Energy Quant
Values and hedges physical energy assets such as batteries, plants and gas storage, which are options in disguise.
The job
An energy quant at a utility, trading house or asset owner values and hedges physical assets and contracts: power plants, batteries, gas storage, tolling and supply deals. A battery is an option on the spread between cheap and expensive hours; a storage lease is an option on the spread between seasons. Things go wrong with weather-driven price spikes, forecasts that miss, physical limits such as ramp rates and wear, and hedges that do not match the asset.
- Supports
- Power and gas trading, origination and structuring
- Stages
- 5 to learn, then proof of work
- Two routes
- Free resources, or SelfMadeQuant programmes and projects
The roadmap
The same stages either way. The free route uses university courses, lectures, books and papers that are free online. The SelfMadeQuant route maps each stage to our programmes and ends in projects on real market data.
Stage 01 · Foundations
Commodity and power markets
Cost of carry, forwards, and how power and gas markets clear.
- Cost of carry
- Forwards and futures
- Power and gas market design
- Electricity explained (opens in a new tab)
U.S. Energy Information Administration
How power is generated, traded and delivered, and what drives wholesale prices.
- Natural gas explained (opens in a new tab)
U.S. Energy Information Administration
Gas supply, storage, pipelines and prices, the physical facts behind gas forwards.
- Forwards, Swaps, Futures and Options (opens in a new tab)
Martin Haugh / Columbia University
Cost of carry, forward pricing and futures mechanics with storage and convenience yield.
Stage 02 · Core
Price forecasting
Forecasting prices with strong seasonality and spikes.
- Seasonal time series
- Regression on load and weather
- Electricity price forecasting
- Forecasting: Principles and Practice (3rd edition) (opens in a new tab)
Rob J Hyndman and George Athanasopoulos / Monash University
Seasonal decomposition, ARIMA and regression with multiple seasonalities, with evaluation done properly.
- Forecasting day-ahead electricity prices: A review of state-of-the-art algorithms, best practices and an open-access benchmark (opens in a new tab)
Jesus Lago, Grzegorz Marcjasz, Bart De Schutter and Rafał Weron / arXiv
Benchmark models, open datasets and testing practice for day-ahead power price forecasts.
- Deep Learning for Electricity Price Forecasting: A Review of Day-Ahead, Intraday and Balancing Electricity Markets (opens in a new tab)
Runyao Yu, Derek W. Bunn et al.
Maps which deep learning architectures work in each power market and where they fall short.
Stage 03 · Core
Optimisation
Dispatch as a linear programme, and decisions made over time.
- Linear programming
- Dynamic programming
- Constraints such as ramp rates
- Convex Optimization (opens in a new tab)
Stephen Boyd and Lieven Vandenberghe / Stanford
Linear programming, duality and convex modelling for dispatch and portfolio problems.
- Dynamic Programming (opens in a new tab)
Thomas J. Sargent and John Stachurski / QuantEcon
Bellman equations and value iteration with runnable code, the core of storage and dispatch.
- 6.231 Dynamic Programming and Stochastic Control (opens in a new tab)
MIT OpenCourseWare (Dimitri Bertsekas)
Finite-horizon and stochastic dynamic programming at graduate level.
- CVXPY (opens in a new tab)
CVXPY developers
Python library for writing dispatch and hedging problems as convex programmes.
Stage 04 · Specialist
Real options and Monte Carlo
Valuing assets you can exercise again and again.
- Least-squares Monte Carlo
- Spread options
- Storage valuation
- Valuing American Options by Simulation: A Simple Least-Squares Approach (opens in a new tab)
Francis Longstaff and Eduardo Schwartz / UCLA
The least-squares Monte Carlo method that storage and swing contract valuation build on.
- Gas Storage Valuation Using a Monte Carlo Method (opens in a new tab)
Alexander Boogert and Cyriel de Jong / Birkbeck
Extends least-squares Monte Carlo to gas storage with injection and withdrawal constraints.
- Pricing and Hedging Spread Options (opens in a new tab)
René Carmona and Valdo Durrleman / Princeton
Survey of spread option pricing, from closed-form approximations to numerical methods, as used for spark spreads.
