CAT and MBA Entrance Preparation Module 1 – Start here: CAT and the other MBA entrance tests
Start with a diagnostic and an error log
Sit a 15-question CAT diagnostic, score it by section and question type, sort lost questions into five causes, and keep an error log that sets next lessons.
What you will learn
- Calculate a diagnostic's score by section and by question type
- Classify every error by its cause, concept, method, reading, calculation or time
- Maintain an error log that chooses your next lessons and re-attempts
Before you start
What you’ll learn · Before you start
What you will learn
- Calculate a diagnostic's score by section and by question type
- Classify every error by its cause, concept, method, reading, calculation or time
- Maintain an error log that chooses your next lessons and re-attempts
Before you start
On this page
Fifteen questions cannot predict a CAT score, and a diagnostic does not try to. Sat before you study, it shows where your marks go: in which section, on which kind of question, and for which of five reasons each lost question was lost.
This lesson gives you an untimed diagnostic of five questions for each section, every answer explained the moment you give it, a program that scores it as CAT 2026 does, five causes to sort every lost question into, and an error log: A record of every question answered wrongly or skipped, with its cause, the lesson that fixes it and a later day to re-attempt it. that turns them into next week’s lessons and re-attempts.
A diagnostic shows where to start, not what you will score
A diagnostic test: A short test taken before preparation starts and scored by section, question type and cause of error, to show where to begin; it measures a starting point, not a final score. is measured by what it tells you to do next. Fifteen questions, one to three of each kind, are enough to rank the three sections and to flag kinds of question, and too few to measure any one of them: a single counting question you miss is a lead to check, not a verdict on counting. Three rules keep the diagnostic honest:
| Rule | Why |
|---|---|
| No clock, but note your minutes for each section | Speed comes later; the minutes show which section is slow |
| Leave a question you cannot work out rather than guess | A blind guess hides the cause the log needs |
| Answer each question once, and note what the checkpoint says | Your first answer is the one you score |
Use rough paper, and a basic calculator if you like: CAT gives an on-screen calculator throughout the test. Each checkpoint below tells you at once whether you were right and why, so note the result, right, wrong or skipped, before you read on.
Not affiliated
The marking used to score this diagnostic is the one the IIMs publish for CAT 2026. MySmartCoPilot has no connection with the IIMs, and every question here is this track’s own.
Checkpoint
Check: What a diagnostic measures
Sit the diagnostic: five questions for each section
The questions are original and follow CAT’s three sections in order, with MCQs and typed answers as CAT 2026 has them. Typed answers are the questions with a number box.
VARC: a passage, a jumble and an odd sentence
Read the passage, then answer questions 1 to 3:
Passage for questions 1 to 3
For most of the last century, a broken radio in an Indian town went to a repair stall before it went anywhere else. The stall was rarely a business plan. It was a store of knowledge, kept in one person’s hands and passed on by apprenticeship. When manufacturers began to seal their products, glue in their batteries and stop selling spare parts, they did not only make repair harder. They cut the line along which that knowledge travelled. A skill that is not used is not merely paused; within a generation it is gone, and buying it back costs more than keeping it would have.
Defenders of sealed design point out, fairly, that it makes devices thinner, tougher and cheaper to build. But the price of a device is not the same as its cost. When a phone that could have been fixed for a few hundred rupees is replaced instead, the buyer pays for a new phone and the town loses one more reason to keep a repair stall open. The case for a right to repair is therefore less about any single device than about whether a town can still look after the things it owns.
Checkpoint
Diagnostic, VARC: questions 1 to 3
Question 4 gives four sentences out of order:
- A new valve that cost ₹300 ended the problem that night.
- For months the town’s water tank lost a third of its water every night, and nobody could find a leak.
- A retired plumber noticed that the loss stopped on the nights when the pump was switched off at the mains.
- That pointed to the pump’s return valve, which had worn through and was letting the tank drain back into the well.
Question 5 gives five sentences, four of which make a paragraph:
- When the library began lending tools as well as books, it expected drills and ladders to be borrowed most.
- Instead, the most borrowed item in the first year was a sewing machine.
