Your mock average is 91 percentile. You walk into the exam confident. The computer flashes your CAT paper. Four hours later, you exit in a daze. Your score? 83 percentile. An 8-percentile drop. Worse, your accuracy was fine; your attempts were low. You were not dumb. You were panicked. In my years coaching CAT, the gap between mock performance and actual CAT performance is almost never about learning gaps. It is about exam-hall mistakes—tactical blunders that compound under pressure.
This post maps the section-by-section mistakes that cost the most percentiles on CAT exam day. Not conceptual errors. Execution errors. You will learn which mistakes repeaters make, which mistakes first-timers make, and which mistakes everyone makes under exam pressure. Then I will give you the inoculation: a pre-exam mental framework and a section-specific decision tree to navigate the exam hall without sabotage.
The VARC Exam-Hall Trap: Time Bankruptcy
VARC is your 40-minute war. You have 24 questions (estimated), split between Reading Comprehension (RCs) and Verbal Ability (VA). Mistake #1: spending 20 minutes on two dense RCs, then rushing through the third RC and all 8 VA questions in 20 minutes. Your accuracy collapses. You attempt 16-18 questions with 60% accuracy instead of 18-20 with 85% accuracy. Lower score.
The mistake happens because of passage selection. In mocks, you had the privilege of reading through all passages and choosing which ones felt manageable. In CAT, you are reading passages blind. A dense philosophy passage shows up first, you dive in, and before you know it, 15 minutes are gone.
The Fix: Spend the first 2 minutes skimming all RC passages and VA questions without solving. Identify the three easiest RCs. Allocate 12-13 minutes per RC (read + answer). Spend 10 minutes on all VA questions (vocab + grammar are faster than RCs). This discipline costs you 2 minutes but saves you 15 minutes of panic. I have seen students jump from 68th to 82nd percentile in VARC just by implementing this skimming discipline.
The DILR Set-Selection Catastrophe
DILR is your 40-minute gamble. You have 20 questions (estimated), likely 5 sets of 4 questions each. Mistake #1: jumping into Set 1, spending 7 minutes understanding a set that has no clear breakpoint. You get two questions right, two wrong. Mentally, you are rattled. You move to Set 2, still shaky, and underperform on a set you could have mastered. By the time you reach Set 4 (which you would have aced), you are out of time and attempt only 3 questions.
Your mock shows you solved 14-15 questions with 80% accuracy. Your CAT shows 12 questions with 65% accuracy. You had the knowledge; you had the time. You did not have the discipline to select the right sets in the right order.
The Fix: Spend 5 minutes reading all 5 DILR sets without solving. Rank them by clarity: (1) Most obvious, solvable in under 10 minutes. (2) Clear, solvable in 10-12 minutes. (3) Ambiguous, likely to cost time. (4) Hard, skip unless time permits. (5) Impossible, skip. Solve in order: 1, 2, 4 (if time). Skip 3 and 5. You will attempt 12-14 questions with 80%+ accuracy. Higher score than attempting 15-16 with 60% accuracy.
The Quantitative Aptitude Panic: Guessing Under Pressure
QA is 40 minutes, 22 questions (estimated). You have 8 TITA (Type In The Answer) and 14 MCQs. Mistake #1: When a QA question takes 2 minutes and you are not seeing the path, you panic-guess. Or worse, you spend 5 minutes on a geometry proof that collapses your time budget. By question 18, you are guessing on three consecutive questions because you are running out of time. Two wrong guesses, one silly mistake. Your score dips.
Additionally, MCQ guessing is toxic. TITA is honest—no negative marking, no penalties. If you do not know a TITA answer, you skip it. But MCQs carry -1 marking. A wrong guess here costs you a mark (you take -1 instead of +3), and a few of those add up fast. Many aspirants guess on a few MCQs without realising how quickly those negative marks add up.
The Fix: Establish hard time limits per question. Arithmetic and Algebra: 1.5 minutes max. Geometry and Modern Math: 2 minutes max. If you have not seen the path by then, mark for review or skip. Spend your time on questions you can solve, not on problems designed to trap you. In CAT, your job is not to solve every question; it is to solve the right questions quickly. Second, NEVER guess on MCQs. If you are unsure, skip. Mark it, review it at the end if time permits. Guessing on MCQs is a -EV strategy. TITA questions are your guessing ground (zero penalty). Use this asymmetry.
Section-Wise Mistakes: The Attempt vs. Accuracy Trade-Off
Every CAT section is a game of attempts versus accuracy. Your job is to find the sweet spot. Here is the mistake most aspirants make:
| Section | Mistake (What Not To Do) | Target (What To Do) | Why |
|---|---|---|---|
| VARC | Attempt 22+ questions with 70% accuracy (14 correct) | Attempt 18-20 questions with 85%+ accuracy (15-17 correct) | 15-17 correct > 14 correct. Higher attempts, lower accuracy = lower percentile. |
| DILR | Attempt all 20 questions, struggling on hard sets (10 correct) | Attempt 12-14 questions, solving easy sets cleanly (10-11 correct) | Fewer, accurate attempts beat more, inaccurate attempts in DILR specifically. |
| QA | Attempt 20+ questions with 65% accuracy (13 correct) | Attempt 16-18 questions with 85%+ accuracy (14-15 correct) | Fewer accurate attempts scale better. One silly mistake costs you 4 points (miss +3, take -1). |
Your exam strategy must reflect this. If you are a natural speed solver (rare), you can push 20+ attempts. If you are a natural accuracy solver (common among engineers), you pull back to 14-16 attempts and maximize accuracy. Neither approach is wrong. But mixing them—attempting 20 questions with 60% accuracy—is a losing strategy.
