← Morgan Stanley Interview Insights

Morgan Stanley·Data Scientist·Technical Phone Screen·Senior

SeniorPrefer not to say
May 2026

Summary

Morgan Stanley data scientist interview with a heavy quant finance slant. Three questions, all centered on Black-Scholes mechanics and hedging PnL, which I was not expecting to be grilled on at this depth for a DS role.

Questions Asked (3)

Q1

For a vanilla European option under Black-Scholes, how would you mentally estimate delta, gamma, and vega without a calculator?

Technical Trade-offsProduct Analytics & Metrics
Author's notes

This was more doable than it sounds.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by recalling the Black-Scholes formulas for delta, gamma, and vega, then use simple approximations and rules of thumb to estimate them mentally. Focus on the key drivers: moneyness, time to expiration, and volatility, and relate them to intuitive ranges.

Pro tip: Mention that for at-the-money options, delta is approximately 0.5, gamma is highest and roughly 0.4/(Sσ√T), and vega is around 0.4S√T. This shows you know the practical shortcuts used on trading floors.

1. Recall the formulas

State the exact Black-Scholes formulas for delta (N(d1) for call, N(d1)-1 for put), gamma (N'(d1)/(Sσ√T)), and vega (S√T N'(d1)).

2. Identify key variables

Note that d1 depends on S, K, r, σ, and T. For mental math, assume r=0 and focus on moneyness (S/K) and total volatility σ√T.

3. Use approximations for N(d1) and N'(d1)

For at-the-money (ATM), d1≈0, so N(d1)≈0.5 and N'(d1)≈0.4. For away-from-the-money, use linear approximations or known values (e.g., delta≈1 for deep ITM call).

4. Estimate gamma and vega

Gamma is highest ATM and decreases as you move away. Vega is also highest ATM and increases with √T. Use the ATM approximations: gamma≈0.4/(Sσ√T), vega≈0.4S√T.

5. Sanity check with ranges

Ensure estimates are within plausible bounds: delta between 0 and 1 (or -1 and 0 for put), gamma positive and typically <0.1 for equities, vega positive and often in the tens or hundreds depending on S and T.

Key Points to Mention

  • Delta is the risk-neutral probability of finishing in-the-money, so it ranges from 0 to 1 for calls and -1 to 0 for puts.
  • Gamma is the rate of change of delta and is highest for at-the-money options, decreasing as the option moves in- or out-of-the-money.
  • Vega is the sensitivity to volatility and is also highest for at-the-money options, increasing with the square root of time to expiration.
  • For ATM options, delta ≈ 0.5, gamma ≈ 0.4/(Sσ√T), and vega ≈ 0.4S√T.
  • Away from the money, delta approaches 1 (ITM call) or 0 (OTM call), and gamma and vega approach 0.
  • Mention that these are approximations and actual values depend on the specific parameters, but the mental math gives a quick sanity check.

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.

Q2

If you can't remember exact values of the Gaussian CDF, what quick approximations or rules of thumb let you estimate the option Greeks on the fly?

Technical Trade-offsAlgorithms & Data Structures
Author's notes

Blanked for a second.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by acknowledging that exact Gaussian CDF values are rarely needed for quick estimates; instead, use approximations like the 68-95-99.7 rule and the logistic approximation. Then explain how these approximations translate to estimating Greeks (delta, gamma, vega, theta) by leveraging the relationships between the CDF, PDF, and option sensitivities. Finally, emphasize the importance of understanding the qualitative behavior of Greeks and how they change with moneyness, time to expiration, and volatility.

Pro tip: Mention that in practice, traders often use the '1-2-3' rule for delta: ATM delta ≈ 0.5, 1 standard deviation ITM/OTM delta ≈ 0.84/0.16, and 2 standard deviations ≈ 0.98/0.02. This shows practical intuition beyond textbook formulas.

1. Recall the 68-95-99.7 rule

Use the empirical rule for normal distributions: about 68% of values lie within 1 standard deviation, 95% within 2, and 99.7% within 3. This gives quick CDF estimates at key points: N(1) ≈ 0.84, N(2) ≈ 0.975, N(3) ≈ 0.9985.

