Grading Scale Charts for 10 to 100 Question Tests

How many wrong answers each grade allows

Most people arriving here want one number: how many can be wrong before the grade drops. That depends only on the length of the test, so it fits in a single table — one row per test length, one column per grade band.

Every figure below is computed at build time by the same module our test grade calculator runs on. No result below was typed by hand and none was copied from another chart, which is the only reason the two can be trusted to still agree next term.

The second column is the one worth reading twice. On a 10-question quiz a single wrong answer moves the score 10 percentage points — a full letter grade and then some. On a 100-question exam it moves it 1. The same student, the same proportion of mistakes, and a completely different amount of damage per slip. Short quizzes are not gentler than long ones; they are harsher, and the arithmetic is why.

How to read one of these charts

Count the mistakes rather than the correct answers — on most papers there are far fewer of them, and it is the count these charts are indexed by. Find that number in the first column and read across: the score out of the total, the percentage, and the letter.

The percentage is (total − wrong) ÷ total × 100, rounded to one decimal place, and the letter is read off the rounded figure afterwards. That order is worth knowing, because it is what decides a borderline paper. It also explains the ragged percentages on lengths that do not divide neatly: 2 wrong out of 15 is 86.666…%, which the chart shows as 86.7% and grades as a B.

The scale these charts use

No US body publishes the letter-grade scale — cutoffs are a district, school or teacher decision, as the FAQ above says — so this needs saying plainly rather than being left as an assumption. (Elsewhere it can be a published standard: Ofqual sets England’s GCSE grades and the ECTS scale is defined across the EU. Neither governs an American gradebook, which is what these charts are for.) The charts use the standard US 13-band scale the test grade calculator grades with — it is fixed, not a setting — and that scale is a convention rather than a rule.

If your school draws its lines somewhere else, the charts still work — the percentage column is arithmetic and does not care about anybody’s policy. Only the last column is a convention. Read the percentage, apply your own cutoffs, and the chart has still saved you the division.

A short test cannot produce every letter grade

Here is something these charts make obvious and almost nothing else does. A test of N questions has exactly N + 1 possible scores, spaced evenly apart. Where those steps are wider than a grade band, no score lands inside the band at all — and that grade simply cannot happen on that test.

On a 10-question quiz, 7 of the 13 letters are unreachable, and the clearest hole is in the Bs. 8 correct is 80%, a B-, and 9 correct is 90%, an A-. Nothing a student can score falls between them, so the B+ and B bands are both empty while B- survives. Marking a ten-question quiz on a plus/minus scale means handing out 6 possible grades while writing them down as though there were 13.

The comparison that surprised us is 25 against 30. The longer test loses more letters: 25 questions misses only C+, while 30 questions misses B+, C+, D+. Step size is not the whole story — where the steps land matters as much. A 25-question test moves 4 points at a time and happens to land on 88% and 68%, which sit inside the B+ and D+ bands. A 30-question test moves 3.3 points and lands on 90%, 80%, 70% and 60% — the floors of the four minus bands — while stepping clean over the three plus bands below A+.

Of the 10 lengths charted here, 5 produce the whole scale: 35, 40, 50, 60, 100 questions. The other 5 lose at least one letter. If you grade with plus and minus bands and you have any say over how long the test is, that is a real argument for 35 questions over 30.

The charts

One chart per length, every row the calculator produces, from a perfect paper down to the first failing score. They stop there on purpose: past that row every remaining result is another F, and the 100-question chart alone would otherwise carry 59 of them.

When the chart beats the calculator, and when it does not

A chart wins when you are marking the same test thirty times. There is no count to set and no score to enter, so your hands stay on the papers — glance at the row, write the grade, move on. It is also why the cardboard slide-graders survived so long: for one repeated test length, a fixed table is genuinely faster than any interface.

The test grade calculator wins everywhere else. It handles any length from 1 to 500 questions rather than these 10, it applies the two curves teachers actually use (flat added points, or rescaling so the best paper counts as full marks), and its Print chart button produces a self-contained sheet for your exact question count. This page has no print stylesheet of its own, so printing it brings the site’s navigation along for the ride — the calculator’s print sheet does not.

One habit is worth more than either tool: settle the scale before you start marking, not after. Every chart here reads the same percentages, and the one thing that can invalidate a whole stack is discovering afterwards that your department draws the A line at 93% rather than 90%.

The rest of the assessment day is covered elsewhere: the exam clock runs the test itself on a projector, and spelling test practice and the multiplication timed test generate the papers that end up in the stack. The rest live in classroom tools.