Prepare · Application Season

The Assessment Centre

What actually happens on the day, what each exercise is really testing, and how to practise for it. Built from what candidates report back after sitting them.

💡
Read this first. No two firms run the same day. A systematic fund screens very differently from a long-only asset manager or a market maker. Treat the stages below as the menu, not the fixed order, and always ask your recruiter what your specific day looks like. They usually tell you if you ask directly.
The Process
From application to offer
Stage 01

Online tests

Numerical and logical reasoning, sat at home under time pressure. Quant and systematic seats add a coding round on HackerRank or CodeSignal. This stage is a filter, not a differentiator: firms set a cutoff and everyone below it disappears. The good news is that it is the most improvable part of the whole process, because the question types repeat.

⏱ Usually 30 to 60 minutes
Stage 02

Recorded video interview

Pre-set questions, 30 seconds to think and 90 seconds to answer, no human on the other end. Almost entirely a fit and communication screen. The single biggest fix is treating it like a presentation rather than a conversation: structure every answer, look at the lens and not your own face, and record a few practice runs before the real one.

⏱ 15 to 25 minutes
Stage 03

First round with the desk

Usually 30 to 45 minutes with one or two people who actually do the job. Expect your CV walked line by line, a market view you have to defend, and light technicals. Anything you wrote on your CV is fair game, so know your own projects cold.

⏱ 30 to 45 minutes
Stage 04

The assessment centre

The full day, either on site or over video. Some mix of group exercise, case study or trade pitch, mental maths, brainteasers, and back to back interviews. This is where most of the decision gets made, and it is what the rest of this page covers.

⏱ Half day to full day
Stage 05

Final round and offer

A senior conversation, often with the PM or a partner. Less technical, more about whether they want to sit next to you for the next few years. Have your questions ready and make sure they are questions you could not have answered with ten minutes on their website.

⏱ 30 minutes, then a decision in days
On the day
Every exercise, and what it is really testing

The group exercise

Six to eight candidates, one scenario, thirty minutes, and two assessors saying nothing while they write. Typical briefs: allocate a fixed budget across competing projects, agree an asset allocation for a mandate, or decide which of five companies to invest in. The scenario is deliberately underdetermined. There is no correct answer, and that is the point.

✅ What scores

  • Bringing in the person who has not spoken yet, by name
  • Watching the clock out loud: "we have ten minutes, let us lock the first two and move on"
  • Building on someone else's point rather than restarting the discussion
  • Changing your mind when the argument against you is good, and saying why
  • Writing the group's conclusion down so the summary is accurate

❌ What fails

  • Talking the most and listening the least
  • Silence, then one long speech at minute 25
  • Steamrolling a quiet candidate to win a point
  • Running out of time with no decision reached
  • Repeating a point already made, louder

Marked on

Influence without dominance Listening Time management Structured reasoning Handling disagreement

The trade idea presentation

Either prepared in advance or built in a 45 minute prep room, then pitched for five to ten minutes with questions after. This is the exercise that separates candidates most sharply, because a mediocre idea argued rigorously beats a fashionable idea argued loosely, every single time.

✅ A pitch that lands

  • One sentence up front: the trade, the direction, the horizon
  • A view that differs from consensus, and you say what consensus is
  • The catalyst, with a date or a trigger attached
  • Explicit sizing, entry, stop, and target
  • The strongest argument against you, addressed before they raise it
  • What would make you exit and admit you were wrong

❌ What sinks it

  • A company description instead of a thesis
  • "It is undervalued" with no mechanism for the gap to close
  • No risk section, or a risk section of generic boilerplate
  • Numbers you cannot defend when questioned
  • An idea from a podcast you have not pressure tested
  • No answer to "why now"

Marked on

Thesis clarity Risk awareness Conviction under challenge Commercial instinct Knowing what would prove you wrong

The case study

A pack of materials, a quiet room, and a deliverable. On the PE side it is usually an LBO or a commercial assessment of a target. On the fund side it is more often a data set with a question attached, or a portfolio you have to critique. The pack always contains more than you can use, which is itself part of the test.

✅ How to work it

  • Read the question first, then the pack, and only take what answers it
  • Spend the first five minutes planning the output, not reading
  • State your assumptions on the page where you use them
  • Sense check every number: does this order of magnitude make sense
  • Leave time to write the recommendation properly, it is what gets read

❌ Where people lose

  • Building an elaborate model and never reaching a conclusion
  • Describing the data back instead of deciding something
  • Precision theatre: four decimal places on a guessed input
  • Missing the one exhibit the whole case turns on
  • Running out of time in the appendix

Marked on

Prioritisation Numeracy under time Assumption discipline Clear recommendation Defending it in Q&A

Mental maths and the numerical test

Standard at market makers and trading desks, increasingly common elsewhere. Usually 60 to 80 questions in eight minutes, no calculator, heavily weighted toward percentages, fractions, and sequences. Speed matters, but so does knowing when to skip. The drill below is built on the same question shapes.

✅ Techniques that pay

  • Learn fraction to decimal conversions to sixteenths by heart
  • Percentages by decomposition: 17% is 10% plus 5% plus 2%
  • Round, compute, then correct the error at the end
  • Squares to 30 and powers of 2 to 4096 on instant recall
  • Skip anything that stalls you for more than a few seconds

❌ What costs marks

  • Grinding one hard question while ten easy ones expire
  • Long division when estimation would have been accepted
  • Practising untimed, then freezing when the clock appears
  • Mental arithmetic without a scratch pad habit

Marked on

Raw speed Accuracy under pressure Triage instinct Composure after a mistake

Technical interview

Scoped to the seat. Rates and macro get curves, carry, and duration. Equities get valuation and accounting. Quant seats get probability, statistics, and code. The common thread is that they push until you reach the edge of what you know, so reaching that edge is expected and not a failure. What matters is what you do when you get there.

