CAT accuracy vs attempts chart

Ask an aspirant how to improve their score and most will say the same thing: attempt more questions. It feels obviously true. More attempts, more marks. But CAT has negative marking, and that one rule turns the intuition into a trap. Past a certain point, attempting more questions lowers your score rather than raising it. Understanding CAT accuracy vs attempts as a single tradeoff, not two separate goals, is what separates aspirants who plateau from those who climb.

This guide puts numbers on the tradeoff. You will see the expected-value math of a single MCQ, the confidence threshold that should decide whether you attempt, and the section-wise accuracy benchmarks that match different percentile targets. None of it is complicated; it just runs against instinct.

Why "more attempts" is often the wrong move

Picture two aspirants with identical ability. The first attempts 45 questions at 80 percent accuracy. The second, chasing volume, attempts 55 but their accuracy on those extra ten low-confidence questions is only 30 percent. The second aspirant attempted more and scored less, because the extra attempts brought in wrong answers, and each wrong MCQ does not just earn zero, it subtracts a mark. Attempts and accuracy are not independent dials you turn up separately. They move together, and pushing one too hard drags the other down.

This is why "attempt more" is good advice only at the bottom of the curve, where an aspirant is leaving easy marks untouched. Once you are already attempting the questions you can reasonably solve, the next attempt comes from a shakier pool, and that is exactly where negative marking starts working against you. Knowing which side of that line you sit on is the heart of a smart CAT attempt strategy.

The CAT accuracy vs attempts math

CAT scores a correct MCQ at plus three and a wrong one at minus one. Type-in-the-answer questions carry no penalty. For an MCQ, if your chance of being right is p, the expected value of attempting is three times p, minus one times the chance of being wrong. Written out, that is four times p, minus one. A blank scores zero. So attempting beats skipping whenever four p minus one is above zero, which happens once p passes 25 percent.

The table says something surprising: in pure marks, a random four-option guess breaks even, and eliminating even one option makes an attempt positive. So why do low-confidence attempts wreck scores in practice? Because the marks math is only half the story, and the other half is where most aspirants lose. The elimination logic behind the middle rows is worth practising on its own, which we cover in the CAT educated guess strategy.

The real confidence threshold to attempt

The pure math says attempt anything above a random guess. Real exam conditions push that threshold much higher, for two reasons the math leaves out.

First, time. A question you are 35 percent sure of is rarely a quick one; it usually takes two or three minutes of struggle. Those minutes could have banked a near-certain mark elsewhere, so the true cost of a shaky attempt is not just its small negative risk but the sure mark you did not collect instead. Second, calibration. Aspirants consistently overrate their confidence. A question that feels like a 60 percent chance is often closer to 40 once you check against results. Stack those two effects together and the safe attempt threshold climbs from the textbook 25 percent to something nearer two-thirds confidence in practice.

Section-wise accuracy benchmarks

How accurate you need to be depends on your target and varies a little by section. The benchmarks below are accuracy on attempted questions, not on the whole paper, since unattempted ones neither help nor hurt. Use them as a check: if your accuracy sits below the row for your target, the fix is to attempt fewer, surer questions, not more.

DILR accuracy tends to run highest at every target, and the reason is selection. Strong scorers attempt only the sets they have understood, so once they commit, they are usually right. That is the tradeoff working as intended: fewer, better-chosen attempts produce higher accuracy and a higher net score than a wider, shakier net. The gap between your mock accuracy and these benchmarks is also one of the clearest signals of how your real CAT might differ from your practice scores, which we unpack in why your mock score is not your CAT score.

Common questions on accuracy vs attempts

Stop chasing attempts as if they were the score. They are one half of a tradeoff that negative marking governs, and past your real confidence line, each extra attempt costs more than it pays. Learn the expected-value math, run the bet test on borderline questions, and hold your accuracy at the benchmark for your target. Do that and you will often attempt slightly fewer questions and score noticeably more, which is the whole point. Drill the habit on the Optima Learn question bank until skipping a coin-flip feels as natural as solving a sure thing.