Russian Roulette Odds

Russian Roulette win probability is a medium quant interview question on Conditional Probability, reported to have been seen at Jane Street.

Difficulty Medium Topic Conditional Probability Reported at Jane Street

MyQuantPartner is not affiliated with, endorsed by, or sponsored by these companies, and all trademarks belong to their respective owners.

This Russian Roulette question is about conditional probability in a sequential game with only one randomization at the start. It forces you to think carefully about how an initial random state governs all later outcomes, even as the players alternate in a seemingly dynamic process. Instead of being a simple coin-flip model, it's a structured sequence tied to a hidden configuration.

It trains you to map initial states to deterministic paths and to count which of those paths are favorable, a key skill in quant prep. You practice building a clean probability space, understanding dependence between events, and resisting the temptation to re-randomize when the model says you cannot.

This matters for quant interviews because real trading systems often involve one initial random draw followed by deterministic evolution. Interviewers want to see that your probabilistic intuition matches the model, not the narrative.

What it tests

When dealing with sequential probabilistic games where the randomization occurs only once at the start, the key is to view the process as a deterministic sequence governed by the initial random event. Each possible outcome is mapped directly to a specific sequence of events, so the probability of winning is determined by counting the favorable initial configurations, not by recalculating probabilities at each step. This approach works because, after the initial randomization, the process is entirely predictable—no further randomness is introduced. The structure of the game is thus a mapping from initial states to outcomes, and the symmetry or periodicity of the sequence often allows for simple counting arguments. The principle holds because the lack of re-randomization means the future is fully determined by the initial state, making the problem one of combinatorial enumeration rather than dynamic probability updating.

Practise this question with written feedback, or hear it in a spoken mock interview.

Get started free