Phase 1 Foundations
Milestone: Your own research data lake.
Unit 1 · The Quant Landscape
Quant finance is engineering for decisions under uncertainty with money on the line — and the job is mostly not fooling yourself.
| Session | What you walk away with | Builds on | Status |
|---|---|---|---|
| 1.1 · What Quants Actually Do | Map the roles, the firms and how each one actually makes money — so you know which game you are training for. | — | ✅ published |
| 1.2 · Where Returns Come From | The five honest sources of edge — risk premia, behavioural anomalies, liquidity provision, information and speed — and why most apparent edges are none of them. | 1.1 | ✅ published |
| 1.3 · The Quant Research Lifecycle | The master workflow of the whole course — hypothesis to data to signal to validation to sizing to paper to live to decay — and the gates between the stages. | 1.2 | ✅ published |
| 1.4 · The Engineer's Edge, Blind Spots & Honest Odds | What transfers from software and ML, what breaks in markets, and the evidence on who actually makes money. | 1.3 | ✅ published |
| 1.5 · Set Up Your Quant Lab | A reproducible workbench — uv, Python 3.12, Jupyter, DuckDB, the quantnb package, a research repo layout and a data cache. | 1.3 | ✅ published |
Unit 2 · Markets & Instruments
Every instrument is a contract with a cash-flow profile — learn the plumbing and the P&L arithmetic before you model anything.
| Session | What you walk away with | Builds on | Status |
|---|---|---|---|
| 2.1 · Market Anatomy & the Life of an Order | Follow one order from your code through broker, venue, clearing house and custodian — and see where every fee and risk sits. | 1.1 | ✅ published |
| 2.2 · Equities, Indices & ETFs | Shares, dividends, splits, index construction, ETF creation/redemption, and the mechanics and costs of shorting. | 2.1 | ✅ published |
| 2.3 · Returns, P&L, Compounding & Leverage | The arithmetic every later session relies on — simple vs log returns, compounding, annualisation, volatility drag, Sharpe, drawdown and leverage. | 2.2 | ✅ published |
| 2.4 · Futures: Carry, Basis, Roll & Margin | How futures are priced from cost of carry, why curves slope, what roll yield really is, and how margin creates leverage and liquidation risk. | 2.3 | ✅ published |
| 2.5 · FX & Commodities | Currency quoting, forwards and covered interest parity; commodity term structures, storage and convenience yield. | 2.4 | ✅ published |
| 2.6 · Crypto Market Structure | Centralised vs decentralised venues, AMMs, perpetual futures and funding, basis trades, and the custody and counterparty risks that traditional markets hide from you. | 2.4 | ✅ published |
| 2.7 · Options Primer | Calls, puts, payoffs, put-call parity, intrinsic vs time value, and a first look at implied volatility — enough to read any options discussion. | 2.3 | ✅ published |
Unit 3 · The Quant Math Toolkit
The math is a toolkit, not a hurdle — every tool here earns its place by answering a trading question.
| Session | What you walk away with | Builds on | Status |
|---|---|---|---|
| 3.1 · Probability & Expected Value | Random variables, expectation, variance, conditioning and Bayes — the language of every bet and every interview question. | 2.3 | ✅ published |
| 3.2 · Distributions & Fat Tails | Normal, lognormal and Student-t; skewness, kurtosis and tail indices — and why the normal distribution will eventually lie to you. | 3.1 | ✅ published |
| 3.3 · Linear Algebra for Portfolios | Weights as vectors, risk as a quadratic form, PCA as the hidden structure of a market — linear algebra through the portfolio lens. | 3.1 | ✅ published |
| 3.4 · Convex Optimization | Objectives, constraints, convexity, quadratic programs and Lagrange multipliers — and how to state a portfolio problem so cvxpy can solve it. | 3.3 | ✅ published |
| 3.5 · Simulation & Monte Carlo | Simulate prices, strategies and uncertainty with seeded Monte Carlo — random walks, GBM, bootstraps and variance reduction. | 3.2 | ✅ published |
| 3.6 · Bet Sizing, Kelly & Ruin | Why the size of a bet matters as much as its edge — Kelly growth, the cost of over-betting, fractional Kelly and risk of ruin. | 3.5 | ✅ published |
Unit 4 · Financial Data Engineering
Your backtest can only be as honest as your data — point-in-time, or it didn't happen.
