How to Attempt CAT 2026: Question Selection Strategy Live
A real-time in-exam decision tree for CAT 2026 that covers the first-90-second scan protocol, the attack-flag-skip triage, the 3-sweep solving pattern, section-specific attempt orders for VARC, DILR, and QA, the 4-minute DILR abandonment rule, and the conservative last-3-minute cleanup protocol that separates 95 percentile candidates from 80 percentile candidates.
How to Attempt CAT 2026: Question Selection Strategy Live
Two candidates with identical prep walk into the same CAT slot. One finishes with 22 attempts at 90 percent accuracy. The other finishes with 28 attempts at 65 percent accuracy. Same syllabus knowledge. Different in-exam decision tree. How to attempt CAT 2026 is the question that decides this difference. It is the live decision tree that runs in the candidate's head between minute 0 and minute 40 of each section, deciding which questions to attack, which to flag, and which to skip without regret.
This guide is the in-exam playbook for how to attempt CAT 2026. It covers the first 90-second scan protocol, the flag-skip-attack triage, the section-specific attempt order for VARC, DILR, and QA, the 4-minute DILR abandonment rule, and the last-3-minute cleanup protocol. Pair with the CAT 2026 exam day mistakes guide for the broader cognitive-error checklist.
Every section starts with a 90-second scan that tags questions as ATTACK, FLAG, or SKIP. Three sweeps follow: confident solves first, flagged borderline questions second, last 3 minutes for cleanup. Section attempt orders: VARC starts with para-summary and odd-one-out, then the easiest RC; DILR ranks all 4 sets in the first 90 seconds then attacks the 3 most accessible; QA starts with arithmetic and basic algebra. The 4-minute DILR abandonment rule prevents sunk-cost losses. Last 3 minutes are conservative cleanup, never desperate guesses.
The First 90 Seconds: Scan Before You Solve
The first 90 seconds of each CAT section are not for solving. They are for building a difficulty map. The candidate who skips the scan starts attacking question 1 regardless of its difficulty, often loses 3 minutes on a Tier 3 question, and never recovers the time. The candidate who runs the scan routes attention to the highest-yield questions first.
In VARC: read each RC passage's first 100 words and the question stems; flag the easiest passage. In DILR: read each set's setup paragraph in 20 to 30 seconds; tag each set ATTEMPT, BORDERLINE, or DECLINE. In QA: scan all 22 questions; tag each as INSTANT-SOLVE, STANDARD, or DEFER. The 90-second investment pays back across the remaining 38 minutes.
Flag, Skip, Attack: The 3-Way Triage
Confident topic, clear setup, recognition trigger fires in under 30 seconds. Expected hit rate above 85 percent. First sweep priority.
Familiar topic but specific setup unclear, or familiar setup with one twist. Expected hit rate 60 to 75 percent. Mark for review.
Tier 3 weak topic + unfamiliar setup + no recognition trigger in 30 seconds. Skip without negotiating with yourself.
The triage decision should finish in 30 seconds per question. Once you start negotiating with yourself ("maybe I can solve this if I just try the third approach"), the question has already cost too much time. Triage is unemotional. Practice the 30-second decision in 6 mocks before CAT day.
The 3-Sweep Solving Pattern
Sweep 0: Scan
Read every question stem fast. Triage all 22 to 24 questions. Build the attack list mentally.
Sweep 1: Attack
Solve all ATTACK questions in order of confidence. Target 90 percent plus accuracy. Move on the moment a question stalls beyond 90 seconds; flag and continue.
Sweep 2: Flagged
Return to FLAGGED questions. Expected hit rate 70 percent. Cap each flagged question at 2 minutes; if no progress, move to the next flagged item.
Sweep 3: Cleanup
Review marked answers, verify submissions, pick at most 1 easy SKIP question if it has become obvious. No wild guesses.
Section-Specific Attempt Orders for CAT 2026
VARC Attempt Order
VARC is fixed at 24 questions (typically 16 RC plus 8 non-RC). The recommended order:
- 1. Para-summary and odd-one-out first: non-RC questions average 90 seconds; high accuracy possible. 6 to 8 marks locked in the first 12 to 15 minutes.
- 2. Easiest RC passage second: from the 90-second scan, attack the passage with the most familiar topic and clearest first 100 words.
- 3. Two more RC passages in descending order of accessibility: leave the hardest passage for last.
- 4. Hardest RC passage attempted only if 4 plus minutes remain: otherwise skip; the marks-per-minute is too low.
DILR Attempt Order
DILR is the section where attempt order matters most. The 4 sets typically include one accessible, two medium, and one hard set.