Negative Marking and the 5-Question Guessing Trap
Here is the math most aspirants do not do. You have 14 MCQs in QA (example). You know 11 of them clearly. 3 are ambiguous. Your impulse: guess all 3, hope for 1-2 hits. Expected outcome: 1 correct (EV), 2 wrong. Net: 1 × 3 = +3 (1 correct), 2 × (-1) = -2 (2 wrong). Total: +1 point for guessing. Sounds okay. But here is the truth: you do not know which question you will get right. In reality, you will likely get 0 or 1 correct, leaving you with 0 points or -2 to +1 points. The variance kills you.
The Fix: If your guess rate on MCQs is below 50% (you get fewer than half your guesses right), never guess. Your expected value is negative. Skip instead. If your guess rate is above 50%, you can guess selectively. Most aspirants are below 50%. Skip your ambiguous MCQs. Period.
The Mental Game: Exam-Hall Psychological Mistakes
You walk in confident. Question 3 of VARC is a dense philosophy passage. You spend 8 minutes reading it and understand 40%. You are suddenly doubting your prep. You panic-solve the passage questions, get 1 out of 4 wrong. Your confidence tanks. You carry that shaken feeling into DILR and QA. Mental error compounds into execution error.
The fix? Detach from individual questions. You will encounter one or two impossible questions in CAT. That is by design. When you hit an impossible passage or a set that makes no sense, flag it, move on, and never look back. Do not let one question destabilize your strategy. Mentally, you need to accept: I will skip 2-4 questions. I will get 1-2 questions wrong. That is fine. This acceptance is freedom.
Your Pre-Exam Strategy Lock
| Step | Timing | Action | Outcome |
|---|---|---|---|
| 1 | First 2 minutes | Skim all RCs without reading deeply. Identify and rank 3 easiest. | Your reading order is locked. No panic. |
| 2 | Next 12-13 min per RC | Read chosen RCs deeply. Answer all 4 questions. Then VA questions (10 min). | 18-20 attempts, 85%+ accuracy. |
| 3 | First 5 minutes (DILR) | Read all 5 sets without solving. Rank by clarity. Solve Sets 1, 2, 4. | 12-14 attempts, 80%+ accuracy. |
| 4 | QA throughout | 1.5 min per Arithmetic/Algebra. 2 min per Geometry. Skip if unclear. Never guess MCQ. | 16-18 attempts, 85%+ accuracy. |
| 5 | Last 5 minutes | Review skipped questions only if time permits. Do not re-solve. | Capture 1-2 additional points if possible. |
Frequently Asked Questions
If I attempt 16 questions and get all right, versus 20 questions and get 16 right, which scores higher?
Both score the same (16 correct = 48 points). But the first approach—16 accurate attempts—is less risky. Fewer chances for silly mistakes. In exams, fewer, more confident attempts beat more, less confident attempts.
Should I use the on-screen calculator in QA, or solve mentally?
Use it for verification only, not solution. If you can solve mentally in under 1 minute, do so and verify with the calculator. If it takes 2+ minutes to solve, you are on the wrong track; skip and move on. The calculator should save you time, not consume it.
What if I encounter a set in DILR that I have never seen before?
Do not panic. DILR has infinite variations. You will not have seen the exact set before. If a set logic is unclear after 2 minutes of reading, move on. You do not have time to reverse-engineer a new set in 2 minutes. Your job is to solve sets whose logic is obvious, not to debug new logic under pressure.
Should I read all questions in a DILR set before the data, or read the data first?
Read the data first, understand the structure, then read questions. Questions often repeat the same information; reading data first gives you context. This saves re-reading time and reduces errors.
How many questions should I attempt in my first pass through the exam?
Aim to attempt all questions you plan to solve in your first pass. No revisits or review rounds. If you have flagged a question as skip, and you have solved all others, only then revisit. But most often, your first-pass attempt count is your final attempt count.
The Exam-Hall Formula
CAT is not about solving every question. It is about solving the right questions with surgical accuracy under time pressure. Every percentile above 90 comes from discipline: skipping the unsolvable, solving the solvable, and accepting that 2-4 questions will remain unsolved. This acceptance is strength, not weakness. Own it. On November 29, 2026, your mocks will not matter. Your discipline in the exam hall will.
Want a final rehearsal before CAT exam day? Book a free counselling call with Percentilers. Let us lock your exam-hall strategy and eliminate costly mistakes.