2. Use the logistic approximation

Approximate the normal CDF with the logistic function: N(x) ≈ 1/(1 + exp(-1.702x)). This is easy to compute mentally for any x and is accurate to within 0.01 for most x.

3. Relate CDF to Greeks

For a European call, delta = N(d1), so the CDF approximation directly gives delta. Gamma and vega involve the PDF, which can be approximated by the derivative of the logistic or by using the peak value at ATM (≈0.4) and decaying exponentially.

4. Estimate d1 and d2 quickly

Recall that d1 = [ln(S/K) + (r + σ²/2)T] / (σ√T) and d2 = d1 - σ√T. For quick estimates, ignore small r and use ln(S/K) ≈ (S-K)/K for near-the-money options.

5. Apply qualitative rules for Greeks

Delta is roughly 0.5 ATM, approaches 1 deep ITM and 0 deep OTM. Gamma peaks ATM and decreases as you move away. Vega peaks ATM and increases with time to expiration. Theta is most negative ATM and decreases with time.

Key Points to Mention

  • The 68-95-99.7 rule for quick CDF values at 1, 2, and 3 standard deviations.
  • The logistic approximation N(x) ≈ 1/(1 + exp(-1.702x)) for more precise estimates.
  • Delta as N(d1) for calls, and the symmetry for puts (delta = N(d1) - 1).
  • Gamma and vega are proportional to the normal PDF, which peaks at 0.3989 and decays quickly.
  • Qualitative behavior: delta ranges from 0 to 1, gamma and vega are highest ATM, theta is most negative ATM.
  • The impact of time to expiration and volatility on the width of the distribution and thus on Greeks.

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.

Q3

If realized market volatility differs from the constant implied volatility you used to price and delta-hedge an option, what does the PnL of your hedged position look like, and when does it make or lose money?

Technical Trade-offsRoot Cause Analysis
Author's notes

This is the one I actually enjoyed.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by explaining that the PnL of a delta-hedged option position under constant implied volatility but stochastic realized volatility is driven by the difference between implied and realized variance. Then describe how the PnL can be decomposed into a gamma/theta trade-off and a vega-like exposure to realized volatility, and specify the conditions under which the position makes or loses money.

Pro tip: Emphasize that the sign of the PnL depends on whether realized volatility exceeds implied volatility, and that the magnitude scales with the option's gamma and the square of the volatility difference. Mention that in practice, transaction costs and discrete hedging can erode profits, so the theoretical edge may not be fully realized.

1. Set up the PnL decomposition

Explain that the PnL of a delta-hedged option position can be approximated by the difference between the option's theta and the gamma-weighted realized variance. Under constant implied volatility, the expected PnL is zero if realized volatility equals implied volatility.

2. Introduce realized vs implied volatility

Define realized volatility as the actual volatility of the underlying over the hedging period, and implied volatility as the constant volatility used to price and hedge. The PnL depends on the difference between these two volatilities.

3. Derive the PnL formula

Present the standard result: PnL ≈ 0.5 * Gamma * S^2 * (σ_realized^2 - σ_implied^2) * Δt, summed over the hedging period. This shows that the position profits when realized volatility exceeds implied volatility and loses when it is lower.

4. Discuss the role of gamma and hedging frequency

Explain that the magnitude of PnL is proportional to the option's gamma, which is largest for at-the-money options near expiration. Also note that discrete hedging introduces hedging errors that can add noise to the PnL.

5. Conclude on profitability conditions

Summarize that the hedged position makes money if realized volatility is higher than implied volatility (long gamma) and loses if realized volatility is lower (short gamma). The breakeven point is when realized volatility equals implied volatility.

Key Points to Mention

  • The PnL of a delta-hedged option is approximately proportional to the difference between realized variance and implied variance.
  • The position is long gamma, so it benefits from high realized volatility and suffers from low realized volatility.
  • The PnL is scaled by the option's gamma and the square of the underlying price.
  • Discrete hedging and transaction costs can reduce or eliminate the theoretical PnL.
  • The breakeven point occurs when realized volatility equals the implied volatility used for pricing and hedging.
  • In practice, the PnL can be decomposed into a gamma/theta trade-off and a volatility risk premium.

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.