JEE Counselling Logo
JEE Counselling

2026

JEE Main Normalization Formula Explained — Full Methodology

A deep, technical breakdown of how NTA converts raw JEE Main marks into a fair percentile score across shifts, with a worked example.

Why Raw Marks Cannot Be Compared Directly Across Shifts

JEE Main is conducted by the National Testing Agency (NTA) across multiple shifts spanning several days within each session, simply because no single testing infrastructure in India can accommodate over a million candidates in one sitting. Each shift receives a different but equivalent-difficulty question paper. Despite NTA's rigorous paper-setting and moderation process, perfect difficulty parity across every single shift is practically impossible — some papers will always be marginally tougher or easier than others purely due to the inherent variability of question design.

If NTA simply ranked all candidates by raw marks regardless of which shift they appeared in, a candidate in an easier shift would have an unfair statistical advantage over an equally capable candidate who happened to be assigned a tougher shift. To eliminate this bias, NTA does not use raw marks for the final merit list. Instead, it converts each candidate's performance into a percentile score, calculated independently within their own shift, and it is this percentile — not raw marks — that determines rank and eligibility for JEE Advanced.

The Normalization Formula

The percentile score for a candidate in a given shift is calculated using the following formula:

Percentile Score = (Number of candidates in the shift with raw score ≤ the candidate's raw score ÷ Total number of candidates in that shift) × 100

In simpler terms, your percentile tells you what percentage of candidates within your specific shift scored equal to or lower than you. This calculation is performed separately and independently for every shift across both sessions. The result is reported up to seven decimal places on the official scorecard to minimize the number of candidates who end up with an identical percentile.

Worked Example

Suppose a particular shift had exactly 100,000 candidates, and a candidate named Rohan scored raw marks such that 97,500 candidates in that same shift scored equal to or lower than him. Applying the formula:

Percentile = (97,500 ÷ 100,000) × 100 = 97.5000000 percentile

Now consider a different shift with 80,000 candidates, where another candidate, Priya, had 78,000 candidates scoring equal to or below her raw marks:

Percentile = (78,000 ÷ 80,000) × 100 = 97.5000000 percentile

Even if Rohan's and Priya's raw marks were numerically different (because their shifts had different difficulty levels and different total candidate counts), they end up with the identical percentile of 97.5, because each outperformed the same relative proportion of candidates in their own shift. This is precisely the fairness that normalization is designed to achieve.

Common Misconceptions About Normalization

MisconceptionReality
Percentile = percentage of marks scoredPercentile is relative standing among your shift's candidates, unrelated to percentage of marks.
NTA adjusts/inflates your raw marksRaw marks are never altered. Only the percentile calculation uses relative comparison.
Easier shifts always give unfairly higher percentilesNormalization specifically exists to neutralize this — equal relative performance yields equal percentile regardless of shift difficulty.
My percentile depends on candidates across all shifts combinedPercentile is calculated purely within your own shift, not against the entire national candidate pool at once.
100 percentile means a perfect score100 percentile only means you were among the top scorer(s) in your specific shift — not necessarily a perfect 300/300.

How Percentile Feeds Into Your Final Rank

Once every candidate's shift-wise percentile has been computed for both Paper 1 subjects combined and for individual subjects, NTA uses the overall percentile (and, where needed, subject-wise percentiles and other tie-breaking criteria) to prepare the final merit list and All India Rank (AIR). If a candidate appeared in both Session 1 and Session 2, the higher of their two overall percentiles is used for this final ranking, as explained in detail on our multi-session percentile page.

This percentile-driven rank is what determines JEE Advanced eligibility (typically the top approximately 2.5 lakh candidates, category-wise) and is also the score used by JoSAA, CSAB, and most state counselling bodies that accept JEE Main scores for NIT/IIIT/GFTI and affiliated college admissions.

JEE Main
JEE Advanced
DASA / NRI

Know which college you can get

Predict IITs, NITs, IIITs & GFTIs · 135+ institutes · JoSAA 2026 data

Predict My College

Frequently Asked Questions – JEE Main Normalization Formula

Common queries about NTA's percentile normalization methodology

1
Is JEE Main percentile the same as percentage of marks?
No, and this is the single most common misconception. Percentile reflects your relative standing compared to all other candidates in your shift, while percentage of marks is simply your marks out of the total marks. A candidate can score 250/300 (about 83% marks) but still have a percentile below 99 if many other candidates in that shift scored similarly high or higher. Percentile and percentage are entirely different measures.
2
Why doesn't NTA just compare raw marks directly across shifts?
Because JEE Main is held across dozens of shifts spread over multiple days, and no two shifts have exactly identical question papers or exactly identical difficulty levels, even with careful paper-setting. A shift with a marginally tougher paper would unfairly disadvantage its candidates if raw marks were compared directly against a shift with an easier paper. Percentile normalization corrects for this by measuring relative performance within each shift rather than absolute marks across shifts.
3
Can two candidates with different raw marks have the same percentile?
Yes, if they appeared in different shifts. A candidate scoring 220/300 in a tougher shift might get the same percentile as a candidate scoring 235/300 in an easier shift, because both may have outperformed the same proportion of candidates within their respective shifts. This is the entire point of normalization — to make cross-shift comparison fair.
4
Can the highest scorer in a shift ever get less than 100 percentile?
No. By definition, the highest scorer(s) in each shift receive exactly 100 percentile for that shift, since no candidate in that shift scored higher than or equal to them among the rest — they represent the top of the distribution. If multiple candidates tie for the top marks in a shift, all of them receive 100 percentile.
5
Does normalization change my actual raw marks?
No. Your raw marks (out of 300 for Paper 1) remain exactly as scored and are shown separately on your scorecard alongside your percentile. Normalization only affects how your performance is translated into a percentile score for the purpose of preparing the merit list — it does not alter or 'curve' your actual marks.
6
Is normalization unique to JEE Main, or used elsewhere too?
Percentile-based normalization across multiple shifts/sessions is a common statistical technique used by NTA for several large multi-shift exams it conducts, not just JEE Main. It is a widely accepted method in large-scale standardized testing wherever a single-day, single-shift administration is not logistically feasible.
7
Where can I verify the exact number of candidates in my shift?
NTA does not typically publish a shift-wise breakdown of candidate counts publicly in real time, though aggregate session-wise numbers are usually shared in official press notes after each session. Your percentile itself, shown on your scorecard, already reflects this calculation — you do not need the raw shift-wise count to interpret your result.

JEE College Predictor 2026

IITs · NITs · IIITs · GFTIs · 135+ colleges

Predict

Disclaimer

Methodology explained is based on NTA's publicly stated normalization approach. Verify specifics for your session at jeemain.nta.nic.in.