| Session | What you walk away with | Builds on | Status |
|---|---|---|---|
| 4.1 · The Data Landscape, Vendors & Licensing | Bars, ticks, quotes, books, fundamentals and macro — who sells what, what it costs, and what the licence lets you do with it. | 2.1 | 🟡 in progress |
| 4.2 · Point-in-Time Data & Data Biases | Look-ahead, survivorship, restatements, vintages and delistings — the data biases that manufacture fake alpha, and how to engineer them out. | 4.1 | 🟡 in progress |
| 4.3 · Corporate Actions, Security Master & Continuous Futures | Adjust prices correctly, keep identifiers straight through ticker changes, and stitch futures into continuous series without inventing returns. | 4.2, 2.4 | 🟡 in progress |
| 4.4 · Time, Calendars, As-of Joins & Bars | Timestamps and time zones, exchange calendars, as-of joins that never peek, and information-driven bars (volume, dollar, imbalance). | 4.2 | 🟡 in progress |
| 4.5 · The Research Storage Stack | Parquet, Arrow, DuckDB and polars — a fast, local, columnar research stack — plus when kdb+ and managed warehouses earn their cost. | 4.4 | 🟡 in progress |
| 4.6 · Lab: Build Your Research Data Lake | Build an idempotent, tested, point-in-time data pipeline with quality checks and lineage — the foundation for every later lab. | 4.3, 4.5 | 🟡 in progress |
Phase 2 Research Craft
Milestone: A strategy report that survives DSR/PBO.
Unit 5 · Statistics & Time Series of Returns
Returns are noisy, fat-tailed and barely predictable — good statistics tells you exactly how little you know.
| Session | What you walk away with | Builds on | Status |
|---|---|---|---|
| 5.1 · Stylized Facts of Returns | The empirical regularities every model must respect — fat tails, volatility clustering, the leverage effect, and near-zero autocorrelation. | 3.2, 4.6 | 🟡 in progress |
| 5.2 · Regression & Inference That Doesn't Lie | OLS alpha and beta, HAC standard errors, overlapping observations, the standard error of a Sharpe ratio, bootstraps and multiple testing. | 5.1 | 🟡 in progress |
| 5.3 · Bayesian Thinking: Priors, Shrinkage & Updating | Priors, posteriors and shrinkage — the machinery behind Ledoit-Wolf, Black-Litterman and Kalman filters, and the honest answer to "is my Sharpe real?" | 5.2 | 🟡 in progress |
| 5.4 · Stationarity & ARMA Models | Unit roots, ADF and KPSS, AR/MA/ARMA, ACF and PACF — and why returns are close to unpredictable while prices are not stationary. | 5.2 | 🟡 in progress |
| 5.5 · Modeling Volatility | Realized-volatility estimators from OHLC data, EWMA and GARCH forecasts — volatility is the one thing you can forecast well. | 5.4 | 🟡 in progress |
| 5.6 · Cointegration & Mean Reversion | Engle-Granger and Johansen tests, hedge ratios, spreads, the Ornstein-Uhlenbeck process as a continuous AR(1), and half-lives. | 5.4 | 🟡 in progress |
| 5.7 · Kalman Filters & Regime Models | State-space models, the Kalman filter for time-varying hedge ratios, hidden Markov regimes and structural breaks. | 5.3, 5.6 | 🟡 in progress |
Unit 6 · Backtesting & Research Discipline
A backtest is a unit test that can lie — the discipline is designing it so that it can't.
