How IIMs shortlist after CAT: composite scores

A 98.5 percentile does not decide an IIM interview call on its own. For the 2027–29 batch, IIM Ahmedabad gives CAT 65 % of its shortlisting score, IIM Bangalore 55 of 100 points and IIM Calcutta 56 of 85, and each fills the rest from your school and degree marks, your work experience and gender diversity, in its own way.

This lesson works the three published formulas on one invented profile, so you can compute a composite scorecomposite score (admissions): A single number an institute computes from several inputs, such as CAT scores, academic marks, work experience and gender diversity, each with a published weight, to rank applicants for a shortlist or an offer. yourself, see how its parts trade off and read a published minimum percentile for what it is. It also marks the figures of the applicant pool that every formula needs and no candidate can know, which is why no calculator can tell you whether you will be shortlisted.

Each IIM sets its own stages after one common test

Every IIM reads the same CAT result, and that is where the common ground ends. The CAT 2026 Information Bulletin says that each IIM sets its own eligibility criteria, including academic cut-offs and relative weights, and follows its own selection process. CAT performance is an important component, and an IIM may also use your earlier academic record and relevant work experience. IIM Ahmedabad’s policy adds that it shortlists independently of the other IIMs, so the same candidate can be on one IIM’s list and not on another’s.

The stages still follow one shape. Step through it with the three policies of the 2027–29 batch:

Two differences in the order matter. IIM Bangalore and IIM Calcutta remove candidates below their minimum percentiles before they compute anything, while IIM Ahmedabad computes every composite first and applies its CAT minimums to the normalised scores. And each institute computes a new score after the interviews, with new weights, so the composite that earns a call is not the score that decides an offer. The next lesson works through those final formulas.

IIM Ahmedabad rates your record out of 38

IIM Ahmedabad’s first step is an application ratingapplication rating (IIM Ahmedabad): IIM Ahmedabad's points for an applicant's record: bands of 10th, 12th and bachelor's marks, months of work experience and gender diversity, added to a rating of at most 38 and then divided by the average of the pool's 50 highest ratings. built from five entries of your CAT form: your 10th, 12th and bachelor’s percentages, your months of full-time work experience and your gender. Each of the three marks earns points by band, from one table:

Work experience earns nothing below 12 months, then 0.2 points for every month above 11, which reaches 5 points at 36 months; more than 36 months still earns 5. Every candidate who is not male receives 3 points for gender diversity. The highest possible rating is therefore 10 + 10 + 10 + 5 + 3 = 38.

Step through the rating of an invented profile: 92 % in the 10th, 88 % in the 12th, 74 % in the degree, 24 months of work, male. The program works in tenths of a point, so every value it holds is a whole number:

# IIM Ahmedabad's rating rules for
# PGP 2027-29, in tenths of a point
BANDS = [(90, 100), (80, 80),
         (70, 50), (60, 30),
         (55, 20)]
rating = 0
for pct in [92, 88, 74]:
    points = 10
    for floor, tenths in BANDS:
        if pct > floor:
            points = tenths
            break
    rating += points
months = 24
if months > 36:
    rating += 50
elif months >= 12:
    rating += 2 * (months - 11)
male = True
if not male:
    rating += 30
print("rating:", rating / 10)

rating: 25.6

The profile scores 10 + 8 + 5 + 2.6 + 0 = 25.6 of 38. Its degree costs it most: 74 % is worth 5 of 10, and every degree mark above 70 and up to 80 earns the same 5.

The rating is weighed against CAT, 35 to 65

A rating of 25.6 means little until it is compared with other applicants’ ratings, and that is the next step. IIM Ahmedabad divides each rating by the average of the 50 highest ratings in the applicant pool, then adds 35 % of that normalised rating to 65 % of the overall CAT percentile divided by 100:

In words: a rating equal to the top-50 average together with a 100th percentile gives exactly 1. For the sample profile at the 98.5th percentile, with an assumed top-50 average of 36, it is 0.35 × 0.7111 + 0.65 × 0.985 = 0.8891. Last, each raw composite is divided by the average of the top 1 % of raw composites in the candidate’s own undergraduate discipline, one of 20 the policy lists, among them Engineering & Technology and Commerce, Accounting & Finance; that top 1 % holds at least 5 and at most 100 scores. The program runs all three published formulas on the profile and marks every figure that no policy publishes:

"""Three published shortlisting formulas for the 2027-29 batch, worked for one profile with exact fractions:
IIM Ahmedabad's composite score, IIM Bangalore's pre-interview score and IIM Calcutta's Stage II composite, each as
its own admission policy prints it. Every figure of the applicant pool that a formula needs but no policy publishes is
marked ASSUMED: change one and the result moves, which is why no program can predict a shortlist.
Press Edit, put in your own profile, and run it again."""
from fractions import Fraction as F

