The NTA does not publish marks. It publishes a percentile, and every JEE Main aspirant then spends a week trying to work out what that percentile means for their rank, their category and their chances.

Here is how the two relate, and the arithmetic that turns one into the other.

Percentile is not percentage

This is the misunderstanding behind most of the panic every January.

Percentage is how much you scored: 180 out of 300 is 60%.

Percentile is how many candidates you beat. A 99.2 percentile means you performed better than 99.2% of everyone who sat the exam. It says nothing at all about your marks.

In a hard session, 160 marks might be 99 percentile. In an easy one, the same 160 might be 96. That is the whole point of the system: it corrects for the fact that different shifts get different papers.

Turning percentile into rank

The formula is simple arithmetic:

Rank ≈ (100 − percentile) ÷ 100 × total candidates

With roughly 14 lakh candidates and a 99 percentile: (100 − 99) ÷ 100 × 14,00,000 = 14,000.

PercentileApproximate rank (14 lakh candidates)
99.91,400
99.57,000
9914,000
9828,000
9570,000
901,40,000
852,10,000

Notice how much a decimal is worth at the top. Between 99.9 and 99.5 sit around 5,600 candidates. At that end of the table, the third decimal place decides which institute you attend.

The JEE Main rank predictor does this conversion and also estimates your category rank, which is the number that actually governs a reserved seat.

Why two sessions complicate it

JEE Main runs in sessions, and most candidates sit more than one. Your final percentile is the best of your sessions — not the average, and not the latest.

Two things follow:

  • A weak first session costs you nothing except confidence, provided you sit the second.
  • Percentiles are computed within each session, then merged. This is why a session-1 percentile and a final percentile for identical marks can differ.

Category rank matters more than AIR for most candidates

The Common Rank List is one list. JoSAA seat allocation runs on several, and if you belong to a reserved category, your category rank is the one that decides your seat for reserved quota.

Category ranks are dramatically better than the CRL, because each list contains only that category's candidates. An approximate feel:

CategoryRoughly, for a CRL of 50,000
General (CRL)50,000
GEN-EWS~5,500
OBC-NCL~16,000
SC~7,000
ST~3,000

These proportions shift each year with the candidate mix, so treat them as a range rather than a figure. The predictor gives a band for exactly that reason.

Two more factors people forget

Home state quota. Half the seats at each NIT are reserved for candidates domiciled in that state. Your chances at the NIT in your own state are meaningfully better than your all-India rank suggests, and this is often the single biggest lever available to a mid-rank candidate.

Female supernumerary seats. NITs and IIITs run additional seats for female candidates, typically around 20% of intake. Those close at noticeably higher ranks than the general pool.

The JEE Advanced cut-off is a percentile, by category

Qualifying for Advanced means being inside the top ~2.5 lakh of Main, expressed as a category-wise percentile cut-off. Recent years have sat approximately at:

CategoryApproximate qualifying percentile
General~93.2
GEN-EWS~81.3
OBC-NCL~79.6
SC~60.1
ST~46.7

These move a point or two each year. The official cut-off is published with the results, and that is the only one that counts.

What a predicted rank is good for

It is good for building a JoSAA choice list before the counselling window opens, so you are not researching institutes under time pressure. It is good for deciding whether a state counselling application is worth submitting as a backup.

It is not good for deciding whether to drop a year. Predictions carry a real error band, sessions vary, and the official result arrives soon enough. Make irreversible decisions on the official number, never an estimate — including ours.