- Its librarians explained that a drill is needed once a year, while a sewing machine is needed every week by the same few families.
- The library’s reading room is open until nine at night.
- The finding changed what it bought next: fewer one-off tools and more of the ones people need often.
Checkpoint
Diagnostic, VARC: questions 4 and 5
DILR: a table and a seating
Five friends sat the same practice paper. The table gives each one’s marks in the three sections, under CAT’s marking. Answer questions 6 to 8 from it:
| Friend | VARC | DILR | QA |
|---|---|---|---|
| Asha | 45 | 25 | 38 |
| Bilal | 33 | 40 | 30 |
| Chitra | 51 | 31 | 22 |
| Dev | 30 | 34 | 47 |
| Esha | 38 | 29 | 41 |
Checkpoint
Diagnostic, DILR: questions 6 to 8
The same five then sit in a row of five chairs, numbered 1 to 5 from the left, to talk about the paper. Answer questions 9 and 10 from these clues:
- (i) Dev sits in chair 1.
- (ii) Asha sits immediately to the right of Esha.
- (iii) Chitra sits at neither end of the row.
- (iv) Bilal sits next to Dev.
Checkpoint
Diagnostic, DILR: questions 9 and 10
QA: five questions from five topics
Questions 11 to 15 cover arithmetic, algebra, the number system, geometry and counting, one each:
Checkpoint
Diagnostic, QA: questions 11 to 13
Checkpoint
Diagnostic, QA: questions 14 and 15
Every DILR and QA answer above is worked out by the program below from the question itself, by brute force where a count is asked for, so the key does not rest on anyone’s arithmetic. The verbal answers are argued in each question’s explanation instead, since no program can prove an inference.
"""Proves the answer key of the diagnostic: every DILR and QA answer is worked out here from the question itself,
by brute force where a count is asked for. The verbal answers are argued in the lesson, not computed.
"""
from fractions import Fraction
from itertools import permutations
from math import isqrt
# Questions 6 to 8: the five friends' section marks.
MARKS = {
"Asha": (45, 25, 38),
"Bilal": (33, 40, 30),
"Chitra": (51, 31, 22),
"Dev": (30, 34, 47),
"Esha": (38, 29, 41),
}
totals = {name: sum(m) for name, m in MARKS.items()}
best = max(totals.values())
print("Q6 highest total:", [n for n, t in totals.items() if t == best], best)
dilr = [m[1] for m in MARKS.values()]
print("Q7 highest DILR above the lowest:", Fraction(max(dilr) - min(dilr), min(dilr)) * 100, "%")
qualify = [n for n, m in MARKS.items() if min(m) >= 30 and sum(m) >= 100]
print("Q8 every section 30 or more and a total of 100 or more:", len(qualify), qualify)
# Questions 9 and 10: the same five in chairs 1 to 5, numbered from the left.
names = ["Asha", "Bilal", "Chitra", "Dev", "Esha"]
clues = {
"Dev in chair 1": lambda s: s.index("Dev") == 0,
"Asha just right of Esha": lambda s: s.index("Asha") == s.index("Esha") + 1,
"Chitra at neither end": lambda s: s.index("Chitra") not in (0, 4),
"Bilal next to Dev": lambda s: abs(s.index("Bilal") - s.index("Dev")) == 1,
}
every = [s for s in permutations(names) if all(c(s) for c in clues.values())]
print("Q9 seatings with all four clues:", len(every), "->", every[0], "chair 4:", every[0][3])
without = [s for s in permutations(names) if all(c(s) for k, c in clues.items() if k != "Bilal next to Dev")]
print("Q10 seatings without the Bilal clue:", len(without))
# Questions 11 to 15.
price = Fraction(1500) / (Fraction(5, 4) * Fraction(4, 5))
print("Q11 price before the rise and the discount: Rs", price)
ages = [(r, b) for b in range(1, 60) for r in range(1, 120) if r == 3 * b and r + 6 == 2 * (b + 6)]
print("Q12 Neha and her brother now:", ages)
print("Q13 divisors of 720:", sum(1 for d in range(1, 721) if 720 % d == 0))
a, b = 9, 12
c = isqrt(a * a + b * b)
assert c * c == a * a + b * b
print("Q14 inradius of the 9-12-15 triangle:", Fraction(a * b, 2) / Fraction(a + b + c, 2), "cm")
apart = [p for p in permutations("PQRS") if abs(p.index("P") - p.index("Q")) != 1]
print("Q15 rows of four books with P and Q apart:", len(apart))
