Quantitative Analytics interview prep.

Behavioural questions to expect

1. Walk me through your CV.
2. Tell me about your most rigorous piece of energy / commodity quant work.
3. Why energy / commodity trading quant - and not pure financial-markets quant or pure power-systems engineering?
4. Why this commodity / market mix - the sector, power vs gas vs LNG vs emissions, US vs European vs Asian?
5. Why the firm?
6. What's your read on our trading book and modelling stack?
7. How do you think the firm manages model risk + market risk?
8. Walk me through how you'd build a hub-price forecast for the sector - power, gas, or LNG - over the next 12-24 months.

Technical concepts to master

• The fundamentals stack

  Merit order + marginal fuel · Load + weather forecasting · Renewables + ELCC · Outages + storage · Transmission + transport constraints

• Spreads + option pricing

  Spread as option · Margrabe + Kirk approximations · Vol calibration · Path-dependent + constrained options · Greeks + hedging

• Asset valuation - intrinsic + extrinsic + LSMC

  Intrinsic value · Extrinsic value · Least-squares Monte Carlo (LSMC) · Operational constraints · Hedge ratios + rebalancing

• Backtest hygiene + model risk for energy quant

  Look-ahead via revisions · Regime + fuel-switch dependence · Realistic transmission / transport costs · Tail-event dependence · Model validation + kill criteria

Practical drills

• A gas peaker has heat rate 9.0 MMBtu/MWh and VOM $3/MWh. Front-month power forward is $50/MWh, gas $3.50/MMBtu. Power vol 60%, gas vol 40%, correlation 0.6. (a) Compute the intrinsic spark spread per MWh. (b) Compute the spread vol. (c) Roughly estimate the extrinsic value for a 1-month at-the-money option assuming 21 trading days.
• Walk through how you'd value a 10 Bcf gas storage facility with 200 MMcf/day max injection + 300 MMcf/day max withdrawal, currently 50% full, against TTF or the US national gas hub forward curves. 8-10 min delivery, 10-15 min Q&A.
• A candidate shows a model: Sharpe 3.5 over 2019-2023 on a basis-trading strategy in PJM (West hub vs Eastern hubs). Backtested by trying ~500 specifications, run on revised ISO load data, valued at hub-to-hub spread with no transmission cost, no FTR cost. What's wrong, and how would you actually validate it?

Smart-question anchors

• Book + asset access - which commodities + markets the desk trades + which physical assets the quant supports
• Modelling stack + tooling - fundamentals models, price + spread models, LSMC + risk, build vs buy stance
• Research-to-deployment pipeline - how a model goes from research to live capital + who owns model validation
• Risk + model validation - market risk + model risk regime, stress + scenario, kill criteria for a degrading model
• Data edge - alt data, weather data, fundamentals data, vintage-stamped discipline

Sourced from

Eydeland & Wolyniec - Energy and Power Risk Management · Burger, Graeber, Schindlmayr - Managing Energy Risk · Longstaff & Schwartz - Valuing American Options by Simulation · FERC + ISO / RTO market design + EIA data · Wall Street Oasis + Global Derivatives community (energy desk threads) · Bailey & Lopez de Prado - Deflated Sharpe Ratio + backtest overfitting