Test Grade Calculator
Questions right or questions wrong — enter whichever you were given — and get the percentage, the letter, and the thing a score alone never tells you: how many more you could have missed before the letter changed.
The paper in front of you
How many questions the test had. Every one counts the same here — a test scored out of points needs a points calculator instead.
Teachers mark the misses and students count the hits. Switch between the two and the number below is converted for you.
Whole questions only. A half-credit answer is a points score, not a question count.
Schools disagree. A 93 is an A here and an A− wherever A starts at 94; a 90 is an A− here and a straight A on the ten-point scale. Check your syllabus.
34 of 40 right. Every question on this test is worth 2.5%, and a B survives up to 6 wrong.
- Percentage
- 85%
- Letter grade
- B
- Questions right
- 34
- Questions wrong
- 6
- Each question is worth
- 2.5%
| Letter | Needs at least | Miss at most | Score there |
|---|---|---|---|
| A | 93% | 2 | 95% |
| A− | 90% | 4 | 90% |
| B+ | 87% | 5 | 87.5% |
| B | 83% | 6 | 85% |
| B− | 80% | 8 | 80% |
| C+ | 77% | 9 | 77.5% |
| C | 73% | 10 | 75% |
| C− | 70% | 12 | 70% |
| D+ | 67% | 13 | 67.5% |
| D | 63% | 14 | 65% |
| D− | 60% | 16 | 60% |
| F | 0% | 40 | 0% |
- Every question counts the same. Partial credit, extra credit and questions worth more than one point are all points scoring, and none of them is modelled here.
- The letter is read off the percentage rounded to two decimals, so the number above and the letter beside it can never disagree at a cutoff.
How the score, the letter and the headroom are worked out
The percentage is one division. The interesting part is the third formula, which is where the table under the result comes from.
p = (n − w) ÷ n × 100 • step = 100 ÷ n • wₘₐₓ = the largest w whose p still reaches the band- n, w
- Questions on the test, and questions answered wrong. Enter the number right and w is n minus it
- p
- The score as a percentage, rounded to two decimals — the letter is read off this figure, not the raw one
- step
- What one question is worth, and therefore the gap between two adjacent achievable scores
Reading the letter off the rounded percentage rather than the raw one is a small decision with a visible consequence: a paper shown as 85% is labelled from 85%, so the number on screen and the letter beside it can never disagree at a cutoff.
For the headroom column, every possible score on the test is walked from a perfect paper downward and each one is assigned its letter. The last score to land in a band is that band’s worst passing paper, and the number of wrong answers behind it is how much room the letter has. Some bands never get a score at all — that is the next section, and it is not a bug.
Six wrong out of forty, and the B with no room in it
Forty questions, six missed. That is 34 ÷ 40 = 0.85, so 85%, and on the plus/minus scale where B starts at 83, a B. Each question is worth 100 ÷ 40 = 2.5 percentage points.
Now look at what that 2.5 does to the B band. The band runs from 83 up to but not including 87. The achievable scores near it are 87.5 (five wrong), 85 (six wrong) and 82.5 (seven wrong) — and only one of those, the 85, is inside the band. On a 40-question test the entire B grade consists of one possible paper. Miss one fewer and you are a B+; miss one more and you are a B−.
This is the number the score alone hides. Eighty-five percent sounds like a comfortable B with a couple of points of cushion. It has none.
A 20-question test cannot produce an 87%
A test does not produce a continuous percentage. It produces one of n + 1 values, spaced 100 ÷ n apart, and nothing in between. On twenty questions the reachable scores are 100, 95, 90, 85, 80, 75 and downward. There is no 87, no 92, no 88. A teacher asking for an 87 on that quiz is asking for a 90 without realising it.
The consequence for letter grades is sharper than most people expect. The plus/minus bands are three or four percentage points wide, so a step of five jumps clean over some of them. On a 20-question test, B+, C+ and D+ are unreachable: no paper can score inside them. On a 10-question quiz it is worse — six of the twelve letters (B+, B, C+, C, D+ and D) have no score that lands in them at all, because the only possible marks are 100, 90, 80, 70, 60 and below.
