Fast Mental Math For Quant Interviews
This conversation was generated using Cook'd AI's career assistant and reviewed by Fallon Han from our editorial team for accuracy.
I have a quant research internship interview in Sydney, and I keep getting tripped up by fast estimation questions. I know the math, but I freeze when the interviewer wants a back-of-the-envelope answer. How should I handle it cleanly? This is for a hedge fund, and I do not need visa sponsorship.
Yes. For these questions, the interviewer usually cares more about structure than a perfect number. Start by restating the problem in simple terms, define the unit you are estimating, and then build the answer in layers. For example: identify the population, make a few reasonable assumptions, do one rough multiplication at a time, and say your logic out loud. If you are unsure, give a range and explain what drives the upper and lower bound. That shows judgment, which matters a lot in quant interviews.
Can you give me a concrete framework I can use in the interview so I do not ramble?
Use a four-step structure: 1. Clarify the target quantity. 2. Break it into factors you can estimate. 3. Make assumptions explicit and round aggressively. 4. Sanity-check the result against what should be plausible. For instance, if asked about daily trading volume or number of market events, think in terms of participants, frequency, and size. Keep each step short so the interviewer can follow the chain.
What if the question is something like, estimate how many shares are traded on the ASX in a day? I worry I will miss the scale.
That is exactly where structure helps. You could say: first estimate the number of active listed names, then the average shares traded per name, then separate large-cap and small-cap behavior if needed. For example, suppose the ASX has a few hundred actively traded stocks, but a small set does most of the volume. Estimate heavy names at very high turnover and the rest at much lower turnover. The point is not to nail the exact market total, but to show a sensible decomposition and a believable order of magnitude.
How do I avoid sounding like I am just making up numbers?
Anchor each number to a reference point. Say things like, 'I am assuming around X because this seems consistent with a liquid developed market,' or 'I will round to keep the arithmetic simple.' Also, if you can compare against a familiar benchmark, do it. A good candidate will sound calm, state assumptions clearly, and adjust if the interviewer pushes back. That is often stronger than trying to sound overly precise.
So in the interview, should I optimize for speed or for showing the reasoning?
Prioritize reasoning first, speed second. A clean, organized answer in 90 seconds is better than a rushed one in 30 seconds that feels shaky. If you can, verbalize the structure before computing: 'I will estimate this from the number of names and their average turnover.' Then do the math in chunks and conclude with a quick sanity check. That style fits quant research, especially for a fresh graduate with a math background, because it shows disciplined thinking rather than memorized tricks.