What Grade Is X Out of Y? The Percentage Is Easy, the Letter Isn't

The arithmetic half takes one line. Divide what you got right by how many there were:

percentage = questions right ÷ questions total × 100

46 out of 50 is 92%. 4 wrong out of 50 is the same thing said backwards. That number is not in dispute anywhere, and if it is all you came for, you already have it.

But almost nobody asking this question wants the number 92. They want to know whether 92 is an A. And that half of the question has no arithmetic answer at all — it has an institutional one, decided by somebody who is not you and published in a document that is not your test.

92% is at least three different grades

Three real published scales, all currently in force at accredited American universities:

One point of raw percentage is the entire distance between an A and an A− at Columbia, and no distance at all at Pace. Nothing about the test changed. The only thing that changed is whose document you are reading.

Pace is the most useful of the three, because it says the quiet part in its own catalogue: the percentage equivalents are shown only as guidelines for faculty. Not a rule the university enforces — a suggestion to instructors, who may set their own. Missouri makes the same point by omission, publishing what each letter is worth in grade points while declining to say what percentage earns it.

This is why the test grade calculatorhas an editable cut-off table rather than a built-in scale. A site that prints “92% = A−” is not giving you a fact; it is giving you one school's policy without telling you whose.

Which document decides yours

Two, in this order.

• Your course syllabus governs the grade in the course. If it prints a cut-off table, that table is the answer, and no general scale outranks it.
• Your registrar's published scale governs what the letter is worth once it reaches your transcript — the grade points behind your GPA. That is a separate lookup, chosen by a different office, and the two do not have to agree.

If your syllabus prints no cut-offs, ask before the exam rather than after. It is an ordinary question, and the answer is one line long.

A cut-off includes its own number, and rounding is a policy

If the A starts at 93, then 93.0 is an A and 92.99 is not. Arithmetic has no opinion about how close that is. Plenty of instructors round, and plenty of syllabi say so in prose — “final percentages are rounded to the nearest whole number” — but that is a stated policy of your course, not a property of the number. Nothing here assumes it, because assuming it would hand you a letter you might not have.

The useful reframe: count questions, not percentages

Percentages are the wrong unit for the decision you are actually making. On a 50-question test, one question is exactly 2 percentage points, so the cut-offs turn into whole answers:

Note the first row. 93% of 50 is 46.5, and half a right answer does not exist, so the A costs 47 — always rounded up, never to nearest. Rounding 46.5 down to 46 would be telling you that you have a letter you do not have. At 46 right you are on 92%, which is the A− we started with.

Counting in questions also makes the margin legible. “Three” is a number you can hold in your head while checking your work; “93%” is not.

Where this stops

A curved score is a different question. If your instructor curved the test, the raw percentage is not what goes on your record, and the method used is almost never written down anywhere — four documented curve methods give four different answers from the same raw score.

Extra credit reads above 100 here, uncapped. 52 right out of 50 is 104%, because that is what an extra-credit mark does to a score, and clamping it to 100 would silently delete points you earned.

And one test is not a course grade. A 92% on an exam worth 25% of your course contributes 23 points, not 92. What the whole term adds up to is a weighted calculation, and what a single remaining exam still has to do is a different one again.