| Session | What you walk away with | Builds on | Status |
|---|---|---|---|
| 6.1 · What a Backtest Is (and Isn't) | A backtest is a historical simulation under assumptions, not a research tool — what it can prove, what it cannot, and why the gap to live trading exists. | 1.3, 5.2 | 🟡 in progress |
| 6.2 · Build a Vectorized Backtester | Signals to positions to returns in a few lines of pandas/polars — and the shift discipline that keeps the future out. | 6.1 | 🟡 in progress |
| 6.3 · Build an Event-Driven Backtester | An event loop with orders, fills and portfolio state — slower, but realistic, and the same shape as a live trading system. | 6.2 | 🟡 in progress |
| 6.4 · The Bias Hall of Fame | Live demos of every classic way a backtest lies — look-ahead, survivorship, data snooping, selection, stale and untradable prices. | 6.2, 4.2 | 🟡 in progress |
| 6.5 · Transaction Costs & Capacity | Spread, commission, slippage, market impact, borrow and funding — and how costs cap how much capital a strategy can ever run. | 6.2, 2.1 | 🟡 in progress |
| 6.6 · Validation: Walk-Forward, Purged K-Fold & CPCV | Walk-forward analysis, purging and embargo, and combinatorial purged cross-validation — train/test splits that respect time. | 6.4 | 🟡 in progress |
| 6.7 · Overfitting Statistics: PSR, DSR, MinTRL & PBO | The probabilistic and deflated Sharpe ratios, minimum track record length, and the probability of backtest overfitting — how to price in the number of things you tried. | 6.6, 5.2 | 🟡 in progress |
| 6.8 · Lab: The Research Protocol | Pre-register a hypothesis, log every trial, sweep parameters, score them with PBO and DSR, and write the go/no-go memo. | 6.7 | 🟡 in progress |
Unit 7 · Asset Pricing & Factor Investing
Factors are the vocabulary of returns — but most of the dictionary is fake.
| Session | What you walk away with | Builds on | Status |
|---|---|---|---|
| 7.1 · Efficient Markets & the Limits to Arbitrage | What market efficiency really claims, where it bends, and why mispricings can persist even when smart money sees them. | 1.2, 5.2 | ⬜ planned |
| 7.2 · CAPM, APT & Factor Regressions | From one factor to many — CAPM, APT and time-series factor regressions that separate alpha from exposure. | 7.1 | ⬜ planned |
| 7.3 · Building Factors: Sorts & Fama-MacBeth | Portfolio sorts, long-short factor construction, Fama-French methodology and Fama-MacBeth cross-sectional regressions — done without look-ahead. | 7.2, 6.4 | ⬜ planned |
| 7.4 · The Factor Zoo & the Replication Debate | Hundreds of published factors, the t > 3 hurdle, post-publication decay, and the two sides of the replication debate — risk vs mispricing. | 7.3, 6.7 | ⬜ planned |
| 7.5 · Causal Factor Investing | Causal graphs, confounders and colliders — why a factor model can be statistically significant and still causally wrong. | 7.4 | 🟡 in progress |
Unit 8 · Strategy Families
A handful of economically motivated ideas explain most systematic returns — the craft is in the implementation.
| Session | What you walk away with | Builds on | Status |
|---|---|---|---|
| 8.1 · Time-Series Momentum & Trend Following | Trend following across futures with volatility-scaled EWMA crossovers — the most robust systematic strategy family, built properly. | 6.8, 2.4, 5.5 | 🟡 in progress |
| 8.2 · Cross-Sectional Momentum & Reversal | Ranking winners against losers, short-term reversal, and the momentum crashes that make position sizing matter. | 7.3, 8.1 | ⬜ planned |
| 8.3 · Pairs Trading & Statistical Arbitrage | Distance, cointegration and PCA-residual stat arb — entries, exits, hedge ratios and why the edge has shrunk. | 5.6, 5.7, 6.5 | ⬜ planned |
| 8.4 · Carry Across Asset Classes | FX carry, futures roll yield, bond carry and roll-down — one unifying definition of carry and its crash risk. | 2.4, 2.5, 8.1 | ⬜ planned |
| 8.5 · Value, Quality & Fundamental Quant | Point-in-time fundamentals from EDGAR XBRL, value and quality signals, accruals — and why value has been so painful. | 7.3, 4.2 | ⬜ planned |
| 8.6 · Event-Driven Strategies | Event studies, post-earnings-announcement drift, index rebalances — trading around information with precise timestamps. | 8.5, 4.4 | ⬜ planned |
| 8.7 · Combining Signals | Information coefficient, the fundamental law of active management, forecast scaling and combination — many weak signals into one strong book. | 8.1, 8.2, 8.3, 8.4 | ⬜ planned |
Phase 3 Portfolio, Risk & Derivatives
Milestone: A risk-managed multi-strategy portfolio.
