Expected Cards 'til King Ace Pair
Expected cards to draw for king ace is a hard quant interview question on Conditional Probability, reported to have been seen at Citadel.
MyQuantPartner is not affiliated with, endorsed by, or sponsored by these companies, and all trademarks belong to their respective owners.
This question is about random stopping times in a deck, where drawing halts as soon as a particular configuration of high-value cards shows up. It lives at the intersection of conditional probability, combinatorics, and expectations, and forces you to reason about how competing stopping events interact as cards are gradually revealed without replacement during a quant prep session.
It trains your ability to compute expected values in systems where multiple triggers can end the process, and where symmetry and conditioning must be used carefully. You practice formalizing a random time as a function of the entire sequence, decomposing it into cases, and tracking how the composition of remaining outcomes changes as information is revealed.
This matters for quant interviews because real trading and risk problems often involve competing exit conditions and path-dependent stopping rules. Top trading firms want to see if you can model these correctly, manage conditional structures under uncertainty, and justify each probability step under pressure.
What it tests
This problem class is governed by the principle of linearity of expectation and the use of 'dividers' to model waiting times for the first occurrence of special events in a sequence without replacement. When tracking the first appearance of one or more types of objects (like specific cards), the expected waiting time can be found by partitioning the sequence into intervals defined by the positions of those objects. The expected number of non-special items before the first special one is distributed evenly among the gaps, regardless of the actual order, due to symmetry. Conditioning on the identity of the first special object allows you to recursively break the problem into smaller, similar subproblems, each with updated populations and stopping rules. This structure holds because the process is memoryless in terms of the remaining deck composition, and the stopping condition is always triggered by a specific event whose probability can be recalculated after each draw.
Practise this question with written feedback, or hear it in a spoken mock interview.
Get started free