Why This Topic Matters
What CAT actually tests
In CAT, "Logical Reasoning" does not mean syllogisms or blood relations — those have not appeared since the IIMs took over the paper. It means multi-question reasoning sets: a block of conditions, then 4–6 questions you can only answer once you've cracked the underlying arrangement. On average about 4 LR questions per slot (plus another ~3 per slot in arrangement-heavy sets), typically 2–4 sets per paper. One well-chosen set is worth 8–12 marks; one badly-chosen set can swallow 15 minutes for nothing. Selection and representation matter as much as raw logic.
What the last five years of LR sets looked like
| Set flavour | 2021 | 2022 | 2023 | 2024 | 2025 | Avg/slot |
|---|---|---|---|---|---|---|
| Puzzles (allocation, grids, mixed) | – | – | – | 5.0 | – | 1.0 |
| Scheduling | 2.0 | – | – | 1.7 | – | 0.7 |
| Games & tournaments | 1.3 | – | – | – | 1.3 | 0.5 |
| Relations / family-tree style | 2.0 | – | – | – | – | 0.4 |
| Distribution of objects | 2.0 | – | – | – | – | 0.4 |
| Voting | – | – | 1.7 | – | – | 0.3 |
| Routes | – | – | 1.7 | – | – | 0.3 |
| Other classic LR | 3.3 | – | – | – | – | 0.7 |
Read the rotation, not the rows. No LR flavour repeats reliably — 2023 brought a voting set and a routes set, 2024 a 15-question puzzles wave (coach–player allocation, a constrained number grid, an election scenario), 2025 a tournament set (counting "taps" across rounds). What is stable: every set rewards the same five moves — pick a representation, anchor on the most restrictive clue, branch narrow, prune on contradiction, re-read the question stem. Train the moves, not last year's costume. And note what's absent: truth-teller/liar sets have not appeared even once in 2021–2025.
The recurring set archetypes
Recognising the type in the first read tells you which diagram to draw and where the marks hide.
| Archetype | What it looks like | Crack it with |
|---|---|---|
| Arrangements | Order people/objects in a line, circle, or grid | Slot diagram; lock fixed positions first |
| Distribution / Allocation | Hand out items or people under min–max rules | Anchor the total; squeeze with min/max and "all distinct" |
| Matching | Map entities to attributes (one-to-one) | Tick/cross matrix |
| Scheduling | Tasks vs. time/days with precedence & capacity | Time-resource grid |
| Games & Tournaments | Matches, scores, rankings | Match counts + the scoring rule; tie-breakers carry marks |
| Routes & Networks | Paths, flows, connections | Enumerate every complete route; never trust a greedy guess |
Step 1 — pick the right representation
Most sets are won or lost at the diagram. Match structure to template before you start deducing:
- Matrix — entities × attributes, one-to-one (who ordered what)
- Slot diagram — a linear order (positions 1…n)
- Circular template — seating, cycles (mind clockwise vs. anticlockwise)
- Table — entities × several attributes at once
- Case tree — when a rule forces an either/or split
Step 2 — deduce, branch, and prune
- Start from the most restrictive clue, not the first one. The clue that fixes a value or kills the most options is your anchor.
- When forced to split, split on the constraint that creates the fewest branches. Carry all the conditions down each branch.
- One contradiction kills a branch — prune it immediately and don't look back.
- Conditionals fire only when their trigger is met. "If A then B" says nothing until A is known. Its contrapositive is free information: if A then B ≡ if not B then not A.
A worked matching set
Four friends — Asha, Bittu, Chen, Dia — each ordered a different drink: coffee, tea, juice, soda. (1) Asha ordered neither coffee nor tea. (2) Bittu ordered neither juice nor soda. (3) Chen ordered tea. (4) Dia did not order juice.
Deduce in matrix form:
- (3) Chen = tea. Tea is now gone for everyone else.
- (2) Bittu ≠ juice, ≠ soda, and ≠ tea (taken) ⇒ Bittu = coffee.
- (4) Dia ≠ juice; coffee & tea are taken ⇒ Dia = soda.
- Asha takes what's left ⇒ Asha = juice (consistent with clue 1).
Result: Asha–juice, Bittu–coffee, Chen–tea, Dia–soda. Every "who/which" question on this set now answers itself from the grid.
Timing and traps
A short set is not a safe set. Under-constrained ("vague conditional") sets force heavy enumeration; numeric, additive constraints usually mean a determinate, faster set. The 2024 number-grid set looked tiny — ten numbers, six conditions — but each condition interacted with all the others. Judge sets by constraint type, not by paragraph length.
- Greedy guessing on routes/distribution is the classic killer — enumerate completely.
- Carry every constraint into every branch. Most wrong answers come from a rule dropped during a case split.
Watch this
2IIM lays out the full CAT DILR landscape — what the sets look like and how to read them:
Checklist
- Name the archetype, then draw the matching template
- Anchor on the most restrictive clue
- In mixed sets, exhaust structure rules before number rules
- Split on the fewest-branch constraint; prune on contradiction
- Use contrapositives as free clues
- Don't marry a set — if there's no traction in ~3–4 minutes, bank the easy sub-questions and move on
Sample Questions
22 practice questions
Sign in for full access
Create a free account to access all 22 practice questions on this topic.
CAT PYQ Spotlight
Actual CAT questions on this topic
Sign in for full access
Create a free account to access all 8 CAT PYQs on this topic.
Continue Your Prep