Unit 9 · Portfolio Construction & Risk
Returns are forecast, risk is measured; the portfolio is where the two meet — under costs and constraints.
| Session | What you walk away with | Builds on | Status |
|---|---|---|---|
| 9.1 · Mean-Variance & Why It Breaks | Markowitz, the efficient frontier and the tangency portfolio — and why naive mean-variance is an error-maximiser. | 3.4, 8.7 | ⬜ planned |
| 9.2 · Covariance Estimation | Sample covariance and its failure modes, Ledoit-Wolf and OAS shrinkage, random-matrix denoising and EWMA covariance. | 9.1, 5.3 | ⬜ planned |
| 9.3 · Risk Parity & HRP | Risk contributions, inverse-vol and equal-risk-contribution portfolios, hierarchical risk parity and HERC — allocation without return forecasts. | 9.2 | ⬜ planned |
| 9.4 · Factor Risk Models | Build a mini Barra-style model — exposures, factor covariance and specific risk — and use it to decompose where your risk comes from. | 9.2, 7.3 | ⬜ planned |
| 9.5 · Sizing, Vol Targeting & Allocating Across Strategies | Carver-style volatility targeting, position buffering, diversification multipliers and allocating capital across strategies with fractional Kelly. | 3.6, 8.7 | ⬜ planned |
| 9.6 · Optimization with Constraints & Costs | Real portfolio optimisation — turnover penalties, cost-aware objectives, leverage and exposure constraints, and Black-Litterman views. | 9.4, 6.5 | ⬜ planned |
| 9.7 · VaR, ES, EVT & VaR Backtests | Value at risk and expected shortfall — historical, parametric and filtered — extreme value theory for the tail, and testing whether your VaR is honest. | 5.5, 3.2 | ⬜ planned |
| 9.8 · Drawdowns, Copulas, Tail Risk & Stress Tests | Drawdown statistics, tail dependence and copulas, and stress tests from 2008, March 2020 and 2022 — with LTCM and the 2007 quant quake as cautionary tales. | 9.7 | ⬜ planned |
| 9.9 · Performance, Attribution & When to Switch Off | Measure performance honestly, attribute it to factors and skill, monitor live decay, and decide in advance when a strategy is retired. | 9.4, 6.7 | ⬜ planned |
Unit 10 · Derivatives & Volatility
Price by replication, hedge by the Greeks, and trade the gap between implied and realized.
| Session | What you walk away with | Builds on | Status |
|---|---|---|---|
| 10.1 · No-Arbitrage & Binomial Trees | Replication and the law of one price, risk-neutral probabilities, and pricing European and American options on a binomial tree. | 2.7, 3.1 | ⬜ planned |
| 10.2 · Stochastic Calculus for Engineers | Brownian motion, quadratic variation, SDEs and Itô's lemma — built by simulation first, formalism second. | 10.1, 3.5 | ⬜ planned |
| 10.3 · Black-Scholes-Merton | Derive BSM two ways (PDE and risk-neutral expectation), implement it, and list precisely which assumptions markets violate. | 10.2 | ⬜ planned |
| 10.4 · The Greeks & Hedging P&L | Delta, gamma, vega and theta; delta hedging in simulation; and the P&L identity that links hedging to realized minus implied variance. | 10.3 | ⬜ planned |
| 10.5 · The Implied Volatility Surface | Smiles, skews and term structures; fitting SVI and SSVI; and checking the surface for static arbitrage. | 10.4 | ⬜ planned |
| 10.6 · Stochastic Volatility & Numerical Methods | Local vol vs stochastic vol, the Heston model, Monte Carlo and finite-difference pricing, calibration — with QuantLib as the reference. | 10.5 | ⬜ planned |
| 10.7 · Trading Volatility | The volatility risk premium, straddles and variance swaps, dispersion, the VIX complex — and the crash risk that pays for all of it. | 10.4, 10.5 | ⬜ planned |
Unit 11 · Rates, Credit & Macro
Money has a price that changes with time and with trust — curves encode both.