- 1. Rank all 4 sets in the first 90 seconds: ATTEMPT, BORDERLINE, BORDERLINE, DECLINE.
- 2. Attack the ATTEMPT set first: target 4 out of 4 in 12 to 15 minutes; lock in 12 marks.
- 3. Move to the better BORDERLINE set: target 2 to 3 out of 4 in 10 to 12 minutes.
- 4. Attempt the second BORDERLINE set if time permits: cap at 8 minutes; partial attempts are worth the time only if 1 to 2 questions clear.
- 5. The DECLINE set stays declined: attempting it usually loses more in time cost on other sets than it gains.
QA Attempt Order
QA is 22 questions in 40 minutes. The recommended order:
- 1. Arithmetic and basic algebra cluster first: percentages, ratios, SI/CI, averages, time-work, basic equations. Target 10 to 12 questions in the first 22 minutes.
- 2. Geometry and mensuration next: triangles, circles, mensuration, coordinate geometry. Target 3 to 4 questions in the next 8 minutes.
- 3. Number system and modern math last: combinations, logarithms, sequences. Target 2 to 3 questions in the final 8 minutes.
- 4. Last 2 minutes: review flagged questions; do not start a new question with under 2 minutes remaining.
Myth
The candidate with the most attempts wins. Attempt every question; CAT scoring rewards volume.
Reality
CAT scoring is +3 / -1 in MCQ format. 22 attempts at 90 percent accuracy yields 60 marks; 28 attempts at 65 percent accuracy yields 50 marks. Quality beats volume across every percentile band.
Want a CAT 2026 mock attempt analyser that tracks your sweep timing, triage accuracy, and DILR abandonment discipline?
Build My CAT Attempt AnalyserThe 4-Minute DILR Abandonment Rule
The single biggest in-exam DILR mistake is the sunk-cost trap: a candidate is 6 minutes into a set with zero solves, has invested setup work, and refuses to abandon. The data is unforgiving: 4 minutes in with zero solves means the set is failing. The expected return on another 4 minutes is roughly 1 to 2 marks at 50 percent confidence; the expected return on starting a fresh ATTEMPT set is 8 to 10 marks at 70 percent confidence.
"I have already built the setup; abandoning is wasteful." The setup time is already gone. The decision is forward-looking: which next 8 minutes yield more marks. Almost always, a fresh ATTEMPT set yields more than continuing a failing BORDERLINE set. The discipline to abandon is what separates 95 percentile DILR scores from 80 percentile DILR scores.
The Last 3 Minutes Protocol
The last 3 minutes of each CAT section are not for desperate guesses. They are for conservative cleanup:
- Minute 1: scan the question palette for any marked-for-review item; resolve by either confirming or moving on.
- Minute 2: verify every confident solve is properly submitted; check the answer-selection state.
- Minute 3: pick at most one easy SKIP question that has become obvious in retrospect; solve it or leave it.
Wild guesses on Tier 3 questions in the final minute typically cost 1 to 3 marks net (negative-marking on wrong answers exceeds the small probability of lucky hits). The discipline to NOT guess wildly is what separates 95 percentile candidates from 80 percentile candidates.
- Start every section with a 90-second scan before solving any question.
- Triage every question as ATTACK, FLAG, or SKIP within 30 seconds; do not negotiate with yourself.
- Run three sweeps: confident solves first, flagged questions second, conservative cleanup last.
- VARC opens with para-summary and odd-one-out; DILR opens with the most accessible ranked set; QA opens with arithmetic.
- The 4-minute DILR abandonment rule overrides sunk-cost emotion; abandon at 4 minutes with zero solves.
- Cap each flagged question at 2 minutes in sweep 2; cap the hardest QA question at 90 seconds in sweep 1.
- The last 3 minutes are conservative cleanup; never burn them on wild guesses on Tier 3 questions.
CAT rewards better decisions, not more attempts. Practise the triage in mocks until it runs in your sleep.
Build My CAT 2026 Attempt Strategy
Get a mock-by-mock triage tracker, DILR set selection diagnostics, and a personalised sweep-timing plan tailored to your current percentile band. Join the structured CAT 2026 prep sprint.
Build My CAT Attempt Plan
Optima Learn Editorial Team
CAT preparation specialists publishing structured guides on Quant, VARC, DILR, and IIM admissions. We build in-exam decision trees, mock-attempt diagnostics, and exam-day playbooks calibrated to the CAT 2026 format and past-cycle attempt data.
Build your CAT 2026 study plan
Personalised daily plan that adapts to your section-wise mock scores.