MARKS = {"10th": 92, "12th": 88, "degree": 74}  # percentages as the CAT form records them
MONTHS = 24  # months of full-time work experience
GENDER = "male"  # "male", "female" or "transgender"
PERCENTILE = F("98.5")  # overall CAT percentile

# Figures of the pool that the policies do not publish, each a guess you can change.
TOP_50_RATING = F(36)  # ASSUMED: IIM Ahmedabad, average of the 50 highest application ratings
TOP_1_COMPOSITE = F("0.95")  # ASSUMED: IIM Ahmedabad, average of the top 1 % raw composites in your discipline
POOL = {"10th": (88, 6), "12th": (85, 7), "degree": (72, 8)}  # ASSUMED: IIM Bangalore, (mean, sd) of the first shortlist
SCALED_SHARE = F("0.60")  # ASSUMED: IIM Calcutta, your total scaled score as a share of the maximum possible


def num(x, places=4):
    return f"{float(x):.{places}f}".rstrip("0").rstrip(".")


print("IIM Ahmedabad: an application rating, then the composite score")
BANDS = [(90, 10), (80, 8), (70, 5), (60, 3), (55, 2)]  # above the floor, up to the next band


def band(pct):
    return next((points for floor, points in BANDS if pct > floor), 1)


work = F(0) if MONTHS < 12 else F(5) if MONTHS > 36 else F(2, 10) * (MONTHS - 11)
diversity = 0 if GENDER == "male" else 3
rating = sum(band(m) for m in MARKS.values()) + work + diversity
for name, pct in MARKS.items():
    print(f"  {name} {pct} %: {band(pct)} points")
print(f"  work experience, {MONTHS} months: {num(work)} points; gender diversity: {diversity}")
print(f"  rating: {num(rating)} of 38")
normalised = rating / TOP_50_RATING
raw = F(35, 100) * normalised + F(65, 100) * PERCENTILE / 100
print(f"  ASSUMED top-50 average rating {num(TOP_50_RATING)}: normalised rating {num(normalised)}")
print(f"  raw composite: 0.35 x {num(normalised)} + 0.65 x {num(PERCENTILE / 100)} = {num(raw)}")
print(f"  ASSUMED top 1 % average in your discipline {num(TOP_1_COMPOSITE)}: composite {num(raw / TOP_1_COMPOSITE)}")
per_point = (F(35, 100) / TOP_50_RATING) / (F(65, 100) / 100)
print(f"  one more rating point is worth {num(per_point, 2)} CAT percentile points under these assumptions")

print()
print("IIM Bangalore: every part standardised against the first shortlist")


def standardise(value, mean, sd, weight):
    return max(F(0), min(F(weight), F(weight, 2) + F(value - mean, sd) * F(weight, 6)))


total = F(0)
for name, (mean, sd) in POOL.items():
    points = standardise(MARKS[name], mean, sd, 10)
    total += points
    print(f"  {name} {MARKS[name]} %, ASSUMED mean {mean} and sd {sd}: {num(points, 3)} of 10")
work_b = F(10) if MONTHS >= 36 else F(10 * MONTHS, 36)
total += work_b
print(f"  work experience, {MONTHS} months (10 x months / 36, not standardised): {num(work_b, 3)} of 10")
print(f"  these four parts: {num(total, 3)} of 40; CAT adds up to 55 and gender diversity up to 5")
print("  one sd above the mean adds weight / 6: VARC 3.167, DILR 3.5, QA 2.5, each mark (weight 10) 1.667")

print()
print("IIM Calcutta: points for school marks and a scaled CAT score")
ten = next((p for floor, p in [(80, 10), (75, 8), (70, 6), (65, 4), (60, 2)] if MARKS["10th"] >= floor), 0)
twelve = next((p for floor, p in [(80, 15), (75, 12), (70, 9), (65, 6), (60, 3)] if MARKS["12th"] >= floor), 0)
gender_c = 4 if GENDER in ("female", "transgender") else 0
cat_c = 56 * SCALED_SHARE
print(f"  class 10: {ten} of 10; class 12: {twelve} of 15; gender diversity: {gender_c} of 4")
print(f"  CAT: 56 x ASSUMED share {num(SCALED_SHARE)} = {num(cat_c)}; composite {num(ten + twelve + gender_c + cat_c)} of 85")
print("  no degree marks and no work experience at this stage")

IIM Ahmedabad: an application rating, then the composite score
  10th 92 %: 10 points
  12th 88 %: 8 points
  degree 74 %: 5 points
  work experience, 24 months: 2.6 points; gender diversity: 0
  rating: 25.6 of 38
  ASSUMED top-50 average rating 36: normalised rating 0.7111
  raw composite: 0.35 x 0.7111 + 0.65 x 0.985 = 0.8891
  ASSUMED top 1 % average in your discipline 0.95: composite 0.9359
  one more rating point is worth 1.5 CAT percentile points under these assumptions