| Session | What you walk away with | Builds on | Status |
|---|---|---|---|
| 11.1 · Yield Curves & Bootstrapping | Discount factors, zero, forward and par rates, day counts, and bootstrapping a curve from market instruments. | 2.3, 3.4 | ⬜ planned |
| 11.2 · Duration, Convexity & Curve PCA | DV01, duration, convexity and key-rate risk — and the three PCA factors (level, slope, curvature) that drive almost every curve. | 11.1, 3.3 | ⬜ planned |
| 11.3 · Short-Rate Models & Swaps | Interest-rate swaps and SOFR, Vasicek, CIR and Hull-White short-rate models, and what a swaption is. | 11.2, 10.2 | ⬜ planned |
| 11.4 · Credit: Spreads, CDS & Merton | Credit spreads, default intensities and recovery, credit default swaps, and the Merton structural model of default. | 11.1, 10.3 | ⬜ planned |
| 11.5 · FX & Macro Regimes | Uncovered interest parity and the forward-premium puzzle, nowcasting with point-in-time macro data, and growth-inflation regimes across assets. | 2.5, 5.7, 8.4 | ⬜ planned |
Phase 4 Machines & Markets
Milestone: A paper-trading system running.
Unit 12 · Machine Learning for Trading
In markets, machine learning is a low signal-to-noise game where validation — not model capacity — decides who wins.
| Session | What you walk away with | Builds on | Status |
|---|---|---|---|
| 12.1 · Why Finance ML Is Different | Non-IID samples, concept drift, adversarial markets and tiny signal-to-noise — why ML habits from vision and NLP backfire here. | 6.6, 1.4 | ⬜ planned |
| 12.2 · Labeling: Triple Barrier & Meta-Labeling | Fixed-horizon labels and their flaws, the triple-barrier method, and meta-labeling to size bets from a primary model. | 12.1 | ⬜ planned |
| 12.3 · Features & Sample Weights | Stationary features that keep memory (fractional differentiation), sample uniqueness, the sequential bootstrap and time-decay weights. | 12.2, 5.4 | ⬜ planned |
| 12.4 · Gradient Boosting for Cross-Sectional Alpha | Predict cross-sectional returns with LightGBM on point-in-time features, validated with purged CV and scored on IC after costs. | 12.3, 8.7, 6.7 | ⬜ planned |
| 12.5 · Feature Importance & Interpretability | MDI, MDA, single-feature importance and SHAP — and clustered importance to survive correlated features. | 12.4 | ⬜ planned |
| 12.6 · Deep Learning, RL & the Complexity Debate | Sequence models, reinforcement learning and the "virtue of complexity" debate — where deep models help, and where they are hype. | 12.4 | ⬜ planned |
Unit 13 · NLP, LLMs & Alternative Data
Text and alternative data are new sources of information — and new ways to leak the future into your backtest.
| Session | What you walk away with | Builds on | Status |
|---|---|---|---|
| 13.1 · Alternative Data | A taxonomy of alternative data, how to evaluate a dataset before buying it, and the legal lines (MNPI, scraping, privacy) you must not cross. | 4.2, 6.7 | ⬜ planned |
| 13.2 · Text as Data | From dictionaries (Loughran-McDonald) to TF-IDF to embeddings — turning filings, calls and news into point-in-time features. | 13.1 | ⬜ planned |
| 13.3 · LLMs for Return Prediction — and the Look-Ahead Trap | What the evidence says about LLM signals, and how training-data leakage fakes their performance — with a lab that tests only after the model's cutoff. | 13.2 | ⬜ planned |
| 13.4 · LLM Agents as Research Assistants | Using LLM agents to speed up research without letting them inflate your trial count — guardrails, logging and mandatory deflation. | 13.3, 6.8 | ⬜ planned |
Unit 14 · Microstructure, Execution & Market Making
Every trade has a counterparty who may know more than you — the order book is where you pay for that.
