How to Explain Quant Interview Solutions
How to explain quant interview solutions clearly: state assumptions, narrate choices, show checks, and recover from mistakes.
Candidates who can solve practice problems but struggle to communicate live.
Say the model before the math
The first job is to make the model visible. Say what the random variable is, what event you are counting, whether order matters, and what you are assuming. This costs about fifteen seconds and it lets the interviewer correct your setup before you spend five minutes computing the wrong thing. It also changes what a wrong answer means: an interviewer who watched you build a correct model reads a slip as a slip, while an interviewer who saw only a number reads it as luck running out.
Narrate decisions, not noise
You do not need to read arithmetic aloud. Explain the choices that carry information: why the complement is easier here, why linearity applies even though the variables are dependent, why this problem needs a state recursion, why the market should be wide. A useful test is whether a listener could reconstruct your solution from the narration alone, without seeing your work. If they could, you narrated the decisions. If they could not, you narrated the digits.
Use checks as part of the explanation
A check is not an apology; it is part of the answer. Say whether the result lies between zero and one, whether the units match the question, whether an extreme case behaves correctly, or whether a fair price sits inside the payoff range. Doing this unprompted signals that you would catch your own errors on a desk, which is a large part of what the interview is trying to find out.
Concrete example
For "what is the chance a five-card hand contains exactly two aces," a clear explanation runs like this: "Order does not matter, so I will count unordered hands. The denominator is C(52,5). For the numerator I choose two of the four aces and three of the remaining forty-eight cards, so C(4,2) x C(48,3) = 6 x 17,296 = 103,776. Over C(52,5) = 2,598,960 that comes to about 0.04, so roughly four percent." The model came first, the formula second, and the plausibility check - four percent is reasonable for a fairly specific hand - came last.
Recover cleanly from mistakes
If you spot an error, name it and repair it out loud: "I treated the second draw as independent, but the deck changes, so the denominator should be 51." Interviewers are explicitly looking for this, because it is what error correction looks like on a real desk. Quietly patching the number, or carrying a broken answer forward and hoping, costs far more than the original mistake did.
Practice path
Record yourself solving one problem per session, or explain it to someone who does not already know the answer. Review for exactly three things: did you state the model before computing, did you justify the method, and did you check the result. If any of the three is missing, redo the explanation on the same problem instead of moving to a new one - the thing you are practicing is the explanation, not the problem.
Practice the pattern
Use the LeetQuidity curriculum and calibration to turn this topic into a focused practice plan.