The notes under the result say how many scores each option touched; switch an option off to see the raw curve.
Are my students’ scores uploaded?
No. The scores, names and the curve are handled in your browser, and the CSV is created on your device.