The Goldman Sachs Markets Simulation: What to Expect & How to Prepare

Goldman Sachs received more than 360,000¹ applications for its 2025 internship class and accepted fewer than 1%¹ of them (Fortune, 16 June 2025). Its markets simulation has become a critical differentiator in how the firm identifies talent. Here is what it tests and how to prepare for it.

The Scale of Goldman Sachs Recruiting

The direction of travel is documented rather than asserted. In April 2022 the bank said a record 236,000 people globally applied to its internship programme, a 17% jump on 2021 (Business Insider, 5 April 2022); by the 2025 class the figure was more than 360,000, up 300% since 2018 (Fortune, 16 June 2025). At its 2023 Investor Day the firm put the ratio at roughly 300 applicants per open position (eFinancialCareers, February 2023).

What most candidates don't realize is that Goldman has been quietly transforming its hiring process. Traditional resume screens and GPA cutoffs are giving way to skills-based assessments. The firm's markets simulation, launched as part of their New Analyst Program, is central to this shift.

What the Markets Simulation Actually Tests

Unlike a typical case interview, the Goldman Sachs markets simulation places candidates in a real-time trading environment where decisions have measurable consequences. Based on interviews with former Goldman analysts and publicly available program details, the simulation evaluates five core competencies:

1. Market Intuition Under Pressure

2. Risk Assessment and Management

3. Quantitative Reasoning

4. Communication and Client Interaction

5. Ethical Decision-Making

How to Prepare: A Data-Driven Approach

Build Market Literacy (4-6 Weeks)

Master the Fundamentals (2-4 Weeks)

Ensure you can explain discounted cash flow, WACC, P/E ratios, and basic options pricing without hesitation. Goldman's recruiting FAQ explicitly states they test "foundational financial knowledge."

Practice Under Time Pressure

Use Bloomberg's Market Concepts (BMC) certification or Coursera's "Financial Markets" by Robert Shiller as structured preparation. The key is simulating the pressure: set a timer and practice making decisions with incomplete information.

Develop a Market Thesis

Come prepared with a view on at least three asset classes. As one Goldman VP told Business Insider in 2024: "We don't care if you're right. We care if you can articulate why you believe what you believe."

What Happens After the Simulation

The simulation is typically one stage in a multi-round process:

• Application Review: Resume + HireVue video (1-2 weeks)
• Online Assessment: Including the markets simulation (1 week)
• Superday: 4-5 back-to-back interviews at Goldman's offices
• Offer Decision: Typically within 1 week of Superday

The Bigger Picture

SHRM found that 56% of employers use pre-employment assessments and 79% say skills assessments are just as or more important than other hiring criteria: a figure for employers generally, not for financial services alone. Source: SHRM, "New SHRM Research Makes the Case for Skills-Based Hiring", 15 August 2022.

Sources

• Fortune (16 June 2025). “A record 360,000 students applied for a Goldman Sachs internship this year, and fewer than 1% were accepted.”
• Business Insider (5 April 2022). “A record 236,000 students are vying to become interns at Goldman Sachs.”
• eFinancialCareers (February 2023). “Goldman Sachs has 300 applicants per open position.”
• SHRM (15 August 2022). “New SHRM Research Makes the Case for Skills-Based Hiring.”

What a markets simulation is actually measuring

Candidates prepare for a Goldman Sachs markets simulation as if it were a knowledge test, and it is not. The exercise compresses a trading floor day into a few hours precisely so assessors can watch decision behaviour under incomplete information: how you size a position relative to conviction, how fast you cut when the thesis breaks, whether you can explain a view in thirty seconds, and whether your stated risk matches the risk you actually take.

Two candidates can end the simulation with identical profit and be scored very differently. The one who articulated a hypothesis, defined the invalidation level in advance, and closed the trade when that level was hit scores well even on a loss. The one who won on an unexplained hunch and doubled after being wrong scores badly on a gain, because the process does not survive repetition. Markets divisions hire for repeatable process, not for lucky outcomes.

The five behaviours assessors record

• Thesis articulation. A view stated as a mechanism (what happens, why, and what would falsify it) rather than a direction.
• Risk expression. Position size proportional to conviction and to the loss you can absorb; a pre-declared stop that you honour.
• Speed of updating. Evidence against the thesis producing a change in behaviour rather than a search for confirmation.
• Client framing. In sales and structuring rounds, whether the candidate explains a product's risk before its upside. Failing to do so reads as a conduct risk, which is the fastest way to lose an offer.
• Composure. A managing director will interrupt, disagree, and add constraints mid-answer. What is being tested is whether you stay analytical while being pressured, not whether you can be pressured.

The technical floor, and where candidates lose marks

The quantitative bar is real but narrower than folklore suggests. Expect mental arithmetic under pressure, bond price and yield intuition, basic option payoff and sensitivity reasoning, and the ability to convert a macro headline into an expected effect across rates, FX, credit and equity. What sinks candidates is rarely the maths; it is unbounded confidence: quoting a number without stating the assumption behind it, or answering a question about magnitude when they only know the direction.

The professional formulation is direction, then magnitude with an explicit assumption, then the thing that would make you wrong. That structure is the vocabulary of the floor, and using it naturally signals a candidate who has done the work rather than memorised outcomes.

A four-week preparation protocol

1. Week 1: build the market map. For each major asset class, write the two or three drivers you would monitor and how they transmit. This turns news into structure rather than noise.
2. Week 2: run a paper book with a written journal. Three positions maximum, each with a thesis, a size rationale, an invalidation level and a review note when it closes. The journal is the artefact that makes your process describable in the interview.
3. Week 3: drill the mechanics. Timed arithmetic, yield and duration intuition, option payoff sketches, and a daily one-minute verbal market summary recorded and reviewed for concision.
4. Week 4: rehearse under interruption. Have someone challenge your view mid-sentence and add constraints. Practise the three moves that work: acknowledge the point, state what would change your view, and give a bounded answer.

The candidates who convert in a markets simulation are not the ones who predicted the market. They are the ones whose reasoning an assessor could reconstruct afterwards and would be comfortable putting in front of a risk committee, which is, precisely, what the exercise was built to find.

Withdrawn figures

On 8 September 2026 this page was audited against its own reference list, document by document. The figures withdrawn from it are the 236,000-applications-for-3,000-positions pairing and the 1.3% acceptance rate, the 33% rise since 2020 attributed to an annual report, the CFA Institute's 0.71/0.21/0.38 correlations, Goldman's internal 2.3x risk-discipline multiple, LinkedIn's 67% and 41% quantitative-skills shares, Harvard Business School's 34% attrition share, the 45% rise in ethics training hours, Handshake's 2.8x reading multiple, Glassdoor's 6.2-week median and its pay figures, SHRM's 78%/23% financial-services pair, and the claim that simulations predict performance three times better than interviews. No primary document published by a named body could be found printing them at any edition a reader can open, so they have been withdrawn rather than left standing on a reference nobody can check. Every figure that remains above is credited in its own sentence to a document a reader can open at a stated date.