IIM Bangalore: every part standardised against the first shortlist
  10th 92 %, ASSUMED mean 88 and sd 6: 6.111 of 10
  12th 88 %, ASSUMED mean 85 and sd 7: 5.714 of 10
  degree 74 %, ASSUMED mean 72 and sd 8: 5.417 of 10
  work experience, 24 months (10 x months / 36, not standardised): 6.667 of 10
  these four parts: 23.909 of 40; CAT adds up to 55 and gender diversity up to 5
  one sd above the mean adds weight / 6: VARC 3.167, DILR 3.5, QA 2.5, each mark (weight 10) 1.667

IIM Calcutta: points for school marks and a scaled CAT score
  class 10: 10 of 10; class 12: 15 of 15; gender diversity: 0 of 4
  CAT: 56 x ASSUMED share 0.6 = 33.6; composite 58.6 of 85
  no degree marks and no work experience at this stage

Under these assumptions one rating point is worth about 1.5 CAT percentile points: it adds 0.35 ÷ 36 to the raw composite, while each percentile point adds 0.65 ÷ 100. So two applicants with 88 % and 91 % in the 12th, alike in everything else, start about 3 percentile points apart, since the second earns 10 points for the 12th instead of 8.

IIM Bangalore scores each part against the first shortlist

IIM Bangalore starts by cutting the pool. Its first shortlist keeps only candidates with a positive raw score in every section of CAT 2026 and a pre-determined minimum in every section, and its policy prints the percentiles it applied to CAT 2025 (the last section compares them). Every later step, it says, is limited to the first shortlist alone. For those candidates it builds a pre-interview score of 100 points:

The policy turns each raw score into points with one formula, a standardised scorestandardised score: A score measured by its distance from a group's mean in standard deviations, so that marks from different boards or tests can be compared. IIM Bangalore turns it into points between 0 and a part's weight. capped between 0 and the part’s weight w:

In words: a candidate exactly at the mean of the first shortlist earns half the weight, each standard deviation above the mean adds a sixth of the weight, and three standard deviations above or below reach the cap or the floor. With weight 10, a mean of 75 and a standard deviation of 10, a mark of 85 earns 5 + 1.667 = 6.667, a mark of 95 earns 8.333 and a mark of 50 earns 0.833. Work experience has a formula of its own: 10 × months ÷ 36 below 36 months, counted up to July of the CAT year, and 10 from 36 months on, so 24 months give 6.667. The program’s second block applies these rules to the sample profile, with means and standard deviations it has to assume.

Two more rules widen the gate. The top ten candidates by total CAT score, by adjusted bachelor’s score in commerce, arts and science, and by professional course score qualify for the interview stage automatically. And a candidate with both a professional course and work experience gets the higher of the two scores, not both.

IIM Calcutta adds points for school marks to a scaled CAT score

IIM Calcutta’s Stage I keeps candidates who meet its section-wise and overall minimum percentiles with no negative raw score in any section. Its Stage II composite has 85 points and leaves out two things the other two policies count, the bachelor’s degree and work experience:

These bands are closed at the bottom, the reverse of IIM Ahmedabad’s: 80.0 % earns IIM Calcutta’s top band but only 5 of IIM Ahmedabad’s points. The sample profile earns all 25 points for its school marks, so everything else it can gain here comes from CAT. The CAT term needs the maximum possible scaled score, which the policy does not state, so the program assumes a share of 0.60 and gets 33.6 + 25 = 58.6 of 85.

Published minimums are floors, and the pool decides the rest

Every policy prints minimum percentiles, and each says in its own words that they are not the cut-offs. For General candidates who are not PwD the minimum percentilesminimum percentile: The lowest CAT percentile, overall or in a section, at which an institute will consider a candidate at a stage. It is a floor for being considered, not the score of the last candidate called. are:

IIM Calcutta writes that the actual cut-offs may be much higher, depending on how many candidates it calls in each category; IIM Bangalore notes under its table that the scores of those called for interviews were higher; IIM Ahmedabad says its cut-offs depend on the applicant pool and the spread of CAT scores, and may change. A floor only lets a profile be ranked. Run four invented profiles against the three tables:

"""The published minimum CAT percentiles for General (open) candidates who are not PwD, as three admission policies
for the 2027-29 batch print them, checked against four invented profiles. IIM Bangalore's row is the table its policy
prints for the previous cycle (CAT 2025); for CAT 2026 it says only that its first shortlist needs a pre-determined
minimum in every section. Clearing a floor only lets a profile be ranked; it shows nothing about a shortlist.
Press Edit, add your own percentiles, and run it again."""

