CAT VARC Score: How to Climb From 75 to 90 Raw Marks

You already know a strengthen question from a weaken question. You can compress a 900-word passage into one accurate sentence. Your reading speed is fine, your inference is usually sound, and yet the VARC section of your last five mocks reads the same way every time: a sectional percentile that will not move. The concepts stopped being the problem a while ago. What is left is the four to six marks that leak out of every paper before you press submit.

At the top of the range, VARC scoring stops being about what you know and starts being about what you execute. Aspirants stuck around the 90th sectional percentile are not missing a comprehension skill. They are losing two or three reading-comprehension questions to careless errors and leaving two or three verbal-ability marks unclaimed. This piece is about closing that specific gap, with honest arithmetic about how CAT VARC is actually scored.

Why VARC scores plateau near the top

There is a point in preparation where the study material stops adding marks. You have read the theory, drilled the question types, and learned to spot the trap options. Past that point, re-reading a grammar module or an RC guide does almost nothing, because the knowledge is already in place. What separates a 90th-percentile VARC score from a 99th-percentile one is rarely a concept. It is execution under a 40-minute clock: how few questions you misread, and how many low-risk marks you refuse to leave on the table.

It helps to be precise about the numbers. The CAT exam VARC section carries 24 questions. Multiple-choice questions score +3 for a correct answer and -1 for a wrong one. The non-MCQ questions, the ones CAT labels TITA (Type In The Answer), score +3 for a correct answer and carry no penalty for a wrong one. That puts the section ceiling at 24 times 3, which is 72 raw marks. You cannot score 90 raw marks in VARC, and no one does. The phrase aspirants search for reflects a percentile mindset, not a real score. Read "75 to 90" as the climb from a strong sectional percentile to a top-band one, where a handful of net marks can cover several percentile points.

The two-lever framework: error elimination and TITA capture

Once concepts are handled, exactly two levers move a top-band VARC score, and both are about execution rather than learning. The first is error elimination: removing the two or three reading-comprehension questions you get wrong not because the passage defeated you, but because you misread the stem, chose an option that overreached, or fumbled the final elimination. The second is TITA capture: making sure every verbal-ability question that carries no negative marking gets an answer, because a blank guarantees zero marks from it.

The arithmetic is unsentimental. A careful aspirant recovers two to three marks by cutting careless RC errors and another two to three by attempting every no-penalty TITA question. That is four to six marks, and in a competitive year that is frequently the whole gap between the 95th and the 99th sectional percentile. You are not chasing a new skill here. You are refusing to donate marks you already earned the right to keep.

• Lever one, error elimination. Convert two to three careless RC losses per paper into correct answers or clean skips. The knowledge is already there; the leak is attention.
• Lever two, TITA capture. Attempt every para-jumble and odd-one-out, because their expected value is never negative and a blank is a guaranteed zero.

The careless-error taxonomy

Careless errors at the top are not random. They fall into a few repeatable patterns with a shared cause. Daniel Kahneman's 2011 book Thinking, Fast and Slow describes two modes of thinking: System 1, fast and intuitive, and System 2, slow and deliberate. Under a 40-minute clock, System 1 answers what feels right and moves on, and you rarely engage System 2 to check whether it even read the stem correctly. Every careless VARC error is a System 1 answer that a two-second check would have caught. The four below account for most of them.

The fix in every row takes about two seconds, and none of it requires new knowledge. Build the four checks into a single pre-submit habit, and the two to three careless marks you have been leaking each paper come back. If strengthen-and-weaken stems are where you slip most often, the four-step method in our guide on CAT RC strengthen and weaken questions drills exactly that stem-reading discipline.

TITA expected value: why every attempt is worth making

The second lever is pure arithmetic, and it is the one most aspirants get wrong out of misplaced caution. Because MCQs carry a -1 penalty, careful aspirants develop a habit of skipping anything they are unsure of. That habit is correct for MCQs and actively harmful for TITA questions. For an MCQ you have no read on, a blind guess among four options has an expected value of 0.25 times 3, plus 0.75 times -1, which comes to 0 marks, so skipping loses nothing. For a TITA question, which has no penalty, the expected value of any attempt is your probability of being right multiplied by 3, a number that can never fall below zero. Even a one-in-five shot is worth 0.6 marks. Leaving it blank is worth exactly nothing.

Most para-jumbles and odd-one-out (odd-sentence) questions in CAT VARC are TITA, which are precisely the questions where the skip habit costs you. Treat them as lottery tickets you have already paid for: narrow the options as far as your reasoning allows, then always commit an answer. Practising these under timed conditions on a large bank of VARC practice questions trains you to reach a defensible answer quickly instead of freezing on the ones you cannot fully solve.

The top-band improvement routine

You cannot fix errors you have not named. The top-band routine is a diagnostic audit of your last five mock VARC sections, and it takes about an hour. Norman Mackworth's 1948 vigilance research, the origin of the vigilance decrement, showed that sustained-attention performance drops measurably within the first half hour of a monotonous task. VARC runs 40 minutes, so your careless-error rate almost certainly climbs in the back third of the section. The audit surfaces both your recurring error types and the point on the clock where they cluster.

Run this after every block of mocks, not once. Feeding your section-wise results into the CAT score predictor shows whether the recovered marks are actually moving your projected percentile, which keeps the audit honest. If the same plateau shows up across sections and not only in VARC, the wider diagnosis in our piece on the CAT percentile ceiling covers the mock-to-mock patterns that keep aspirants stuck.

What not to do at this level

The instinct at a plateau is to add: another test series, a new vocabulary app, a fresh stack of RC sources. At the top of the range, adding material is usually the wrong move, because the constraint is not input. Doing more of what already works will not fix execution errors, and it quietly steals the time a proper error audit needs.

Timing matters as well. In the final stretch before the exam, the audit-and-check routine compounds better than any new material could, which is why our CAT last 30 days plan puts error review ahead of fresh content. Protect the marks you have earned, claim the ones the marking scheme is giving away, and let the percentile follow.

Optima Learn Editorial Team

Optima Learn is an AI-powered CAT preparation platform built on behavioural science and cognitive psychology research. Our editorial team translates findings from attention research, decision science, and learning research into practical section strategy tested against real aspirant data and mock patterns. Every method published here is designed to hold up under the 40-minute VARC clock and across the full CAT 2026 preparation arc.