The threshold is easy to state. The narrowest band on the standard scale is three points wide, so a step of three or less guarantees every band contains something: at 34 questions or more, no letter on that scale is empty. Below 34, gaps are possible, and below about 25 they are routine. The calculator above names them whenever they appear.
None of this is a rounding artefact and no calculator can smooth it away — it is what discreteness does to a scale. It is also a real argument for longer quizzes: a 10-question test does not measure a grade to the nearest letter, it measures it to the nearest ten points. If you want the full score-to-percentage table for a test of any length, the easy grader prints every row at once.
What this calculator assumes
- Every question is worth the same and is either right or wrong. Partial credit, extra credit and questions worth several points are points scoring, and none of them is modelled here.
- Questions are whole. A count of 6.5 is rejected rather than quietly accepted, because a half-question is a sign the test is really being scored out of points.
- No curve is applied. Adding points to every paper in a class is a decision about the class, not a property of yours — apply it to the score before you type it.
- The cutoffs default to the plus/minus scale common in United States schools, and are yours to change. There is no national grading scale, and the same 90% is an A− here and a plain A on the ten-point scale.
- Percentages are rounded to two decimal places and the letter is read off that rounded figure. A teacher who rounds to a whole percent before assigning the letter can land a boundary paper one band higher.
Test grade FAQ
Why can't I get an 87% on a 20-question test?
Because 87 is not one of the scores that test can produce. A test with n questions has exactly n + 1 possible scores, spaced 100 ÷ n apart. On 20 questions that spacing is 5 points, so the achievable scores are 100, 95, 90, 85, 80 and downward — and there is nothing between 85 and 90. On the standard plus/minus scale that empties three whole letters: B+ (87), C+ (77) and D+ (67) have no score that can land in them. If a teacher says you need an 87 on a 20-question quiz, the lowest paper that clears it is a 90, which means missing two rather than three.
How do I calculate a test grade by hand?
Divide the number right by the total number of questions and multiply by 100. Thirty-four right out of forty is 34 ÷ 40 = 0.85, or 85%. If you know the number wrong instead, subtract it from the total first. The useful shortcut is the per-question value: 100 ÷ 40 = 2.5, so on that test every question you miss costs exactly 2.5 percentage points and you can do the rest in your head.
How many can I miss and still get an A?
It depends on the length of the test and on where your school starts its A, which is why this page prints the whole table rather than one number. On a 40-question test with an A starting at 93, the answer is two: three wrong is a 92.5% and an A−. On a 20-question test it is one. On a 100-question test it is seven. The pattern is that shorter tests are harsher — each question carries more of the grade, so the same letter buys fewer mistakes.
Is a 70% a C or a C−?
On the plus/minus scale used by many US schools, a 70 is a C− and a 73 is a C. On the ten-point straight-letter scale that a great many other schools use, a 70 is a plain C. There is no national standard, which is why the cutoffs on this page are editable: set them to match your syllabus before you read the letter, because the same percentage genuinely earns different letters at different schools.
What if some questions are worth more than others?
Then this is not the right tool. A question count assumes every question carries the same weight, which is what makes '6 wrong out of 40' a complete description of the paper. A test where question seven is worth twelve points and the rest are worth four is a points score: add up the points you earned, divide by the points available, and use that percentage. Partial credit works the same way.
Does this round 89.5% up to 90%?
No. The percentage is rounded to two decimal places and the letter is read off that same rounded figure, so the number and the letter can never disagree — but 89.5 stays 89.5 and lands in the band below 90. Plenty of teachers round to a whole percent before assigning the letter, which would make that same paper an A−. If yours does, round the percentage yourself and check it against the cutoff column in the table.
Sources and review notes
- Standards for Educational and Psychological Testing (AERA, APA and NCME) — the joint standards that treat a test score as an estimate carrying measurement error, which is the professional form of the point this page makes about short tests
- NCES Digest of Education Statistics — the standing federal reference for how American schools record grades, and the reason the cutoffs here are a setting rather than a constant
The arithmetic on this page is a division and a sweep over a finite list; neither can go stale. What varies is the scale, which is set school by school and sometimes teacher by teacher. The sources above are cited for why no single set of cutoffs can be treated as correct, not for the calculation. Your syllabus outranks both of them.