Why Your CAT Quant Score Does Not Match Your Accuracy

QUANT

Your accuracy went from 62 percent to 78 percent across four mocks. That is a substantial improvement and you can see it in the report. Your Quant score moved by almost nothing, and the mismatch is baffling in a way that makes people distrust the whole measurement.

The measurement is fine. Accuracy and score are related but they are not the same quantity, and accuracy can rise for reasons that produce no marks at all. This piece shows the four ways the two come apart, the arithmetic that makes it concrete, and which number to actually track, because a candidate optimising accuracy alone can improve on paper while standing still on the thing that matters.

They Are Not Measuring the Same Thing

Start with the definitions, because the confusion lives there.

Accuracy is correct answers divided by attempts. It is a ratio, so it has no information about how many questions you attempted, how long each took, or how many you never saw.

Score is a total under the marking scheme: plus three for a correct answer, minus one for an incorrect MCQ, zero for an unattempted question, and no negative marking on Type In The Answer questions. It depends on how many correct answers you produced in absolute terms, which means it depends on volume as well as on the ratio.

Two candidates can have identical accuracy and very different scores, and the same candidate can raise accuracy while their score sits still. That is not a paradox, it is what happens when you watch a ratio and care about a total.

Four Ways They Come Apart

The First One Explains Most Cases

If your accuracy rose and your score did not, check your attempt count before anything else. It is the explanation more often than the other three combined.

The arithmetic is easy to see. Suppose a candidate attempts 20 questions at 60 percent accuracy: 12 correct at plus three is 36, 8 incorrect MCQs at minus one is 8 off, net 28. Now suppose they attempt 14 at 79 percent: 11 correct is 33, 3 incorrect is 3 off, net 30. Accuracy rose by nineteen points and the score moved by two marks.

Nothing went wrong in that second paper, and the candidate did become more selective. But the report will show a dramatic accuracy improvement attached to almost no score improvement, and if they read only the percentage they will conclude the mock series is broken.

Where the Improvement Landed

The second mechanism is subtler and it is why section-wise review beats overall review.

Accuracy is computed across everything you attempted, so an improvement concentrated in a question type that appears rarely moves the ratio without moving many marks. A candidate who spent a month on a low-frequency area may genuinely be better at it and see very little on the scoreboard.

This is the practical case for sequencing preparation by historical return. Arithmetic and Algebra have consistently been the largest contributors to QA, with Geometry and Number System behind them and Modern Maths a smaller share. An accuracy gain in a high-frequency area converts into marks; the same gain in a rare one largely does not.

What the report shows	Likely cause	What to check
Accuracy up, score flat	Attempts fell	Attempt count across the same mocks
Accuracy up, score slightly up	Gain in a rare question type	Which types the improvement sits in
Accuracy up, unattempted up	Correct answers got slower	Time per attempted question
Accuracy flat, score up	Volume rose at steady accuracy	Nothing, this is the good case
Accuracy down, score up	More attempts, penalties absorbed	Whether the extra attempts paid

The Variable Accuracy Cannot See

Time per correct answer is invisible in the percentage and it decides your total, which makes it the most important number your mock report does not show you.

Two candidates both answer twelve questions correctly at the same accuracy. One took ninety seconds per question and had time to look at the rest of the section. The other took three minutes each and never saw the last six questions. Identical accuracy, and one of them had a ceiling imposed by the clock that the other did not.

Under a hard 40 minute sectional limit with no movement between sections, that ceiling is absolute. Capability you did not have time to deploy scores nothing, and accuracy has no way of telling you that is what happened.

What to Track Instead

The fix is to watch a small set of numbers together rather than one in isolation.

Marks in the section is the outcome and the only one that counts. Attempts tells you the volume behind it. Accuracy tells you the quality of those attempts. Questions never reached tells you whether the clock or your judgement set the limit. Read together, those four explain almost any movement; read alone, any one of them misleads.

The pattern you want is attempts holding steady or rising while accuracy rises. That combination has no way of leaving your score flat. Accuracy rising while attempts fall is the pattern that produces the frustration this piece is about.

Setting a Target That Cannot Backfire

Because accuracy can be gamed by attempting less, a target framed on it alone is a target you can hit while going backwards.

Frame it as a pair instead: raise accuracy while holding attempts at least steady. That version cannot be satisfied by becoming more cautious, which is exactly the loophole you want closed. It also tells you clearly when you have not met it, because a drop in attempts invalidates the accuracy gain rather than being ignored.

Doing that honestly requires recording attempts every mock, which most candidates do not. The number is available in every report and gets skipped because the percentage is more prominent and feels more like a verdict.

When Falling Attempts Are Correct

One honest qualification, because the argument above can be over-applied.

Sometimes attempting fewer questions is the right move and does raise the score. A candidate who was attempting far too many at low accuracy is losing marks to penalties, and cutting the weakest attempts removes those penalties while keeping most of the correct answers. In that situation accuracy rises, attempts fall, and the score genuinely improves.

The distinguishing test is simply whether the score moved. If accuracy rose, attempts fell and the score rose, the selectivity was correct and you should keep going. If accuracy rose, attempts fell and the score did not move, you have traded volume for ratio at no gain, and the next improvement has to come from somewhere else.

One More Number Not to Over-Read

While comparing mock reports it is worth restating what the percentile on them is, because it adds noise to exactly this analysis.

A mock percentile is your position within that mock's test-taking cohort, and those cohorts are self-selected and small. It is not a CAT percentile and it moves with who happened to sit that mock as well as with your own performance.

For the question this piece is about, compare raw marks across your own mocks from the same series. The percentile introduces a variable that has nothing to do with whether your accuracy converted into marks, and it can show movement in either direction for reasons entirely outside your paper. There is also no fixed mapping between marks and percentile in any case, since percentile depends on paper difficulty and normalisation across slots.

The Summary

Accuracy is a ratio and score is a total, so they can move independently without anything being wrong with the measurement.

Four mechanisms produce the gap. Attempts fell, which is the cause in most cases and which the arithmetic makes obvious once you look. The accuracy gain landed in question types that appear rarely, so few marks followed. Correct answers got slower, so fewer questions were reached and the clock capped the total. Or the mix of wrong answers and blanks shifted in a way the percentage records differently from the scoreboard.

Track four numbers together: marks, attempts, accuracy, and questions never reached. Set targets as a pair, raising accuracy while holding attempts steady, because that version cannot be satisfied by simply becoming more cautious. And if accuracy rose while attempts fell and the score still rose, the selectivity was right, which is the one case where the gap is good news.

If your accuracy is rising and your score is not, the answer is almost certainly in a number your report shows and you have not been reading. A CAT preparation strategy review will identify which of the four mechanisms is yours, and a personalised CAT preparation plan targets the one that is actually binding.