Most trading books tell you what to think. This one shows you what was built.

Over two months of evenings, a CFA charterholder and an AI coding agent built a complete options trading system: the data plumbing, the strategies, the backtests, and the discipline. Every line of code is public on GitHub. Every failed strategy stayed on the record next to the one that worked.

The flagship strategy won 84% of its trades on the years it was built from, then 87% on two years of data it had never seen. This book shows you that result, and then spends its best chapters explaining why even that is not proof, and what honest proof would take.

Inside the book:

Set up US brokers, data feeds, and TradingView, and place your first paper trade
Put an AI to work writing and testing your trading code, without letting it flatter you
Run backtests that cannot lie: out-of-sample splits, walk-forward tests, lookahead traps
Read results like a professional: expectancy, drawdown, Sharpe, profit factor
Connect the AI to your broker and your chart with MCP, and watch it draw the S&P levels
Follow a two-week, day-by-day plan to a running paper-trading system of your own
What you need. You will need a coding agent like Claude Code, Codex, or any other AI coding agent — the book teaches the whole method through one, and every chapter runs on real sessions you are meant to repeat. A free broker paper account covers the rest.

Every number in this book is a backtest against stated parameters or a paper trade, the failures are shown next to the wins, and the author plans on half of what the model says. For education only. Nothing in this book is investment advice.

Tejas Jadhav, CFA, FRM, is a capital-markets professional. He built and runs the live multi-strategy system this book documents, and the full source code, twenty-nine research studies, and every correction are published openly on GitHub.