KCET RankPredictor
Built on real KCET rank data

What rank would your marks fetch?

Enter expected Class 12 PCM marks (Karnataka 2nd PUC, CBSE or ICSE) and KCET marks. Your overall score is 50% board + 50% KCET, matched against real ranks from past KCET years.

Your marks

–
/ 300

Karnataka 2nd PUC, CBSE or ICSE · out of 300, used as-is

–
/ 180

Physics, Chemistry, Maths · 60 each, out of 180

Overall score –

–

50% from board PCM %, 50% from KCET %.

– –

Your marks would have fetched

≈–

–

–

See rank curves

Year by year

Left is a better rank · tap for details

    Rank curves

    Rank on a log scale, #1 at the top. Solid lines follow recorded entries; dotted parts are extended from other years. Drag across the chart to compare any score.

    Good to know

    • Overall score = 50% board + 50% KCET

      Board % is your Class 12 PCM marks ÷ 300 × 100, and KCET % is your KCET marks ÷ 180 × 100. The overall score takes 50% of each, the same weighting used for KCET engineering ranks.

    • Every board is treated the same

      Class 12 PCM marks from the Karnataka State Board (2nd PUC), CBSE or ICSE are taken as-is. They aren't normalised or adjusted for how easy or hard a board's exams are.

    • 2024: KCET was scored out of 141

      In KCET 2024, questions found to be erroneous were excluded, so the maximum score was 141 instead of 180. Every 2024 entry in the data reflects this. The engine takes care of it: it compares years on percentages, not raw marks, so 2024 ranks line up with the other years and your KCET marks out of 180 are matched on the same scale.

    Questions about eligibility, documents or the admit card? Read the KCET FAQ
    How the estimate works
    1. Overall score. 50% of your board PCM % (marks ÷ 300) plus 50% of your KCET % (marks ÷ 180). Board marks are used as-is for every board. Years are compared on this percentage, which is why 2024 (KCET out of 141) needs no special handling.
    2. One rank curve per year. Each year's recorded (overall %, rank) entries from the sheet form that year's curve. Entries that are clearly mistyped, such as a % above 100 or a rank that contradicts many nearby entries, are set aside (listed below).
    3. Anchored to the nearest recorded %. Between recorded entries, the rank is interpolated smoothly on a log scale. The curve passes through the real entries and always gives a better rank for a higher score.
    4. Beyond a year's data. Above the best or below the lowest recorded score, the curve continues using the shape of the other years, scaled to match that year. These parts are marked with lower confidence.
    5. Likely range. Reflects how much real entries scatter near your score, and widens the further you are from recorded data. In a hold-out test (hide each real entry, predict it from the rest), the median error was 0.5–3% depending on the year.
    6. Next-year outlook. A band between "same as the latest year" and "the recent year-on-year trend continues". It's the least certain figure here.
    Data used