# Question 4: how many orders four sentences have, before any link is used.
print("Q4 orders of four sentences:", len(list(permutations("1234"))))
# Two questions of the lesson's practice.
four = [s for s in permutations(["Pia", "Ravi", "Sam", "Tara"]) if s.index("Pia") not in (0, 3) and s.index("Ravi") == s.index("Pia") + 1]
print("Practice: seatings of four with Pia inside and Ravi just right of her:", len(four))
print("Practice: two rises of 10 %:", (Fraction(11, 10) ** 2 - 1) * 100, "% above the start")Output
Q6 highest total: ['Dev'] 111
Q7 highest DILR above the lowest: 60 %
Q8 every section 30 or more and a total of 100 or more: 2 ['Bilal', 'Dev']
Q9 seatings with all four clues: 1 -> ('Dev', 'Bilal', 'Chitra', 'Esha', 'Asha') chair 4: Esha
Q10 seatings without the Bilal clue: 4
Q11 price before the rise and the discount: Rs 1500
Q12 Neha and her brother now: [(18, 6)]
Q13 divisors of 720: 30
Q14 inradius of the 9-12-15 triangle: 3 cm
Q15 rows of four books with P and Q apart: 12
Q4 orders of four sentences: 24
Practice: seatings of four with Pia inside and Ravi just right of her: 4
Practice: two rises of 10 %: 21 % above the start
Recorded with Python 3.14.8 on macOS 26 arm64. To run it yourself: mise exec python@3.14.8 -- python3 check_answers.py
Runs on this device, in your browser. The first run downloads Python (about 13.5 MB), which is kept for the next runs.
Your run, in this browser
Score it by section, by format and by type
Score the diagnostic as CAT 2026 marks its questions: 3 marks for a right answer, minus 1 for a wrong MCQ, nothing for
a wrong typed answer and nothing for a blank. The program does it from your fifteen first results. Here are Arjun’s,
an invented candidate; press Edit, type yours into MY_RESULTS, R for right, W for wrong and S for skipped, and run it
again:
"""Score the diagnostic as CAT 2026 marks it: +3 for a right answer, -1 for a wrong MCQ, 0 for a wrong typed answer
and 0 for a blank. Type your first result for each question into MY_RESULTS (R right, W wrong, S skipped), five
letters a section in the order VARC, DILR, QA, and run it again.
"""
MY_RESULTS = "RRWWR RWRRS WRSRW" # an invented candidate's first answers
# (question, section, question type, format, the lesson of this track that teaches it)
QUESTIONS = [
(1, "VARC", "reading comprehension", "MCQ", "reading-comprehension-for-cat"),
(2, "VARC", "reading comprehension", "MCQ", "reading-comprehension-for-cat"),
(3, "VARC", "reading comprehension", "MCQ", "rc-question-types-and-elimination"),
(4, "VARC", "para jumble", "typed", "para-jumbles-without-options"),
(5, "VARC", "odd sentence out", "typed", "para-summary-and-odd-sentence-out-for-cat"),
(6, "DILR", "table", "MCQ", "tables-and-charts-in-di-sets"),
(7, "DILR", "table", "typed", "tables-and-charts-in-di-sets"),
(8, "DILR", "table", "MCQ", "tables-and-charts-in-di-sets"),
(9, "DILR", "arrangement", "MCQ", "arrangement-and-scheduling-sets"),
(10, "DILR", "arrangement", "typed", "arrangement-and-scheduling-sets"),
(11, "QA", "arithmetic", "MCQ", "percentages-and-ratios-for-cat"),
(12, "QA", "algebra", "typed", "linear-equations-and-whole-number-solutions"),
(13, "QA", "number system", "MCQ", "number-system-for-cat"),
(14, "QA", "geometry", "typed", "triangles-for-cat"),
(15, "QA", "counting", "MCQ", "permutations-and-combinations-for-cat"),
]
