CAT question selection strategy: what to attempt, what to skip

Why selection matters more in CAT than in most exams

Three features of the CAT paper make selection decisive. First, each section is timed on its own - in recent papers, 40 minutes each for VARC, DILR and Quant - and you cannot carry spare minutes from one section to the next. Second, multiple-choice questions carry negative marking: +3 for a right answer, −1 for a wrong one. Type-in-the-answer (TITA) questions carry no penalty. Third, every question is worth the same. A two-minute arithmetic question and a ten-minute monster carry the same three marks.

Put together, the paper is less a test of how far you can push on one problem and more a test of how quickly you can spot the questions that give you three marks for the least time. Candidates who "fight" every question they open usually finish with fewer attempts and lower accuracy than candidates who move on early.

The two-pass method

Within a section you can move freely between questions, so use two passes instead of going strictly in order.

1. Pass one - the sure marks (about 60 to 65 percent of the section time). Open each question (or set, or passage) and decide quickly. If you can see the route to the answer, solve it now. If it looks long, unfamiliar or wordy, mark it and move on without starting it.
2. Pass two - the workable ones (the rest of the time). Go back to the marked questions and pick the most promising first. Now you can afford to spend longer, because your sure marks are already banked.
3. Last two minutes - the free attempts. Fill in any TITA question you have a sensible answer for, since a wrong TITA answer costs nothing. Do not spray guesses on multiple-choice questions you have not read.

The point of the first pass is not speed for its own sake. It stops one stubborn question from eating the time that three easier ones needed, which is the most common way good candidates lose marks.

The 45-second triage rule

For every question you open, give yourself about 45 seconds to answer one question: do I know how to start this? Not "can I finish it", just "do I see the first step". If yes, solve it. If no, mark it and leave. Signs that a question belongs in the second pass:

• You have read the stem twice and still cannot say what is being asked.
• The setup needs a long table, a case split, or more than one diagram before any real work starts.
• In verbal questions, the options are long and all four look close on the first read.
• You feel the urge to "just try something" and see where it leads.

The opposite matters too: once you start a question, set a ceiling. If you are past about three minutes on a single Quant question, or past about two minutes on a single RC question, with no end in sight, stop. The time already spent is gone either way; what matters is whether the next two minutes will earn more elsewhere.

When to guess: the negative-marking math

Under CAT marking (+3 right, −1 wrong on multiple choice), you can work out exactly when a guess is worth it. With four options and no idea at all, you are right one time in four: 0.25 × 3 − 0.75 × 1 = 0 marks on average. A blind guess is a coin flip that only adds risk. But eliminate even one option and the numbers turn in your favour.

So the rule is simple: never guess blind on a multiple-choice question; guess freely once you have genuinely removed one or two options; always enter an answer on a TITA question you have worked on. The word "confidently" matters. If your eliminations are themselves shaky, the real odds are worse than the table shows.

Selection in VARC

VARC has 24 questions: in recent papers, four Reading Comprehension passages with four questions each (16 questions) and 8 Verbal Ability questions. The RC passages are two-thirds of the section, so selection here is mostly about order, not skipping. Open all four passages in the first minute or two, rank them from most to least readable for you, and do them in that order. Within a passage, answer the questions you can support from the text first, and leave the one or two that need a long option-by-option comparison for the end. Our guide on how to choose RC passages in CAT walks through the ranking in detail.

For Verbal Ability, para jumbles and odd sentence out are usually TITA, so they are always worth an answer, while para summary is multiple choice and follows the guessing rule above. See CAT VA question selection for a type-by-type plan.

Selection in DILR

DILR comes in sets - typically four or five questions built on one shared puzzle, table or chart. That changes the unit of selection: you are choosing sets, not questions. Spend the first three to four minutes scanning every set before solving any. For each, note the type (arrangement, games and tournaments, charts, routes, scheduling), how many conditions it gives, and whether the questions look answerable once the main table is built.

Then commit to the two or three sets you rank highest and solve them fully. A set you have fully cracked often gives four questions for the time one of them cost. Two warnings: a set that looks short can hide a case split that doubles the work, so if your table is not settling after about eight minutes, stop and switch; and do not leave partly solved sets untouched - even an unfinished table often answers one or two questions directly.

Selection in Quant

Quant questions are independent, so the two-pass method works at its cleanest here. In the first pass, take every question whose method you can see immediately - direct arithmetic, a standard algebra form, a familiar geometry result - and mark everything else. Quant also has a good number of TITA questions, which makes a careful attempt on them low-risk.

The trap in Quant is the question that is almost done. You have a messy expression, you are sure the answer is close, and five minutes go by. Use the options where you can: plugging in values, checking the units digit, or testing extreme cases often settles a multiple-choice question faster than finishing the algebra.

Train selection in mocks, not on exam day

Selection only becomes reliable when it is practised. After each mock or sectional test, go through the paper and sort every question into four boxes: attempted and right, attempted and wrong, skipped but was easy, and skipped and rightly so. The two middle boxes are your selection errors. Over a few weeks you will see a pattern: maybe you keep skipping short tone questions that are actually quick, or keep sinking time into DILR sets of one particular kind.

For VARC, the fastest way to build this judgement is volume on passages you have never seen. Finite question banks get memorised, and a passage you have read before teaches nothing about selection. PracticeRC turns any article - or one of our curated essays - into a timed CAT-style RC set with four questions and an explanation for every option, so every rep is a fresh selection decision.

Common selection mistakes

• Going strictly in order. Question 1 is not special; the order on screen is not the order of difficulty.
• The sunk-cost spiral. Staying with a question because you have already spent four minutes on it.
• Blind guessing on MCQs at the end. On average it adds nothing and adds risk.
• Leaving TITA questions blank. A worked TITA answer costs nothing if it is wrong.
• Skipping by topic label alone. A "hard" topic can hold an easy question, and a friendly passage can hide a trap.
• Changing the plan mid-section. Decide your approach in mocks; the exam is where you run it, not invent it.

Frequently asked questions