A 95th percentile does not mean you scored 95% of the marks; it often means far less. Here is the difference, and why it matters when you read your mock results. Understanding CAT percentile vs percentage will also help you set realistic targets, avoid panic after a tough mock and make sense of the scorecard when results arrive.
Percentage measures a share of marks
If a student earns 60 marks out of a hypothetical 200, the percentage is 30%. The formula is marks obtained divided by maximum marks, multiplied by 100. Percentage says nothing directly about how others performed.
Working out a raw CAT-style score
In recent years, CAT has had about 68 questions. Multiple-choice questions (MCQs) carry +3 for a correct answer and −1 for a wrong one, while type-in-the-answer (TITA) questions carry +3 and no negative marking. Unattempted questions score zero. Always confirm the current marking scheme on iimcat.ac.in.
Suppose that in a mock you attempt 20 MCQs and get 15 right, and attempt 6 TITA questions and get 4 right:
| Question type | Correct | Wrong | Marks |
|---|---|---|---|
| MCQ | 15 × 3 = 45 | 5 × (−1) = −5 | 40 |
| TITA | 4 × 3 = 12 | 2 × 0 = 0 | 12 |
| Total | 52 |
With about 68 questions at 3 marks each, the maximum is around 204. Your percentage is 52 ÷ 204 × 100, roughly 25.5%. On its own, that looks low. Whether it is a good score depends entirely on how everyone else did, which is what a percentile tells you.
Percentile measures relative standing
A percentile is a position within a score distribution. Broadly, a 95th percentile means performance at or above that of about 95% of the comparison group, depending on the precise ranking convention. It does not mean 95% of the maximum marks.
Put simply, percentage compares you with the paper, and percentile compares you with other candidates.
| Measure | What it compares you with | Changes if the paper gets harder? | Changes if other candidates get stronger? |
|---|---|---|---|
| Percentage | Maximum marks | Usually falls | No |
| Percentile | Other test takers | Not necessarily | Yes |
| Scaled score | A normalised scale across slots | Adjusted for slot difficulty | Indirectly |
A simple illustration
Imagine 1,000 test takers ranked by performance. A candidate around the top 50 is near the 95th percentile. Their raw score could vary considerably depending on exam difficulty and the performance of the cohort.
A worked example with ten students
Small groups make the arithmetic visible. Ten students take a mock and score:
20, 35, 41, 41, 52, 58, 63, 70, 77, 90
Consider the student who scored 58. One common convention counts the share of candidates who scored strictly below you. Five students (20, 35, 41, 41, 52) scored below 58, so the percentile is 5 ÷ 10 × 100 = 50. Another convention counts candidates at or below your score, which includes you: 6 ÷ 10 × 100 = 60.
With ten students the two conventions differ by a lot. With hundreds of thousands of candidates, the difference becomes tiny, which is why large exams can report percentiles to several decimal places. The exact method for CAT is set out in the official scorecard information, so rely on that rather than any formula you find online.
Notice also that the student with 58 marks has scored 58 out of whatever the maximum is. If the maximum were 204, that is about 28%, yet they are in the middle of this group. If everyone found the paper easy and the scores were all 30 points higher, the same 58 would sit near the bottom. The marks did not change; the comparison group did.
Why high percentiles need fewer marks than you think
CAT is designed to separate a very large number of candidates, and most find it difficult. In recent CAT papers, high percentiles have often corresponded to well under half of the maximum marks. This is the main reason a student who writes “I got only 30% of the marks” may still have done very well. Judge your performance by standing, not by the percentage you would expect in a school or university exam.
Why CAT reporting needs care
CAT is conducted in multiple sessions and uses a scoring and normalisation process. The official scorecard and methodology govern reported scaled scores and percentiles. Do not assume that a raw score from one year maps to the same percentile the next year.
What normalisation does
In recent years, CAT has been held in more than one slot, and each slot has a different question paper. One slot's paper may turn out a little harder than another's. Normalisation adjusts scores so that candidates are not advantaged or disadvantaged by the slot they happened to take. Results are then reported as scaled scores and percentiles, both overall and by section.
Two practical consequences follow:
- Your raw marks from an answer key are an estimate, not your final scaled score.
- “Percentile predictors” that circulate after the exam are guesses based on limited data. Treat them as rough indications at best.
Sectional percentiles matter too
Many institutes look at sectional performance as well as the overall figure, and some programmes set sectional thresholds. A strong overall percentile built on one excellent section and one very weak section can limit your options. Check each programme's admission policy, and see our CAT exam guide and syllabus guide for how the sections are structured.
What students should track
During preparation, track accuracy, attempt quality, sectional strengths and mock percentile trends. Compare mock percentiles cautiously because different test series attract different groups of students.
A single mock percentile is noisy. A trend over several mocks is far more useful, especially when you read it alongside accuracy. Here is an example of a simple tracker:
| Mock | Attempts | Accuracy | Raw score | Mock percentile | Main takeaway |
|---|---|---|---|---|---|
| 1 | 30 | 63% | 46 | 71 | Too many guesses in QA |
| 2 | 26 | 77% | 54 | 80 | Fewer attempts, better accuracy |
| 3 | 29 | 72% | 55 | 76 | Harder DILR; selection held up |
| 4 | 31 | 81% | 69 | 88 | Accuracy and attempts both rising |
(Illustrative numbers. Accuracy here is correct answers divided by attempts, and every wrong answer is assumed to be an MCQ.)
Mock 3's percentile dipped, but the raw score and accuracy held steady while the student handled a harder DILR section well. Without the other columns, that dip might have caused unnecessary worry. Our guide to analysing CAT mock tests explains how to turn these columns into a plan.
Using percentage the right way
Percentage is still useful, just not as a measure of standing. Use it inside your own data: accuracy by section, accuracy by topic and the share of time spent on questions you got wrong. Percentage problems themselves are also a core QA topic; the percentages formula sheet covers the essentials.
Frequently asked questions
Does 99 percentile in CAT mean 99% marks?
No. A 99th percentile means you performed at or above about 99% of candidates. The marks needed for it depend on the paper and the cohort, and in recent years they have usually been far below 99% of the maximum.
How is the CAT percentile calculated?
CAT normalises scores across slots, converts them to scaled scores and then reports percentiles based on candidates' relative standing. The exact method is described in official CAT documentation on iimcat.ac.in, which is the only reliable source.
What is the difference between raw score, scaled score and percentile?
The raw score is your marks under the marking scheme. The scaled score is that performance after normalisation across slots. The percentile is your standing relative to other candidates.
Is mock percentile a reliable predictor of my CAT percentile?
Only roughly. Different test series attract different groups of students and use papers of different difficulty. Trends across several mocks, read with accuracy and sectional scores, are more informative than any single number.
Try this on Pareto Academy
Our mock analysis shows accuracy, attempts and time for each section, which tell you more about what to fix than a single percentile. Start free to see yours.