Combinatorics Interview Questions

Quant interviews use combinatorics as a speed-and-structure test: count a set cleverly — by arrangement formulas, complements, symmetry or a small recursion — instead of enumerating it.
It shows up directly in brainteaser rounds and indirectly inside most discrete-probability questions, where the counting is the actual work.

Last updated 2 July 2026 · built from the 171 combinatorics problems in our curated bank (each with an authored worked solution) and their candidate-reported firm tags — no individual reviewer is named in our source data.

Core sub-areas
Arrangements & selections · complements / inclusion–exclusion · symmetry & double counting · recursions & Catalan · weighings & adaptive search
Typical difficulty
Evenly split easy/medium with a small hard tail — our bank runs 77 easy · 78 medium · 16 hard
Who leans on it
Trading-desk brainteaser rounds and OA math sections; in our firm tags it clusters most around Jane Street, Citadel, Optiver and SIG
Free practice here
25 free problems (15 easy · 8 medium · 2 hard), each with a full worked solution

Where combinatorics shows up: mostly in quant-trader and quant-research screens — timed OA math sections, phone-screen brainteasers and on-site quick-fire rounds — plus the occasional counting-flavored coding task on SWE tracks. See Jane Street interview questions › · Citadel interview questions ›

Practice combinatorics questions → Probability interview questions Brain-teaser interview questions

The surfaceThe combinatorics sub-areas quant interviews test

Seven families cover essentially every counting question in our bank. Each row pairs a sub-area with its recurring question shape and a representative type — a flavor drawn from our own problems, not a leaked firm question.

Combinatorics sub-areas tested in quant interviews, with the recurring question shape and a representative question type for each.
Sub-areaRecurring shapeRepresentative type
Arrangements & selectionsCount configurations under a constraintSplit a class into groups of fixed sizes, or count five-card hands using exactly two suits — the labeled-vs-unlabeled trap decides the answer.
Complements & inclusion–exclusion“At least one” / “avoid X” countsCount triples whose product is even, or sum every integer that never uses a forbidden digit — count the bad set, not the good one.
Symmetry & double countingCount one quantity two waysRecover individual weights from all pairwise sums by noting how often each element is counted.
Recursions & Catalan structuresFirst-choice recursion on nested structureCount valid parenthesizations or ordered break-downs of stacked objects — the recursion, not a formula, is the scored step.
Weighings & adaptive searchMinimum queries to isolate an itemFind the heavy coin in two balance weighings — a three-outcome information bound sets the strategy.
Digit & number-based countingCount digits or structured integers in a rangeHow many times a digit appears from 1 to N, handled position by position — the usual home of the hard tail.
Pigeonhole & extremal argumentsGuarantee-existence reasoningHow many people force a shared birth month — one line once you name the boxes.

What's confirmed vs. what varies: the sub-area list and difficulty spread come straight from our problem bank; which family a given firm emphasizes is candidate-reported and varies by desk, so treat firm tags as directional rather than a syllabus.

The patternsSignature combinatorics question patterns

Three moves generate most combinatorics interview answers. Each worked box is a 60–90 second micro-example in the interview's actual cadence, distilled from a free problem in the practice set below.

Complementary counting — count the bad set instead

Takeaway: when the condition is “at least one” or “not all”, the complement is usually a single clean count — total minus bad beats case-splitting every time.

Shape. How many 3-element subsets of {1, …, 10} have an even product? (The cadence behind Count Even-Product Triplets.)

1. Flip the condition. A product is odd only when every factor is odd — “even product” is the complement of one tiny event.

2. Count both sets. All subsets: C(10,3) = 120. All-odd subsets come only from the five odd numbers: C(5,3) = 10.

3. Subtract and sanity-check. 120 − 10 = 110. Check the scale: most triples should contain an even number, and 110/120 ≈ 92% agrees. The move: name the complement, count it once, subtract.

Labeled vs. unlabeled — the multinomial trap

Takeaway: divide by the group-orderings factor only when groups are interchangeable — saying out loud whether groups are labeled is the scored step.

Shape. Split 12 students into three groups of 4 (the cadence behind Dividing Students into Labeled Groups).

1. Fill labeled slots. With named groups A, B, C, choose in sequence: 12! / (4!·4!·4!) = 34,650 — the multinomial coefficient.

