In this guide
Every January, the same confusion plays out. Two candidates with similar raw marks end up several percentile points apart. Someone with a "good" score misses a call because of one section. Someone else gets a higher percentile than a friend who solved more questions, because they sat a different slot.
None of this is random. CAT reports your performance through three numbers, and each one does a different job. Once you know what they are, score–percentile tables stop being mysterious and you can set targets that actually help your preparation.
Three numbers, three jobs
| Number | What it is | Who uses it |
|---|---|---|
| Raw score | Marks from your answers: +3 right, −1 wrong MCQ, 0 wrong TITA or unanswered | Nobody outside the exam system |
| Scaled score | Raw score adjusted for differences between slots | Appears on your scorecard |
| Percentile | Where your scaled score stands against all candidates | B-schools, for cut-offs and shortlisting |
Percentiles are reported for the whole test and for each section. The sectional ones matter as much as the overall figure, because most IIMs set sectional cut-offs.
Raw score
The raw score comes straight from the marking scheme.
Worked example: in VARC a candidate attempts 16 MCQs and 3 TITA questions. They get 13 MCQs right, 3 wrong, and 2 TITA right, 1 wrong.
- MCQ: 13 × 3 − 3 × 1 = 39 − 3 = 36
- TITA: 2 × 3 = 6
- VARC raw score: 36 + 6 = 42
The overall raw score is simply the sum of the three sectional raw scores.
Scaling, or why slots are normalised
CAT is held in three slots with three different papers. However carefully the papers are set, one will turn out slightly harder than another. If raw scores were compared directly, a candidate who happened to sit the harder slot would be at a disadvantage.
Normalisation fixes this. Each slot's scores are adjusted so that candidates of similar ability in different slots end up with comparable scaled scores. The key assumption is that each slot contains a similar spread of candidates, which is reasonable when lakhs of people are allotted to slots.
What this means for you:
- A tough slot does not automatically hurt you. If everyone found it hard, a modest raw score can translate into a good scaled score.
- Your scaled score can differ from your raw score, in either direction.
- You cannot calculate your scaled score yourself from the answer key. Coaching estimates are approximations.
Percentile
A percentile tells you what share of candidates scored below you. A 95 percentile means roughly 95% of test-takers had a lower scaled score. It does not mean you got 95% of the marks. In recent years even very high percentiles have typically needed less than half the maximum marks.
Worked example: suppose 2,50,000 candidates take the test.
- 99 percentile: about 1% are at or above you, so roughly 2,500 candidates.
- 95 percentile: about 5%, roughly 12,500 candidates.
- 90 percentile: about 10%, roughly 25,000 candidates.
This shows why the top end is so crowded. The gap between 95 and 99 percentile is about 10,000 people, and a few marks can move you across many of them.
Sectional percentiles
Each section gets its own percentile, calculated the same way from sectional scaled scores. A candidate can have 98 overall and 80 in DILR if their other two sections were strong. At an IIM with an 80 sectional cut-off, that candidate scrapes through; at one with 85, they do not.
Why score–percentile tables change every year
Coaching institutes publish tables such as "X marks for 99 percentile". They are useful as rough guides and dangerous as targets, because the marks needed depend on:
- how hard that year's papers were
- how many candidates appeared
- how well everyone else performed
- how the slot normalisation worked out
A score that gave 99 percentile in a tough year may give noticeably less in an easier year. Use old tables to understand the shape of the curve, not to fix a precise number.
| What people say | What is actually true |
|---|---|
| "I need 50% marks for 99 percentile" | The marks needed change every year |
| "My slot was hard, so I am finished" | Scaling adjusts for slot difficulty |
| "My overall percentile is all that matters" | Sectional cut-offs can knock you out |
| "Percentile is my percentage" | Percentile compares you with others |
Setting targets that help
A precise mark target is less useful than a process target. Aim for:
- Accuracy in every section, around 80% or better on your attempts.
- No weak section. Raising a 70 sectional percentile to 85 often matters more than pushing a strong section higher.
- Consistent mock performance. Track your percentile in a large national mock series, which compares you with many serious candidates.
Our guide to mock tests and analysis shows how to track these numbers week by week.
Practice set
- A candidate gets 15 MCQs right, 5 wrong and 3 TITA right in QA. What is the raw score?
- With 3,00,000 candidates, roughly how many are at or above a 98 percentile?
- A friend says "90 percentile means I scored 90%". Correct them in one sentence.
- Candidate P scores 40 raw in a hard slot; candidate Q scores 44 raw in an easier slot. Can you say Q will have the higher percentile?
Answers:
- 15 × 3 − 5 + 3 × 3 = 45 − 5 + 9 = 49.
- 2% of 3,00,000 = 6,000.
- It means about 90% of candidates scored lower than you, whatever your marks were.
- No. After scaling, P's 40 in the harder slot may be worth as much as, or more than, Q's 44.
What to do next
- Work out your raw score by hand for your last mock, section by section.
- Note your weakest sectional percentile across the last five mocks and make it your main target.
- Read the percentile cut-offs in two IIM admission policies; see how IIMs shortlist.
- Revisit the CAT exam pattern to plan how many questions you need per section.
A note on dates and numbers. Exam patterns, vacancies and schedules change from year to year. Always confirm the current details in the latest notification on the Indian Institutes of Management website .
Get the next CAT guide by email
New guides every week. No spam, unsubscribe any time.