| Session | What you walk away with | Builds on | Status |
|---|---|---|---|
| 14.1 · The Limit Order Book | Price-time priority, queues, tick size, maker-taker fees and auctions — how a modern electronic order book actually matches trades. | 2.1, 6.5 | ⬜ planned |
| 14.2 · Microstructure Models | The Roll, Kyle and Glosten-Milgrom models — spreads as the price of adverse selection, and how informed traders move prices. | 14.1, 3.1 | ⬜ planned |
| 14.3 · Market Impact & the Square-Root Law | Temporary vs permanent impact, the square-root law and its evidence, and measuring impact from your own fills. | 14.2 | ⬜ planned |
| 14.4 · Optimal Execution & TCA | TWAP, VWAP and implementation shortfall; the Almgren-Chriss trade-off; and transaction cost analysis against the arrival price. | 14.3, 3.4 | ⬜ planned |
| 14.5 · Market Making | Inventory risk and adverse selection, the Avellaneda-Stoikov model, and a queue-aware market-making backtest in hftbacktest. | 14.2, 14.4 | ⬜ planned |
| 14.6 · Latency & Performance Engineering | Profiling Python, numba and Rust/C++ hot paths, tick-to-trade latency — and an honest map of what speed buys a non-HFT trader. | 14.1 | ⬜ planned |
Unit 15 · Production Trading Systems
A live trading system is a distributed system that loses money when it is wrong — build it like one.
| Session | What you walk away with | Builds on | Status |
|---|---|---|---|
| 15.1 · Architecture of a Live Trading System | Feed, strategy, portfolio, OMS, EMS and broker — the components, their contracts, and the state machine at the heart of an order. | 6.3, 2.1 | ⬜ planned |
| 15.2 · Backtest-to-Live with NautilusTrader | One strategy class that runs in a backtest, on a simulated exchange and against a paper account — with NautilusTrader. | 15.1 | ⬜ planned |
| 15.3 · Broker APIs & FIX | alpaca-py and ib_async against paper accounts, the FIX protocol's core messages, rate limits and idempotent order submission. | 15.1 | ⬜ planned |
| 15.4 · Pre-Trade Risk & Kill Switches | Fat-finger checks, price collars, position and loss limits, and the kill switch — with Knight Capital as the case study. | 15.1 | ⬜ planned |
| 15.5 · Operations & MLOps | Reconciliation, monitoring, alerting, runbooks and restarts — plus model registries, retraining and drift detection for live models. | 15.4, 9.9 | ⬜ planned |
| 15.6 · Market Conduct & Compliance | Insider trading and MNPI, spoofing and layering, wash trading, front-running and best execution — the rules under SEC/FINRA and MiFID II/MAR. | 13.1, 14.1 | ⬜ planned |
| 15.7 · Lab: Deploy a Paper-Trading Bot | Deploy an end-to-end paper-trading system — scheduled, monitored, reconciled and protected by a kill switch. | 15.2, 15.3, 15.4, 15.5 | ⬜ planned |
Phase 5 The Professional
Milestone: Capstone + interview-ready.
Unit 16 · Career, Interviews & Your Own Book
Skill compounds when it is visible — build the evidence, rehearse the interviews, respect the odds.
| Session | What you walk away with | Builds on | Status |
|---|---|---|---|
| 16.1 · Career Paths & the Quant Job Market | Roles, firms, pay structures and hiring pipelines in 2026 — and the portfolio of evidence that gets an engineer hired. | 1.1 | ⬜ planned |
| 16.2 · Interview Drills: Probability, Mental Math & Market-Making Games | The probability, expected-value, mental-math and market-making-game rounds — with drills and the reasoning interviewers want to hear. | 3.1, 14.2 | ⬜ planned |
| 16.3 · Stats, ML, Coding & Research-Case Interviews | The statistics, ML and coding rounds, trading-system design, and the multi-hour research case — and how to write the report. | 16.2, 12.4 | ⬜ planned |
| 16.4 · Running Your Own Capital | Capital, costs, brokers, taxes and rules by jurisdiction, psychology and discipline, and how to build a verifiable track record. | 9.5, 15.7 | ⬜ planned |
| 16.5 · Capstone: Brief & Rubric | The capstone — research, validate, size, paper-trade and report one strategy end to end — with the rubric professionals would use. | 16.4 | ⬜ planned |
Educational material only — not investment, legal or tax advice. Backtests and simulations are hypothetical. Disclaimer · Text CC BY 4.0 · Code MIT.