Hedge Funds interview preparation

Questions

100

Traced to a firm

39

Firms

16

Updated

September 2026

Asked at

All firmsMan Group10Balyasny Asset Management7Bridgewater Associates3DED.E. Shaw3Apollo Global Management2KKR2Oaktree Capital Management2Point722SCSquarepoint Capital2ACAQR Capital Management1BGBaupost Group1Coatue Management1HPS Investment Partners1Northern Trust1Viking Global Investors1Wolverine Trading1

Topic

All topicsStrategy taxonomy8Stock pitch10Short selling6Portfolio construction8Risk and drawdown8Performance and alpha7Event-driven and merger arb8Distressed and credit5Fund structure and economics7Financing, NAV and operations6Compliance and research process5Quant and systematic6India and Category III AIFs5Career and fit11

Level

AnyCoreIntermediateHard

Type

AnyTechnicalMarket viewBrainteaserCaseFit

1. 011Are you sure your thesis can be backed up? What if their costs do not fall?Stock pitchHardsuperdayApollo Global ManagementInvestments · Remote · 2021

   Say this

   Answer the substance, do not defend the position. Give the evidence behind the cost assumption, quantify what the stock is worth if you are wrong, and say where you would cut. Under pressure, the willingness to concede the weak leg is worth more than conviction.

   Then walk it

   1. Evidence first, and be specific. 'Management guided to it' is weak. 'The input contract repriced in Q2, gross margin already moved 180 basis points, and two quarters of run-rate are visible in the reported numbers' is strong.
   2. Then price the downside. 'If costs stay flat, EBITDA is 12 percent below my number, the multiple compresses to peers, and the stock is worth 48 rather than 85. From 62 that is about 22 percent down.'
   3. Then the asymmetry, which is the real defence. 35 up against 22 down still works at even probability, and I would size it accordingly rather than at a full weight.
   4. Then the monitoring plan. Which disclosure tells you early, and by when. If the cost curve is visible in a monthly input price or a quarterly gross margin line, the thesis is testable in real time and that is what makes it a hedge fund position.
   5. Then concede properly where you should. 'You are right that the cost assumption is the weakest leg, so I would start at half size and add on the first print that confirms it' is a better answer than digging in.
   6. And name the structural hedge. If the cost concern is industry-wide rather than company-specific, you can pair the long against a competitor with the same input exposure and isolate the part you actually have a view on.

   Where candidates lose it

   Defending emotionally, or answering a different question than the one asked. This is a test of whether you update on evidence. Candidates who repeat the bull case with more adjectives fail; candidates who quantify the bear case and name a stop pass even if the interviewer keeps pushing.

   Expect next

   • At what price would you stop out?
   • How would you size it given that uncertainty?
   • What would you have to believe for the bear case to be right?

   Reported by candidates at Apollo Global Management (Investments, Remote, 2021). Source: Wall Street Oasis.

2. 012Take me through a structured investment idea the way you would present it to a portfolio manager.Stock pitchHardcase studyPoint72Investment Banking · London · 2026

   Say this

   A PM has five minutes and wants four things: what the trade is, what you know that they do not, what it is worth if you are right and wrong, and what kills it. Structure it in that order and answer the question asked before you show your work.

   Then walk it

   1. Page one is the trade and the sizing. Long, short or paired, entry, target, downside, horizon, and the risk you want to take in the book. Everything else supports this page.
   2. Page two is the variant view with consensus next to your number. A PM reads this page and nothing else if they are busy.
   3. Page three is the mechanism: the two or three drivers that get you from today's numbers to yours, each with the evidence behind it. Volume, price, mix, cost, capital allocation. No page of company history.
   4. Page four is the risk map: the bear case with a price, the two things that break it, the dated falsifiers, and the hedge if the idea has an unwanted factor or sector exposure.
   5. Then the questions you could not answer. Naming them yourself is a credibility move at a platform, because the PM will find them anyway and would rather find them in your appendix than in the P&L.
   6. Business judgement is what is actually being graded in a case like this. That means industry structure, who has pricing power, where the profit pool sits, and whether the company's advantage is durable. A model with no industry view is a spreadsheet, not an idea.

   Where candidates lose it

   Building up to the recommendation. Analysts trained on client decks lead with company overview and market sizing and lose the room. Put the trade in the first sentence, and make sure something in the case shows judgement rather than arithmetic, because that is the explicit test.

   Expect next

   • What is the single best argument against this idea?
   • How would you express it if you could not short the obvious hedge?
   • What would you need to see to double the size?

   Reported by candidates at Point72 (Investment Banking, London, 2026). Source: Wall Street Oasis.

3. 017How did you get the assumptions and calculations in your case study?Stock pitchHardcase studyDED.E. ShawGeneralist · New York · 2025

   Say this

   Go line by line and separate the three kinds of input: facts from disclosure, estimates built bottom-up from a driver, and judgement calls. Say which is which for every important number, and cite the source for the facts.

   Then walk it

   1. Facts first: pull them from primary disclosure and say where. 'Installed base of 1.4 million units is from the 2025 annual report, segment note 4.' Never from a secondary summary if the filing exists.
   2. Then the built estimates. Show the decomposition rather than the result. Revenue is units times price, units are installed base times replacement rate, and the replacement rate comes from the reported life of the asset. Now the number is auditable.
   3. Then the judgement calls, flagged as such. 'I assumed a 60 percent attach rate on the new product. There is no disclosure, so I triangulated from the competitor who does report it and haircut it for their head start.'
   4. Then sensitivity. Say which assumption the answer is actually sensitive to. Most models have one or two inputs that move the value and ten that do not, and knowing which is which is the sign of someone who has built models rather than filled them in.
   5. Then the cross-check. Does the implied market share exceed the whole addressable market in year five? Does the implied margin exceed the best operator in the industry? A bottom-up model with no top-down reality check is where the embarrassing errors live.
   6. Honest close: name the number you are least confident in, and what you would go find if you had another week. Interviewers at quantitative shops are probing whether you know the difference between a number you derived and a number you liked.

   Where candidates lose it

   Saying 'industry reports' or 'I assumed it grows in line with GDP' for a number that drives the whole answer. That reads as reverse-engineering the model to a conclusion. Label facts, estimates and judgement separately, and volunteer the one assumption the valuation is most sensitive to before you are asked.

   Expect next

   • Which assumption is the answer most sensitive to?
   • What is the implied market share in year five?
   • What would you check if you had another week?

   Reported by candidates at D.E. Shaw (Generalist, New York, 2025). Source: Wall Street Oasis.

Firm tags come from public, anonymous candidate reports on Wall Street Oasis: strong signal, not sworn testimony. Firms are named as the places a question was reported, not as partners of Fin Maverick. Answers are written for this page to show how to think out loud; they are not scripts to recite.