How JEE Main percentiles and normalisation work

Suppose two friends write JEE Main on different days. One scores 170 marks and gets a 97.8 percentile. The other scores 185 and gets 97.2. The second friend is sure something has gone wrong. Nothing has. They wrote different papers, and the percentile is doing exactly what it was designed to do.

The JEE Main scorecard shows a percentile, not marks, and many candidates misread it. Once you understand the calculation, you can read your result calmly, set sensible targets and stop trusting every "marks vs percentile" table you find online.

Why marks alone would be unfair

Each JEE Main session runs over several days, usually with two shifts a day, and every shift has a different question paper. However carefully the papers are set, some will be harder than others. If ranks were based on raw marks, a candidate who happened to get a tough shift would be penalised for bad luck.

NTA's answer is normalisation: compare each candidate only with the others who wrote the same paper, then merge the shifts on that common scale.

The NTA score formula

Within each shift, your raw marks are converted to an NTA score using this formula:

Three things follow from it:

• The top scorer in every shift gets 100, whatever their raw marks.
• A percentile of 97 means 97% of candidates in your shift scored the same as you or less. It is not 97% of the marks.
• Scores are calculated to seven decimal places to reduce ties.

NTA also calculates percentiles for Physics, Chemistry and Mathematics separately. These are used for tie-breaking, not for your main rank.

Worked examples

Example 1: calculating an NTA score. Suppose 1,20,000 candidates write a shift, and 1,14,000 of them score the same as or less than a particular candidate.

NTA score = 100 × 1,14,000 ÷ 1,20,000 = 100 × 0.95 = 95.

Example 2: same marks, different percentiles. Suppose 180 marks is at or above 97% of candidates in an easier shift, but at or above 98.5% in a harder shift. A score of 180 gives about 97 in the first and about 98.5 in the second. Candidates in the harder shift found fewer questions they could solve, so 180 there represents a stronger relative performance.

Example 3: from percentile to approximate rank. A rough estimate uses: rank ≈ ((100 − P) ÷ 100) × N + 1, where P is the percentile and N is the number of candidates.

Suppose N = 12,00,000 (an illustrative figure) and P = 99.

Rank ≈ (1 ÷ 100) × 12,00,000 + 1 = 12,000 + 1 = about 12,001.

This is only an estimate. The real number of candidates changes every year, and ranks come from merged scores across all shifts, not from one formula.

(All numbers in these examples are made up to show the method.)

Two sessions, one rank

If you appear in both sessions, NTA uses the better of your two overall NTA scores to prepare your rank. It does not average them, and it does not mix subjects across sessions.

This has two practical consequences:

1. A poor January session does not hurt you if April goes better.
2. There is little downside to writing both sessions, apart from the time and cost. For most candidates the second attempt is worth it.

Ranks and tie-breaking

After the final session, NTA publishes an All India Rank (AIR) and category ranks. JoSAA counselling for NITs, IIITs and GFTIs uses these ranks.

When two candidates have the same overall NTA score, NTA applies tie-breaking rules published in the bulletin — in recent years based on subject-wise scores and answer accuracy. The order has been revised before, so read it in your year's bulletin.

Percentile, percentage and rank compared

Common misreadings

• "99 percentile means I scored 99%." No. It means 99% of your shift scored the same as or less than you. The marks behind it vary by shift and by year.
• "My friend got more marks but a lower percentile, so NTA made a mistake." Almost certainly not. They wrote a different, easier paper.
• "Last year's table says 150 marks = 97 percentile, so I'm safe." Marks-to-percentile tables shift every year and every shift. Treat them as a rough range, never a promise.
• "Percentile predicts my college." Only rank does, and only in combination with category, home state quota and the previous year's closing ranks at JoSAA.

How to use this in preparation

Set targets in marks, not percentiles. In mocks, track raw marks by subject and section. A steady rise from, say, 140 to 190 over three months is real progress, whatever percentile the test series shows.

Think in ranges. If recent years suggest a band of marks for your target percentile, aim for the top of the band. You cannot control how hard your shift is.

Do not panic about your shift. A paper that feels brutal to you was brutal for everyone in that shift, and normalisation accounts for it. Your job in the hall is only to maximise your own marks. Our time strategy guide helps with that.

Quick self-test

1. In a shift of 90,000 candidates, 88,200 score the same as or less than you. What is your NTA score?
2. You score 96.4 in January and 98.1 in April. Which is used for your rank?
3. Can two candidates with the same raw marks get different percentiles?
4. With an illustrative 10,00,000 candidates, roughly what rank does 98 percentile suggest?

Answers

1. 100 × 88,200 ÷ 90,000 = 100 × 0.98 = 98.
2. 98.1 — the better of the two.
3. Yes, if they wrote different shifts.
4. (2 ÷ 100) × 10,00,000 + 1 = about 20,001.

What to do next

• Read the normalisation section of the current NTA bulletin once.
• Start a mock tracker in marks, split by subject and section. See mock tests and analysis.
• Set a marks target for your next five mocks, not a percentile target.
• Before counselling, study JoSAA closing ranks, not percentiles. Our JoSAA counselling guide explains how.