MARKS = {("R", "MCQ"): 3, ("W", "MCQ"): -1, ("R", "typed"): 3, ("W", "typed"): 0}
results = MY_RESULTS.replace(" ", "").upper()
assert len(results) == len(QUESTIONS) and set(results) <= set("RWS"), "type 15 letters, each R, W or S"
rows = [(q, sec, kind, fmt, lesson, r, MARKS.get((r, fmt), 0)) for (q, sec, kind, fmt, lesson), r in zip(QUESTIONS, results)]
def tally(selected):
right = sum(1 for row in selected if row[5] == "R")
wrong = sum(1 for row in selected if row[5] == "W")
return right, wrong, len(selected) - right - wrong, sum(row[6] for row in selected)
print("The diagnostic, scored with CAT 2026 marking")
print("Section Right Wrong Skipped Marks of")
for name in ("VARC", "DILR", "QA", "Total"):
part = [row for row in rows if name in (row[1], "Total")]
right, wrong, skipped, marks = tally(part)
print(f"{name:<8}{right:>6}{wrong:>7}{skipped:>9}{marks:>7}{3 * len(part):>4}")
print()
for fmt in ("MCQ", "typed"):
right, wrong, skipped, marks = tally([row for row in rows if row[3] == fmt])
lost = f"the wrong ones cost {wrong}" if fmt == "MCQ" else "the wrong ones cost nothing"
label = {"MCQ": "MCQs:", "typed": "Typed:"}[fmt]
print(f"{label:<7}{right} right, {wrong} wrong, {skipped} skipped: {marks} marks; {lost}")
print()
print("By question type, right of asked:")
for kind in dict.fromkeys(row[2] for row in rows):
part = [row for row in rows if row[2] == kind]
print(f" {part[0][1]:<5}{kind:<22}{sum(1 for row in part if row[5] == 'R')} of {len(part)}")
print()
share = {}
for s in ("VARC", "DILR", "QA"):
part = [row for row in rows if row[1] == s]
share[s] = round(100 * sum(1 for row in part if row[5] == "R") / len(part))
print("For the planner, per cent right: " + ", ".join(f"{s} {p} %" for s, p in share.items()))
print()
print("To log, with the lesson that teaches each:")
for q, sec, kind, fmt, lesson, r, _ in rows:
if r != "R":
print(f" Q{q:<3}{sec:<5}{'wrong' if r == 'W' else 'skipped':<8}{lesson}")Output
The diagnostic, scored with CAT 2026 marking Section Right Wrong Skipped Marks of VARC 3 2 0 8 15 DILR 3 1 1 9 15 QA 2 2 1 4 15 Total 8 5 2 21 45 MCQs: 5 right, 3 wrong, 1 skipped: 12 marks; the wrong ones cost 3 Typed: 3 right, 2 wrong, 1 skipped: 9 marks; the wrong ones cost nothing By question type, right of asked: VARC reading comprehension 2 of 3 VARC para jumble 0 of 1 VARC odd sentence out 1 of 1 DILR table 2 of 3 DILR arrangement 1 of 2 QA arithmetic 0 of 1 QA algebra 1 of 1 QA number system 0 of 1 QA geometry 1 of 1 QA counting 0 of 1 For the planner, per cent right: VARC 60 %, DILR 60 %, QA 40 % To log, with the lesson that teaches each: Q3 VARC wrong rc-question-types-and-elimination Q4 VARC wrong para-jumbles-without-options Q7 DILR wrong tables-and-charts-in-di-sets Q10 DILR skipped arrangement-and-scheduling-sets Q11 QA wrong percentages-and-ratios-for-cat Q13 QA skipped number-system-for-cat Q15 QA wrong permutations-and-combinations-for-cat
Recorded with Python 3.14.8 on macOS 26 arm64. To run it yourself: mise exec python@3.14.8 -- python3 diagnostic.py
Runs on this device, in your browser. The first run downloads Python (about 13.5 MB), which is kept for the next runs.
Your run, in this browser
Arjun scored 21 of 45: 8 in VARC, 9 in DILR and only 4 in QA. The format lines show what the marking did: his three wrong MCQs cost him 3 marks, while his two wrong typed answers cost nothing.