FLOORS = {  # (VARC, DILR, QA, overall)
    "IIM Ahmedabad": (85, 85, 85, 95),
    "IIM Bangalore (previous cycle)": (80, 75, 75, 85),
    "IIM Calcutta": (80, 80, 75, 85),
}
SECTIONS = ("VARC", "DILR", "QA", "overall")
PROFILES = {  # invented percentiles
    "Asha": (97.2, 96.4, 94.1, 98.5),
    "Bharat": (99.1, 92.0, 82.3, 97.6),
    "Chitra": (78.5, 99.2, 99.4, 96.8),
    "Dev": (90.0, 88.0, 91.0, 94.6),
}

for person, scores in PROFILES.items():
    print(f"{person}: " + ", ".join(f"{s} {v}" for s, v in zip(SECTIONS, scores)))
    for institute, floor in FLOORS.items():
        short = [f"{s} {v} < {f}" for s, v, f in zip(SECTIONS, scores, floor) if v < f]
        print(f"  {institute}: " + ("clears every floor" if not short else "below the floor: " + "; ".join(short)))
cleared = sum(all(v >= f for v, f in zip(scores, floor)) for scores in PROFILES.values() for floor in FLOORS.values())
print(f"{cleared} of {len(PROFILES) * len(FLOORS)} checks cleared; IIM Calcutta adds that actual cut-offs may be much higher")

Asha: VARC 97.2, DILR 96.4, QA 94.1, overall 98.5
  IIM Ahmedabad: clears every floor
  IIM Bangalore (previous cycle): clears every floor
  IIM Calcutta: clears every floor
Bharat: VARC 99.1, DILR 92.0, QA 82.3, overall 97.6
  IIM Ahmedabad: below the floor: QA 82.3 < 85
  IIM Bangalore (previous cycle): clears every floor
  IIM Calcutta: clears every floor
Chitra: VARC 78.5, DILR 99.2, QA 99.4, overall 96.8
  IIM Ahmedabad: below the floor: VARC 78.5 < 85
  IIM Bangalore (previous cycle): below the floor: VARC 78.5 < 80
  IIM Calcutta: below the floor: VARC 78.5 < 80
Dev: VARC 90.0, DILR 88.0, QA 91.0, overall 94.6
  IIM Ahmedabad: below the floor: overall 94.6 < 95
  IIM Bangalore (previous cycle): clears every floor
  IIM Calcutta: clears every floor
7 of 12 checks cleared; IIM Calcutta adds that actual cut-offs may be much higher

Bharat’s QA percentile of 82.3 clears two floors and misses IIM Ahmedabad’s 85, and Dev’s 94.6 overall misses only IIM Ahmedabad’s 95: the same profile is out at one institute and still ranked at another.

What the floors cannot tell you, the formulas cannot either, because each needs figures of the pool that no candidate sees:

• IIM Ahmedabad: the average of the 50 highest ratings, and the average of the top 1 % of raw composites in each discipline.
• IIM Bangalore: the mean and the standard deviation of every part across its first shortlist.
• IIM Calcutta: the maximum possible scaled score, and the composite cut-off it sets in each category.
• All three: how many candidates each decides to call.

A composite calculator shows how the parts of a published formula trade off against each other. It cannot say who will be shortlisted, and this track never predicts it.

Practise: ten questions on the three formulas

Ten questions on the stages, the three formulas and the floors. They start where your answers have been weakest so far, and a question you miss comes back once before the end:

Interview questions

Warm-up: what is a composite score?

It is the single number an institute computes to rank applicants, from several inputs that each carry a published weight. For the 2027–29 batch, IIM Ahmedabad’s adds 35 % of an application rating, built from the 10th, 12th and degree marks, work experience and gender diversity, to 65 % of the CAT percentile. IIM Bangalore’s adds points for CAT, the three marks, work experience and diversity, out of 100, with CAT and the marks standardised against its first shortlist. IIM Calcutta’s adds a scaled CAT score to points for the school marks and diversity, out of 85. A composite earns or misses the interview call; after the interviews each institute computes a new one, with the interview in it, to make its offers.

Two questions take the formulas further:

Key takeaways

• Each IIM sets its own stages and weights after one common test, and IIM Ahmedabad shortlists independently of the other IIMs, so lists for the same candidate can differ.
• IIM Ahmedabad rates the 10th, 12th and degree marks, work experience and gender out of 38, divides the rating by the top-50 average, and weighs it 35 to 65 against the CAT percentile.
• IIM Bangalore standardises its parts against its first shortlist, so a candidate at the mean earns half of each weight, and DILR carries 21 of CAT’s 55 points.
• IIM Calcutta’s shortlisting composite adds 56 points from the scaled CAT score to up to 29 for school marks and diversity, without the degree or work experience.
• Published minimum percentiles are floors, not cut-offs, and every composite needs figures of the pool that no candidate can see, so no calculator can predict a shortlist.