A Gradely About

Test Grade Calculator

Two numbers: how many questions, how many wrong. The percentage and letter appear instantly, and the plus and minus buttons step through a stack of papers without retyping the question count. Below the tool is a chart of every common quiz length, so you can grade a set without touching a screen at all.

Number wrong
Score88.0%
B

22 right out of 25. On this test each question is worth 4.00 percentage points, so one more mistake drops the score to 84.0%.

The rows either side, for marking a stack of papers without retyping.
WrongRightScoreLetter
12496.0%A
22392.0%A
32288.0%B
42184.0%B
52080.0%B
Scores are shown to one decimal place, rounded half-up. The letter comes from the exact fraction, not the rounded number.

The chart: every common quiz length

Each cell is the percentage for that many wrong answers on that length of test, computed as (total − wrong) ÷ total. Percentages are rounded half-up to one decimal place; the letters beneath them come from the exact unrounded value on a ten-point scale, because rounding before you assign a letter can push a score across a boundary it did not really cross. Blank cells mean more wrong answers than the test has questions.

Score for each number of wrong answers, by test length. Letters use ten-point bands (A 90, B 80, C 70, D 60) applied to the exact fraction.
Wrong 5 questions10 questions12 questions15 questions20 questions25 questions50 questions
0 100.0% A 100.0% A 100.0% A 100.0% A 100.0% A 100.0% A 100.0% A
1 80.0% B 90.0% A 91.7% A 93.3% A 95.0% A 96.0% A 98.0% A
2 60.0% D 80.0% B 83.3% B 86.7% B 90.0% A 92.0% A 96.0% A
3 40.0% F 70.0% C 75.0% C 80.0% B 85.0% B 88.0% B 94.0% A
4 20.0% F 60.0% D 66.7% D 73.3% C 80.0% B 84.0% B 92.0% A
5 0.0% F 50.0% F 58.3% F 66.7% D 75.0% C 80.0% B 90.0% A
6 — 40.0% F 50.0% F 60.0% D 70.0% C 76.0% C 88.0% B
7 — 30.0% F 41.7% F 53.3% F 65.0% D 72.0% C 86.0% B
8 — 20.0% F 33.3% F 46.7% F 60.0% D 68.0% D 84.0% B
9 — 10.0% F 25.0% F 40.0% F 55.0% F 64.0% D 82.0% B
10 — 0.0% F 16.7% F 33.3% F 50.0% F 60.0% D 80.0% B

What one question costs

The most useful thing in that chart is not any single cell, it is the gap between rows. On a test of N questions, each question is worth exactly 100 ÷ N percentage points:

  • 5 questions — one wrong costs 20.00 points
  • 10 questions — one wrong costs 10.00 points
  • 12 questions — one wrong costs 8.33 points
  • 15 questions — one wrong costs 6.67 points
  • 20 questions — one wrong costs 5.00 points
  • 25 questions — one wrong costs 4.00 points
  • 50 questions — one wrong costs 2.00 points

A five-question quiz has no C grade available at all: you score 100, 80, 60, 40, 20 or 0, so the possible letters on a ten-point scale are A, B, D, F, F and F. One mistake is the difference between an A and a B; two is the difference between an A and a D. Short quizzes are enormously high-variance instruments, which is worth remembering both when you set one and when you get one back. On a 50-question test the same single mistake costs 2.00 points and almost never changes a letter.

Where the letter depends on the scale, not the score

One wrong out of twelve is 91.7%. On a ten-point scale that is an A. If your school uses a seven-point scale, where the A band starts at 93, the same paper is a B. On a plus/minus scale it is an A-minus. Same twelve questions, same one mistake, three different letters — and none of them is the "right" one outside the school that set the policy. The grade calculator's scale comparison table lists every percentage where the common scales part company.

The rounding rule matters at the same edges. Two wrong out of fifteen is 86.666…%. Taken exactly, that is a plain B on a plus/minus scale — the B+ band starts at 87 and the score is a third of a point short. Round it to 87% first and it becomes a B+. This page never rounds before assigning a letter, and states the rule next to the chart, so that if your instructor does round you can see exactly which papers it moves.

Grading a stack efficiently

  1. Set the question count once. It does not change between papers in the same stack.
  2. Use the minus and plus buttons for the number wrong rather than retyping — one click per paper.
  3. The five-row table under the tool shows the rows either side of the current answer, so a paper one better or one worse is already on screen.
  4. For a whole class, print the chart column for your test length and mark straight from it.
  5. If questions are worth different amounts, stop using question counts and grade in points — the main grade calculator takes 38/45 directly.

Related calculators

Quiz grading questions

How is the percentage worked out?

Questions right divided by questions total, times 100. Three wrong on a 25-question test is 22 ÷ 25 = 88%. That is the whole calculation, and it assumes every question carries the same number of points — which is true of most multiple-choice quizzes and untrue of most written exams.

What rounding does the chart use?

Percentages are rounded half-up to one decimal place: 11 ÷ 12 is 91.666…, published as 91.7%. The letter grade is taken from the exact unrounded fraction, never from the rounded figure, because rounding first can move a score across a boundary. Two wrong out of fifteen is 86.666…%, which is a plain B on a plus/minus scale — round it to 87 first and it becomes a B+.

Some questions on my test are worth more than others. Can I still use this?

Not directly — this page assumes equal-weight questions. Grade by points instead: add up the points earned and the points available and enter the result as a fraction like 38/45 in the main grade calculator, which converts it and keeps the point values.

How do I handle partial credit?

Count it in points rather than in questions. If a student earned half marks on two questions of a 20-question test and got three others wrong, that is 20 − 3 − 1 = 16 points out of 20, so enter it as 16/20 elsewhere instead of trying to express it as a number of questions wrong here.

Why does the same score show a different letter for two of my students?

It should not, unless you changed the scale selector between them. If two different tests give the same number wrong but different letters, that is the point of the chart above: three wrong is an 88% on a 25-question test and a 70% on a 10-question test. The number wrong means nothing without the number of questions.

Can I print the chart?

Yes — it is plain HTML in the page, so your browser's print function will include it. It covers 5, 10, 12, 15, 20, 25 and 50-question tests up to ten wrong, which is the range most classroom quizzes live in.