✅ The right reflexes

  • Think out loud so they can follow and correct you
  • Say "I do not know, here is how I would work it out" and then do that
  • Define terms before using them
  • Connect the concept to something you have actually built or traded
  • Ask a clarifying question when the problem is genuinely ambiguous

❌ Instant red flags

  • Bluffing a definition you half remember
  • Silence while you think, with no narration
  • Not knowing a project on your own CV
  • Memorised answers that collapse on the follow up
  • Arguing when you are corrected

Marked on

Depth in your stated area Reasoning from first principles Intellectual honesty Coachability

Fit and motivation

The round people under prepare because it sounds easy. At a fund of thirty people, culture risk is real and they will cut a technically strong candidate who feels like a bad fit. Every question is a version of the same three: can you do the job, will you stick around, and do we want you here.

✅ What separates candidates

  • Specific reasons for this firm: a strategy, a paper, a person, a recent trade
  • A real failure with a real lesson, not a strength in disguise
  • A coherent story for why this seat follows from your last one
  • Genuine market curiosity that predates the application
  • Questions that show you read past the homepage

❌ What gets you cut

  • "I am passionate about markets" with nothing behind it
  • Reasons that would apply equally to fifty other firms
  • Blaming a previous employer or team
  • No questions at the end
  • Asking about pay or hours before an offer exists

Marked on

Genuine motivation Self awareness Communication Retention risk
Practise
Mental maths drill

⚡ Sixty second drill

Same shapes as the timed tests: percentages, fractions, and sequences. Answer, press enter, keep going. No calculator.

0
Right
0
Wrong
0
Streak
Press start to begin

Under 20 in sixty seconds means keep drilling. Over 35 puts you in good shape for most desks. Market makers set the bar higher again.

Question bank
Brainteasers they actually ask
Try each one properly before revealing. The method is what gets marked, not the number.
Probability
I flip a fair coin until I see two heads in a row. What is the expected number of flips?
6 flips. Set up states by how many consecutive heads you currently have. Let E0 be the expected flips from zero heads and E1 from one head. From E0 you flip once, then land in E1 half the time and back in E0 half the time, so E0 = 1 + 0.5·E1 + 0.5·E0. From E1 you flip once, finish half the time and return to E0 the other half, so E1 = 1 + 0.5·E0. Solving gives E1 = 4 and E0 = 6. The point of the question is whether you build the state machine rather than guess, so always show the setup.
Estimation
How many petrol stations are there in the UK?
Roughly 8,000 to 9,000. Work from cars, not from memory. About 33m cars, each filling up roughly weekly, so around 33m fills a week. A station serves maybe 500 cars a day, so 3,500 a week. That gives about 9,000 stations. State the assumption behind each step out loud. They are checking whether your structure is sound and whether you sense check the answer, not whether you knew the figure.
Markets
The 2 year yield falls while the 30 year rises. What is the market telling you, and how would you express it?
A bear steepener at the long end, or a front end rally on softer policy expectations. The front end is priced off expected policy, so a falling 2 year means the market pushed rate hikes further out or brought cuts forward. The long end reflects term premium, supply and long run inflation, so a rising 30 year points at issuance pressure or fiscal risk. Expressed as a steepener, paying the long end and receiving the front, sized in DV01 so it is neutral to a parallel shift. Reference something live: this is exactly the 7s30s setup desks have been running.
Probability
Three cards: one red both sides, one white both sides, one red and white. You draw one, the face up side is red. Probability the other side is red?
Two thirds. Count faces, not cards. There are three red faces in total: two on the red-red card and one on the mixed card. Given you are looking at a red face, each of those three is equally likely. Two of them sit on the card whose reverse is also red. So the answer is 2/3, not 1/2. Anyone who answers instantly with a half has pattern matched instead of counting, which is precisely what the question is designed to catch.
Logic
You have 8 balls, one heavier. A balance scale. Find the heavy one in two weighings.
Split 3, 3, 2. Weigh the two groups of three. If they balance, the heavy ball is in the pair, so weigh those two and you are done. If one side is heavier, take those three, weigh one against another, and either one drops or it is the third. The generalisation is what they want next: each weighing has three outcomes, so n weighings distinguish up to 3^n items. That is why 8 fits comfortably in two and 9 is the true maximum.
Markets
Oil falls 30% in a month. Walk me through the second and third order effects.
Move outward in rings and be explicit about the chain. First order: energy equities and producer credit fall, energy importers gain. Second order: headline inflation drops, which changes central bank expectations, which moves the front end of the curve and the currency. Third order: energy exporting sovereigns face fiscal strain, high yield energy spreads widen, capex plans get cut which hits industrial demand next year, and rate expectations feed back into equity multiples. Name a real transmission channel at each ring. The recent Brent collapse toward $60 is a live example worth having ready.
Get ready
Your prep plan

📅 Four weeks out

Build the base

⏰ One week out

Sharpen

🎯 The day before

Land it

Practise it with real feedback

Reading about a group exercise is not the same as sitting one. Our August CV clinics and 1-to-1 mentorship sessions run mock exercises and pitch reviews with people who have screened candidates.

📝 Book into Application Season → Book a 1-to-1 session