2. Ask the labeling question. If the groups are indistinguishable teams, every split was counted 3! = 6 times, once per relabeling.

3. Correct only if needed. Unlabeled answer: 34,650 / 6 = 5,775. Equal group sizes are exactly the case where relabeling matters. The move: compute the multinomial, then decide — explicitly — whether to divide by k!.

Weighings as information — bound first, then construct

Takeaway: a balance weighing has three outcomes, so k weighings distinguish at most 3k cases — state the bound first and the strategy almost writes itself.

Shape. Eight look-alike coins, one heavier; find it in two weighings (the cadence behind Eight Coins, One Heavier: Two Weighings).

1. Bound the problem. Two weighings give 3² = 9 distinguishable outcomes ≥ 8 candidates, so two can suffice — and one weighing (3 outcomes) provably cannot.

2. Split into thirds. Weigh 3 vs. 3, leaving 2 aside. Balance ⇒ the heavy coin is among the 2 left out; imbalance ⇒ it is in the heavier group of 3.

3. Finish in one weighing. From 2: weigh them against each other. From 3: weigh 1 vs. 1; balance points to the third. The move: match every weighing's three outcomes to three near-equal candidate sets.

The hard tail's key insights, in one line each: Catalan-style counts (as in Counting Proper Parenthesizations (Catalan Numbers)) come from the first-return recursion, giving C(2n,n)/(n+1); digit-occurrence counts (as in the hard Counting Digit Occurrences in a Range) are handled one digit position at a time — higher digits set full blocks, the current digit adds a partial block — never by scanning numbers.

Free practiceCombinatorics practice problems by difficulty

All 25 free combinatorics problems from our 171-problem bank, grouped by difficulty — 15 easy, 8 medium, 2 hard, mirroring the bank's easy/medium balance. Every sub-area in the table above appears below (the chip on each problem names its family), and each problem opens with a full worked solution. The free hard tail is thin — most of our 16 hard items are in the paid bank — so treat a clean timed medium as the realistic interview bar.

Easy — warm-ups and formula checks (15)

Medium — the realistic interview bar (8)

Hard — the separator tail (2)

The planHow to prepare for combinatorics questions

Five steps, ordered so each one compounds the last. The goal is not more formulas — it is making a handful of counting moves reflexive under a clock.

  1. Lock the selection toolkit. Permutations, combinations and multinomials until instant — and always say out loud whether order matters and whether groups are labeled before writing a formula.
  2. Default to the complement. Whenever a condition reads “at least”, “not all” or “avoids”, count the bad set first; escalate to full inclusion–exclusion only when the bad sets overlap.
  3. Practice counting twice. Double counting, pairing and symmetry arguments turn brutal sums into one-liners — drill them until “how often is each element counted?” is your first question.
  4. Recurse on the first choice. For nested or ordered structures, condition on the first step and let a small recursion (Catalan-style where the structure splits) do the counting.
  5. Time yourself on the free set. Work the 25 free problems above in difficulty order and write your count before opening the solution; then continue in the full combinatorics bank.

FAQCombinatorics interview questions — frequently asked

How important is combinatorics for quant interviews?

It is one of the standard quant-interview topics — counting is the machinery under most discrete-probability questions, so it is tested both directly and through probability. Our bank holds 171 combinatorics problems, and their firm tags cluster most around Jane Street, Citadel, Optiver and SIG.

What combinatorics topics should I focus on?

Make five moves reflexive: arrangements and selections (permutations, combinations, multinomials), complementary counting and inclusion–exclusion, symmetry and double counting, small recursions (including Catalan-style structures), and information-style weighing or adaptive-search arguments. Digit-counting and pigeonhole round out the tail.

How hard are combinatorics interview questions?

Mostly easy-to-medium with a small hard tail: our 171-problem combinatorics bank splits 77 easy, 78 medium and 16 hard. The 25 free problems on this page mirror that shape (15 easy, 8 medium, 2 hard); a clean timed medium is a fair bar for most interviews.

Are these real quant interview questions?

They are representative practice, not leaked wordings. Problems are drawn from our curated bank and rewritten for clarity with full worked solutions we author ourselves; where a firm tag appears it reflects candidate-reported question shapes, and we never claim any wording is verbatim.