Reading the score as a forecast
“21 of 45 means I will score about 47 % in CAT” turns a starting point into a prediction. Fifteen untimed questions, one to three of each kind, say nothing reliable about a 120-minute test with an unknown number of questions. Use them for what they can tell you: which section first, which kinds of question to test next and which causes to log.
The question types are where the leads are, and they come with a warning: one question of a type is a lead to test with a set of that type, not a measurement. His per cent right by section, 60, 60 and 40, is what the planner of the previous lesson takes as its diagnostic.
Key idea
Score a diagnostic three ways, by section, by format and by question type, and use it to choose where to start; the last list it prints, the questions to log, matters more than its total.
Checkpoint
Check: Scoring by section, format and type
Five causes, five different fixes
Two candidates can lose the same question for different reasons, and the reason decides the fix. Give every lost question one cause: the first one whose fix would have saved it.
| Cause | What happened | The fix, and the sign it worked |
|---|---|---|
| Concept | An idea was missing; no amount of time would have saved it | The lesson, then three fresh questions; you can explain the idea and solve a new one |
| Method | You knew the idea but chose a slow or wrong way | The lesson’s method, then a set of that kind; two in a row come out right |
| Reading | You answered another question: a missed “not”, a unit, a condition | Mark what is asked before solving; a paper with no reading errors |
| Calculation | Right method, wrong number | Estimate first, recheck the last step; fewer slips in each paper |
| Time | The clock ran out, or one question took far too long | A rule for when to leave and come back; no question takes over twice its share |
A wrong guess is not a sixth cause. Log it under the cause that made you guess, usually concept or time, and judge the guess itself by the 25 % break-even of a CAT MCQ from the negative-marking lesson. Follow one lost question, number 10 of the diagnostic, from the moment it is lost to the day it is solved again:
The drawing shows the whole scene; its 6 steps are listed below.
Every step as text
- A skip is logged too. Arjun skipped question 10 of the diagnostic, the count of seatings without clue (iv). A skipped question goes into the log as surely as a wrong one, because the paper could have scored it.
- One cause, the first fix. He asks which single fix would have saved the question. He understood the clues, so it is not concept; he began listing seatings and lost count, so the cause is method.
- A lesson, not a section. The fix is taught in a lesson, arrangement and scheduling sets, whose method fixes the position of the pair first and counts what is left. The row names that lesson, not just DILR.
- A day to try again. The re-attempt goes three days later, on day 4 of his plan. That is far enough for the answer to have faded and near enough to test the method while the lesson is fresh.
- A test, not a reading. On day 4 he solves it from the question alone, with no notes and no solution open. With Dev in chair 1, the Esha and Asha pair in chairs 2 and 3, 3 and 4, or 4 and 5 leaves 1, 1 and 2 seatings, 4 in all.
- Right this time. His answer is right, so the row leaves the list of re-attempts; the log keeps it as a record of what was fixed. A wrong answer would have sent him back to the lesson and into the log again, with a new day.
The re-attempt is a test, not a reading: you solve the question again days later, with no solution in sight, which is retrieval practice: Practice by recalling or solving from memory, as in a test, rather than by rereading the material or a solution.. Roediger and Karpicke’s students remembered prose passages better a week later after recall tests than after rereading, and practice testing is one of the two techniques Dunlosky and colleagues rated most useful.
Checkpoint
Check: The five causes of error
The error log chooses next week’s lessons
One row per lost question, with its day, section, cause, the lesson that fixes it and the minutes it took, is all a log needs. The program below summarises four weeks of an invented candidate’s log, 40 rows from the diagnostic and the weeks after it, and says what to do next:
log/error_log.py
"""What an error log says after four weeks: errors by section and cause, the minutes each cause cost, the three
lessons to revisit first, the re-attempts due this week and the split of next week's hours for the planner.
Each lost question is re-attempted from the question alone 3 days after it was logged (MySmartCoPilot's rule); a
re-attempt that goes wrong is logged again as a new row. Change TODAY, or add your own rows to log_data.py.
