Common Quant Interview Mistakes
Common quant interview mistakes in probability, expected value, market making, mental math, and communication, with practical fixes.
Candidates whose quant interview practice is inconsistent or weaker under live pressure.
Mistake 1: solving before modeling
Most misses start before the first calculation. Candidates hear a probability prompt and reach for a formula without defining the sample space, the event, or the conditioning information. The classic case is the two-child problem: "a family has two children and at least one is a girl - what is the chance both are girls." The answer is 1/3, not 1/2, because the condition eliminates only the boy-boy outcome and leaves three equally likely cases, one of which is girl-girl. The fix is boring and effective: state what the outcomes are, what counts as success, and whether order, replacement, or stopping changes the denominator.
Mistake 2: hiding the reasoning
Silent algebra is risky in quant interviews. If the interviewer cannot see your model, they cannot distinguish a small arithmetic slip from a broken setup, and they will assume the worse of the two. Narrate the decisions rather than every arithmetic detail - "I am using the complement, because at least one is easier to count as none" tells them everything they need. It also buys you the correction: an interviewer who hears your setup stops you at second thirty instead of minute five.
Mistake 3: treating speed as the whole score
Speed matters only once the method is stable. Candidates who rush double-count, assume an independence that is not there, or quote markets far too tightly. A candidate who quotes 790 at 810 on a jar they estimated at 800 with honest uncertainty of plus or minus 20 percent has offered free money to anyone better informed, and the interviewer will take it. A fast wrong answer is rarely better than a slightly slower answer with a clear model and a sanity check.
Concrete example
For "two cards are drawn from a shuffled deck - what is the chance both are aces," the common mistake is using 4/52 twice, which gives 1/169 or about 0.59 percent. The second draw depends on the first, so the answer is (4/52) x (3/51) = 12/2652 = 1/221, about 0.45 percent. The unordered route gives the same number, C(4,2)/C(52,2) = 6/1326 = 1/221, which is itself a useful habit: if two correct framings disagree, one of them has an ordering error in it.
Mistake 4: ignoring review data
Practice only improves if misses change the next session. If ten errors trace back to conditioning and the next block is random market making games, the error log is decorative. Tag every miss by method - setup, method choice, arithmetic, time pressure, communication - and let the tag counts choose the next block. In most error logs two tags account for the majority of the damage.
Repair path
After each session take the single most repeated failure mode, write a one-line rule for it, and solve three variants of that exact pattern before returning to mixed practice. Retest it cold a few days later, because a rule you can only apply when you are expecting it has not been learned. The LeetQuidity calibration can help identify whether the underlying issue is probability setup, expected value, mental math, or communication under pressure.
Practice the pattern
Use the LeetQuidity curriculum and calibration to turn this topic into a focused practice plan.