Oct 08, 2026 by Offer

Citadel Waiting Time Problem 2027: Classic Probability

The Citadel waiting time probability question — commonly reported by candidates as "Buses arrive at random. How do you calculate expected waiting time?" — hinges on what "at random" means. For Poisson arrivals, the memoryless property gives an expected wait equal to the mean headway: buses averaging one every 10 minutes mean a 10-minute expected wait, not 5.

What This Citadel Waiting Time Probability Question Assesses

This tests modeling judgment: the answer depends entirely on the arrival process you assume, and strong candidates say so before computing. Citadel wants to see you distinguish the Poisson case (memoryless, wait equals headway) from the fixed-schedule case (uniform wait, half the headway) — and explain why randomness makes waiting longer.

How to Answer This Citadel Waiting Time Probability Question

Structure the answer around the two models:

Step 1 — Clarify the model. Ask or state your assumption: "I'll model arrivals as a Poisson process with rate λ — is that the intended reading?" This single sentence shows modeling maturity.
Step 2 — Poisson case. Memorylessness means the time since the last bus tells you nothing; the remaining wait is always Exponential(λ) with mean 1/λ. So with a 10-minute mean headway, your expected wait is 10 minutes.
Step 3 — Explain the paradox. Random arrival times oversample long gaps — you're more likely to arrive during a 20-minute gap than a 2-minute one — so the average wait exceeds half the headway. Contrast with a fixed schedule: buses exactly every 10 minutes give a uniform wait averaging 5 minutes.

Example line: "Under Poisson arrivals the wait is memoryless, so I'd expect the full mean headway — 10 minutes for 10-minute average spacing. That's the inspection paradox: random arrivals oversample long gaps."

Common Mistakes

Answering 5 minutes without stating the model. Half the headway is correct only for fixed schedules — presenting it as the answer shows you missed the paradox entirely.
Confusing the two models. Mixing Poisson reasoning with uniform-wait conclusions (or vice versa) is worse than picking one model cleanly.
No intuition. Stating the memoryless property without explaining why random arrivals lengthen the wait misses the insight the interviewer is probing.

Candidates commonly report follow-ups like "what if buses come in pairs?" or "what's the distribution of the wait, not just the mean?" The follow-ups test whether you truly understand the model or just memorized the paradox — so make sure you can explain the gap-oversampling intuition in your own words.

Keep Reading

FAQ

What exactly is the inspection paradox? A random observer samples time intervals proportionally to their length, so long gaps are overrepresented. Your wait isn't the average gap — it's the average gap as experienced by a random arrival, which is longer.

Does the expected wait really equal the full headway? For a Poisson process, yes: E[wait] = 1/λ = mean headway. It's counterintuitive but exact, a direct consequence of memorylessness.

What if arrivals are scheduled but sometimes late? Then the truth lies between the models: more regular than Poisson, less than perfect schedule. You'd need the lateness distribution to compute it exactly.

Where does this appear in trading? Order arrivals, quote updates, and fill times are all modeled as point processes — the same memoryless logic underpins how quants think about waiting for fills.

Preparing for Citadel's interview? Our 2027 Citadel Online Assessment and Coding Challenge Tutorials has practice questions and answers — $79 one-time, instant download.