How the RankMatrix college predictor uses JoSAA cutoffs

What the predictor does

RankMatrix builds a shortlist by comparing your inputs with recorded JoSAA opening and closing ranks. It does not calculate a probability of admission, simulate the full JoSAA allocation process or forecast a future closing rank. The comparisons are developed and run by RankMatrix using its stored dataset, not by a JoSAA tool or prediction service.

Inputs and comparison window

The request includes JEE Main rank, optional JEE Advanced rank, category, seat pool, state-code input, a percentage delta and a result count. The backend computes a lower bound by rounding down rank × (1 − delta / 100), and an upper bound by rounding up rank × (1 + delta / 100). A wider delta is a wider search window, not a confidence level.

For example, rank 10,000 with delta 10 gives a 9,000–11,000 window. These are hypothetical input bounds, not published cutoffs or an assertion of admission chances. Use the rank from the same exam and category list as the cutoff row being compared.

Which year and round are used?

The predictor queries the backend's configured latest JoSAA rank model and its maximum stored round. It does not average several years or default to round four. Older years remain useful for manual comparisons in the rank explorer and institute–programme detail pages. A new model year needs imported data before it can produce useful predictions.

How rows are selected and ordered

1. Match category, seat pool and the maximum stored round; exclude preparatory-course rows.
2. Use the Advanced-rank window for participating IIT and IISc programme rows and the Main-rank window for other rows. IISc is stored in the backend's IIT filter bucket; it is labelled separately in the interface.
3. Identify home-quota institutes using stored institute-state and quota relationships.
4. Prioritise rows whose opening/closing interval overlaps the input bounds, followed by additional rows above the window, using the backend's selection tiers.
5. Limit to the requested count, then display the selected rows ordered by opening rank.

The selection tiers are implementation priorities, not labels of safe, likely or guaranteed admission. The endpoint does not cross-check each result against current seat counts, perform document verification or evaluate every exceptional eligibility rule.

What to verify before filling choices

• Match the exam, category rank list, quota and seat pool in every comparison.
• Inspect the year and round. A historical interval can change in a future allocation.
• Check the published seat matrix and official programme restrictions.
• Arrange eligible choices by your genuine preferences. The shortlist is not a recommended preference order.

Sources, refreshes and corrections

Our JoSAA dataset consists of published opening and closing ranks, seat counts, and institute and programme details. A cutoff row describes an institute, programme, quota, category, seat pool, year, and round. It does not identify an individual candidate: we do not obtain candidate names, roll numbers, application IDs, or personal rank records from JoSAA. The ranks visitors enter for comparisons are separate inputs; their handling is described in the privacy policy.

Rank and seat data originate from JoSAA publications. Importing an official release and refreshing a cached web page are separate operations. Detail pages can revalidate daily, but that does not mean new official data have been imported daily. A displayed JoSAA year identifies the dataset, not the date of the last import. Please report discrepancies with the exact year, round and filters.

RankMatrix is an independent reference tool. Complete registration, choice filling, payments and reporting on the official portal under the current rules. See the terms and privacy policy.