← Back to blog
SIG

SIG Problem Solving Assessment Real Questions: Probability, Geometry & Combinatorics (9 Questions)

2026-03-26

The SIG PSA (Problem Solving Assessment) is one of the more rigorous quant OAs out there, covering probability theory, expected value, Bayesian reasoning, geometric optimization, and combinatorics. Below is a full breakdown of 9 recent questions with answers and solution notes.


Q1: Theatre Meeting Probability

Donald and Goofy each arrive at a theatre at a uniformly random time between 18:00 and 19:00. Donald will only wait 12 minutes for Goofy, but Goofy will wait up to 36 minutes for Donald. What is the probability they meet?

Answer: 3 / 5

Let arrival times D, G ∈ [0, 60] (minutes). They meet when −36 ≤ D − G ≤ 12. The non-meeting region consists of two right triangles with areas 48²/2 = 1152 and 24²/2 = 288. Total non-meeting area = 1440 out of 3600. P(not meet) = 2/5, so P(meet) = 3/5.


Q2: Shortest Path Home Through the Swamp

A dog runs at 3 m/s in the swamp and 5 m/s on the road. What is the shortest time (in seconds) for her to reach home from the position shown in the diagram?

Answer: compute from diagram coordinates

This is a classic refraction/Snell's law optimization. Let the dog travel through the swamp to an entry point on the road, then along the road. Minimizing total time requires sinθ₁/3 = sinθ₂/5, where θ₁ and θ₂ are the angles from vertical in the swamp and road segments respectively. Substitute the coordinates from the diagram to find the optimal entry point and compute total time.


Q3: Expected Turns for Two Consecutive Black Tiles

Iggy draws tiles with probability p of black each turn. What is the expected number of turns to place two black tiles consecutively?

Answer: 12 / 1 (when p = 1/3)

Let E = expected turns from scratch, A = expected turns after just drawing a black tile.

Solving these simultaneously gives E = 1/(p²) + 1/p. For p = 1/3: E = 9 + 3 = 12.


Q4: Age Logic Puzzle

Aaron, Bryce, and Craig have distinct integer ages. When speaking to someone older, the speaker always tells the truth; when speaking to someone younger, the speaker always lies.

Answer: Craig = 27

Since Aaron < Bryce, Aaron tells Bryce the truth: Bryce = 1.7 × Aaron. For integer ages, Aaron must be a multiple of 10. Set Aaron = 20, Bryce = 34. Since Aaron < Craig, Aaron tells Craig the truth: Craig = (20 + 34)/2 = 27. Since Craig > Aaron, Craig lies to Aaron: the statement "I'm at least 8 older" is false, meaning Craig − Aaron < 8. Indeed 27 − 20 = 7 < 8. All conditions check out: Craig = 27.


Q5: Expected Value on a 12-Sided Die

Roll a fair 12-sided die (uniform on {1, …, 12}) 10 times. Let X be the number of rolls divisible by 4. Find E[X].

Answer: 5 / 2

Multiples of 4 in {1, …, 12}: 4, 8, 12 — that's 3 values. Single-roll probability p = 3/12 = 1/4. X ~ Binomial(10, 1/4), so E[X] = 10 × 1/4 = 5/2.


Q6: Probability Largest Number Falls in an Interval

Three numbers are drawn independently and uniformly from U[0, 4]. What is the probability that the largest is between 0.5 and 1?

Answer: 7 / 512

P(max ∈ (0.5, 1)) = P(all ≤ 1) − P(all ≤ 0.5) = (1/4)³ − (0.5/4)³ = (1/4)³ − (1/8)³ = 1/64 − 1/512 = 8/512 − 1/512 = 7/512.


Q7: Placing Items on a Black-and-White Grid

A 3×3 grid has 2 black squares and 7 white squares. Place 2 indistinguishable apples and 1 orange such that at least one item is on a black square (at most one item per square). How many ways?

Answer: 147


Q8: Toad Path Count

A toad moves from A(0,0) to B(13,4), only moving right or up, with step sizes strictly alternating between 1 and 3 (may start with either). How many valid paths exist?

Answer: compute via DP

Total displacement: 13 right + 4 up = 17 units. The alternating step constraint limits valid step sequences. One valid decomposition: 5 steps of size 1 and 4 steps of size 3 (total = 5 + 12 = 17), arranged in alternating order starting with 1 (sequence: 1,3,1,3,1,3,1,3,1). For each valid step sequence, count the number of ways to assign each step a direction (right or up) such that total right = 13 and total up = 4. Enumerate via dynamic programming over step index and cumulative right displacement.


Q9: Bayesian Inference — Most Biased Coin

Three biased coins have heads probabilities 9/10, 4/5, and 11/20 respectively. A coin is chosen uniformly at random and flipped — it lands tails. What is the probability the most biased coin (9/10) was chosen?

Answer: 2 / 15


Summary Table

Question Topic Answer
Q1 Geometric probability 3/5
Q2 Geometric optimization (Snell's law) from diagram
Q3 Markov chain expectation 12
Q4 Logic / truth-lie deduction 27
Q5 Binomial expectation 5/2
Q6 Order statistics 7/512
Q7 Combinatorics 147
Q8 Path counting (DP) enumerate
Q9 Bayes' theorem 2/15

💬 Need OA / VO Support?

Ongoing coverage of quant OA questions from SIG, Jane Street, Optiver, Two Sigma, Citadel, and more. Feel free to reach out with questions.

WeChat: Coding0201


📬 Contact

Email: [email protected]
Telegram: @OAVOProxy