"""
from log_data import LOG
TODAY = 29 # the day of the plan, counted from the diagnostic on day 1
GAP = 3
SECTIONS = ["VARC", "DILR", "QA"]
CAUSES = ["concept", "method", "reading", "calculation", "time"]
print(f"{len(LOG)} logged questions, {sum(r[4] for r in LOG)} minutes spent on them")
print()
print(f"{'Cause':<12}" + "".join(f"{s:>6}" for s in SECTIONS) + f"{'Total':>7}{'Minutes':>9}")
for cause in CAUSES:
rows = [r for r in LOG if r[2] == cause]
counts = "".join(f"{sum(1 for r in rows if r[1] == s):>6}" for s in SECTIONS)
print(f"{cause:<12}{counts}{len(rows):>7}{sum(r[4] for r in rows):>9}")
print(f"{'Total':<12}" + "".join(f"{sum(1 for r in LOG if r[1] == s):>6}" for s in SECTIONS) + f"{len(LOG):>7}{sum(r[4] for r in LOG):>9}")
print()
print("Lessons to revisit first (most rows, then most minutes):")
lessons = {}
for _, _, _, lesson, minutes in LOG:
count, total = lessons.get(lesson, (0, 0))
lessons[lesson] = (count + 1, total + minutes)
for lesson, (count, total) in sorted(lessons.items(), key=lambda kv: (-kv[1][0], -kv[1][1]))[:3]:
print(f" {lesson:<38}{count} rows, {total} min")
print()
due = [r for r in LOG if TODAY <= r[0] + GAP < TODAY + 7]
print(f"Re-attempts due on days {TODAY} to {TODAY + 6}, from the question alone:")
for day, section, cause, lesson, _ in due:
print(f" day {day + GAP}: a {section} question lost to {cause} on day {day}")
print()
counts = {s: sum(1 for r in LOG if r[1] == s) for s in SECTIONS}
units = 22 # the 11 section hours of a 12-hour week, in half-hours
share = {s: c / len(LOG) for s, c in counts.items()}
for s in SECTIONS: # the planner's floor: every section keeps at least a fifth of the section hours
if share[s] < 0.2:
rest = [t for t in SECTIONS if t != s]
spare = sum(share[t] for t in rest)
for t in rest:
share[t] *= 0.8 / spare
share[s] = 0.2
exact = {s: units * share[s] for s in SECTIONS}
got = {s: int(x) for s, x in exact.items()}
for s in sorted(SECTIONS, key=lambda s: exact[s] - got[s], reverse=True)[: units - sum(got.values())]:
got[s] += 1
print("For the planner, the log's share of errors: " + ", ".join(f"{s} {100 * c // len(LOG)} %" for s, c in counts.items()))
print(" which splits 11 section hours as " + ", ".join(f"{s} {got[s] / 2:g}" for s in SECTIONS))Output
40 logged questions, 198 minutes spent on them Cause VARC DILR QA Total Minutes concept 2 0 5 7 25 method 2 7 1 10 52 reading 4 3 2 9 22 calculation 0 2 4 6 18 time 2 6 0 8 81 Total 10 18 12 40 198 Lessons to revisit first (most rows, then most minutes): how-dilr-sets-work 7 rows, 30 min choosing-dilr-sets 6 rows, 64 min percentages-and-ratios-for-cat 6 rows, 14 min Re-attempts due on days 29 to 35, from the question alone: day 29: a DILR question lost to time on day 26 day 29: a DILR question lost to reading on day 26 day 30: a VARC question lost to reading on day 27 day 30: a QA question lost to calculation on day 27 day 31: a DILR question lost to time on day 28 For the planner, the log's share of errors: VARC 25 %, DILR 45 %, QA 30 % which splits 11 section hours as VARC 2.5, DILR 5, QA 3.5
Recorded with Python 3.14.8 on macOS 26 arm64. To run it yourself: mise exec python@3.14.8 -- python3 error_log.py
log/log_data.py
"""An invented candidate's error log for the first four weeks of a plan: the diagnostic on day 1, then the questions of
each week's lessons. One row per question answered wrongly or skipped: (day, section, cause, lesson, minutes spent).