- Introduction to Real Options (opens in a new tab)
Martin Haugh / Columbia University
Frames physical assets as options and shows the lattice methods used to value them.
Stage 05 · Specialist
Energy risk and hedging
Revenue at risk on a forward curve, and explaining a valuation to a commercial team.
- VaR on forward curves
- Hedge effectiveness
- Explaining valuations
- Hedging Exposure to Electricity Price Risk in a Value at Risk Framework (opens in a new tab)
Ronald Huisman, Ronald Mahieu and Felix Schlichter
Chooses power hedge positions by their effect on VaR rather than on variance alone.
- RiskMetrics Technical Document (4th edition) (opens in a new tab)
J.P. Morgan and Reuters / MSCI
Cash-flow mapping and covariance VaR, the template for VaR on a forward curve.
- Energy and Power Risk Management: New Developments in Modeling, Pricing, and Hedging (opens in a new tab)
Alexander Eydeland and Krzysztof Wolyniec / Wiley
Practitioner reference on modelling power and gas prices, forward curve risk and hedging structured deals.
Stage 06 · Proof of work
Value a battery with day-ahead price forecasts
Download hourly day-ahead prices for one bidding zone, build a seasonal forecasting model and feed its forecasts into a linear programme that schedules a battery's charge and discharge. Answer how much of the perfect-foresight arbitrage value a realistic forecast captures.
Stage 01 · Foundations
Commodity and power markets
Cost of carry, forwards, and how power and gas markets clear.
- Cost of carry
- Forwards and futures
- Power and gas market design
- Financial Instruments, Markets and Conventions
- Commodities and the Cost-of-Carry World
- Forwards, Futures, Swaps, and Contracts for Difference
Stage 02 · Core
Price forecasting
Forecasting prices with strong seasonality and spikes.
- Seasonal time series
- Regression on load and weather
- Electricity price forecasting
- Econometrics and Financial Time Series
- Regression Diagnostics and Generalized Least Squares
- Univariate Time Series: ARMA and Box–Jenkins
Stage 03 · Core
Optimisation
Dispatch as a linear programme, and decisions made over time.
- Linear programming
- Dynamic programming
- Constraints such as ramp rates
- Linear Algebra, Optimisation and Numerical Methods
- Optimisation
Stage 04 · Specialist
Real options and Monte Carlo
Valuing assets you can exercise again and again.
- Least-squares Monte Carlo
- Spread options
- Storage valuation
- Derivatives Pricing and Computational Finance
- Monte Carlo Methods and Stochastic Volatility
- Binomial Trees, Barrier and American Options
- Index, Currency, Futures and Multi-Asset Options
Stage 05 · Specialist
Energy risk and hedging
Revenue at risk on a forward curve, and explaining a valuation to a commercial team.
- VaR on forward curves
- Hedge effectiveness
- Explaining valuations
- Quantitative Risk Management
- Market Risk
- Professional Quant Practice
- Communication, ethics and defence
Stage 06 · Proof of work
Build the projects
Each project is built on real market data and reports what the data said, including when the idea failed.
- BSGrid Battery Valuation, Dispatch and HedgingA 100 MW / 200 MWh ERCOT battery’s ten-year value fell from a USD 126.1M peak to USD 13.6M as ancillary-service prices collapsed.
- NGNatural-Gas Storage Valuation, Hedging and LNG Cargo OptionalityScored on 57 monthly offers, the average 1 Bcf storage lease was valued at USD 358k and realised USD 330k.
Practise the interview
Questions this role gets asked, each with a hint and a worked solution.
- Two plots of land are each worth USD 1m today, one earning rent and one earning nothing. Which has the higher one-year forward price?
- Price an option to swap one asset for another. Where does a spark spread or a storage spread fit this shape?
- Your Monte Carlo valuation is too noisy to compare two bids. How do you reduce its variance?
Start the SelfMadeQuant route free: a free account opens one complete topic from each primer and the interview question bank. No card needed.
Start free