The lesson is the one of this track that fixes the cause; the minutes are the time the question took before it was lost.
"""
LOG = [
(1, "VARC", "concept", "rc-question-types-and-elimination", 3),
(1, "VARC", "method", "para-jumbles-without-options", 4),
(1, "DILR", "calculation", "tables-and-charts-in-di-sets", 3),
(1, "DILR", "method", "arrangement-and-scheduling-sets", 5),
(1, "QA", "concept", "percentages-and-ratios-for-cat", 2),
(1, "QA", "concept", "number-system-for-cat", 3),
(1, "QA", "concept", "permutations-and-combinations-for-cat", 4),
(3, "DILR", "method", "how-dilr-sets-work", 6),
(4, "VARC", "reading", "reading-comprehension-for-cat", 3),
(4, "QA", "calculation", "percentages-and-ratios-for-cat", 3),
(5, "DILR", "time", "choosing-dilr-sets", 12),
(5, "DILR", "reading", "how-dilr-sets-work", 3),
(6, "VARC", "concept", "rc-question-types-and-elimination", 4),
(7, "DILR", "method", "arrangement-and-scheduling-sets", 5),
(8, "QA", "reading", "percentages-and-ratios-for-cat", 2),
(9, "VARC", "reading", "reading-comprehension-for-cat", 2),
(9, "QA", "concept", "number-system-for-cat", 4),
(10, "DILR", "method", "how-dilr-sets-work", 6),
(11, "VARC", "time", "sectional-time-plans", 9),
(12, "DILR", "time", "choosing-dilr-sets", 11),
(12, "DILR", "calculation", "tables-and-charts-in-di-sets", 3),
(13, "VARC", "method", "para-jumbles-without-options", 5),
(14, "DILR", "method", "how-dilr-sets-work", 5),
(15, "QA", "calculation", "percentages-and-ratios-for-cat", 2),
(16, "DILR", "time", "choosing-dilr-sets", 10),
(17, "DILR", "reading", "how-dilr-sets-work", 2),
(18, "VARC", "reading", "reading-comprehension-for-cat", 2),
(18, "QA", "method", "linear-equations-and-whole-number-solutions", 5),
(19, "DILR", "method", "arrangement-and-scheduling-sets", 6),
(20, "QA", "concept", "permutations-and-combinations-for-cat", 5),
(21, "DILR", "time", "choosing-dilr-sets", 12),
(22, "QA", "calculation", "triangles-for-cat", 4),
(23, "VARC", "time", "sectional-time-plans", 8),
(24, "DILR", "method", "how-dilr-sets-work", 5),
(25, "QA", "reading", "percentages-and-ratios-for-cat", 2),
(26, "DILR", "time", "choosing-dilr-sets", 9),
(26, "DILR", "reading", "how-dilr-sets-work", 3),
(27, "VARC", "reading", "reading-comprehension-for-cat", 3),
(27, "QA", "calculation", "percentages-and-ratios-for-cat", 3),
(28, "DILR", "time", "choosing-dilr-sets", 10),
]Runs on this device, in your browser. The first run downloads Python (about 13.5 MB), which is kept for the next runs.
Your run, in this browser
Three things stand out. Time cost 8 rows but 81 of the 198 minutes, all of it in VARC and DILR sets where one question or one set was held too long. Reading cost 9 rows in only 22 minutes, and its fix is a habit rather than a lesson. The first lesson to revisit is how DILR sets work, with 7 rows; choosing DILR sets and percentages tie at 6, and the minutes, 64 against 14, put choosing sets first.
The last two lines close the loop with the previous lesson. The log’s share of errors, 25 % VARC, 45 % DILR and 30 % QA, takes the place of the diagnostic in the planner, and splits next week’s 11 section hours as 2.5, 5 and 3.5. The plan now follows the errors you actually make.
Study Planner & Revision Timetable Put next week's lessons and the re-attempt days from your log into a day-by-day timetable.Checkpoint
Check: The error log and its re-attempts
Practise: diagnose and log
Ten questions on what a diagnostic measures, two new questions of the diagnostic’s kinds, the scoring, the five causes and the log. The part where your first answers were weakest is asked first, and a miss is asked again later:
Practice that adapts to you
Practise this lesson
10 questions about 5 ideas of this lesson. Each answer is explained at once, and what you get right is kept in this browser.
Interview questions
Warm-up: what are the five causes of error that an error log records?
Concept, method, reading, calculation and time. A concept error is a question you could not have solved with unlimited time, because an idea was missing. A method error means you knew the idea but chose a slow or wrong way, such as listing cases instead of counting them. A reading error answers a different question from the one asked: a missed “not”, a unit or a condition. A calculation error has the right method and a wrong number. A time error is a question lost because the clock ran out or one question took far too long. Each cause has a different fix, so a log records the first cause whose fix would have saved the question, never only that it was wrong.
Model answers open with any pass: ₹79 for 1 week ($2.99), one payment that never renews.
-
Fresher to mid
After four weeks your error log has 18 DILR rows: 6 lost to time, costing 64 minutes, and 7 to method. A friend says, "Just do more DILR sets." What would you do instead, and how would you know in two weeks whether it worked?
-
Fresher to mid
A friend logs only wrong answers, never skipped questions. In a practice paper she skipped 14 questions and got 9 wrong. What does her log miss, and what would you add to it?
Flashcards · 7 cards
Recall it, then check
Say the answer to yourself before you look: recalling it is what makes it stick.
-
What can a 15-question diagnostic tell you, and what can it not?
It ranks the three sections, flags kinds of question to test next and lists the questions to log. It cannot forecast a CAT score: 21 of 45 untimed does not mean 47 % in CAT.
-
What three rules keep a diagnostic honest?
No clock, but note your minutes for each section; leave what you cannot work out rather than guess, since a blind guess hides the cause; and score only your first answer to each question.
-
What are the five causes an error log sorts lost questions into?
Concept (an idea was missing), method (a slow or wrong way), reading (you answered another question), calculation (right method, wrong number) and time (the clock ran out, or one question took far too long).
-
A lost question has more than one cause. Which do you log?
The first cause whose fix would have saved it, because each cause has its own fix. A wrong guess is not a sixth cause: log the cause that made you guess, usually concept or time.
-
What does one row of an error log hold?
Five entries: the day, the section, the cause, the lesson to revisit and the minutes spent. That is enough to rank lessons and to schedule each question for a second attempt.
-
When and how is a lost question solved again?
About three days later, from the question alone, with no solution in sight. That makes the re-attempt retrieval practice rather than a rereading.
-
How does the error log change next week's plan?
Its share of errors by section replaces the diagnostic in the planner: 25 % VARC, 45 % DILR and 30 % QA split 11 section hours as 2.5, 5 and 3.5.
Key takeaways
- A diagnostic shows where to start, not what you will score: sit it untimed, leave what you cannot work out, and score your first answers.
- Score it by section, by format and by question type under CAT’s marking; one question of a type is a lead to test, not a measurement.
- Give every lost question one cause, concept, method, reading, calculation or time, the first whose fix would have saved it, because each cause has a different fix.
- Log each lost question with its day, section, cause, lesson and minutes, and solve it again from the question alone three days later.
- The log ranks the lessons to revisit and replaces the diagnostic in the planner: 40 rows split 11 hours as 2.5, 5 and 3.5.
Terms in this lesson
- diagnostic test
- A short test taken before preparation starts and scored by section, question type and cause of error, to show where to begin; it measures a starting point, not a final score.
- error log
- A record of every question answered wrongly or skipped, with its cause, the lesson that fixes it and a later day to re-attempt it.
- retrieval practice
- Practice by recalling or solving from memory, as in a test, rather than by rereading the material or a solution.
References
- CAT 2026 FAQs (Indian Institutes of Management (CAT 2026, conducted by IIM Indore))
- Test-enhanced learning: taking memory tests improves long-term retention (Roediger and Karpicke) (Psychological Science (PubMed record))
- Improving students' learning with effective learning techniques (Dunlosky, Rawson, Marsh, Nathan and Willingham) (Psychological Science in the Public Interest (PubMed record))
Related tools
Report a problem with this lesson
Kept only in this browser. Your Learn progress