Python for Algorithmic Trading Cookbook, 2nd Edition — Book Review
Gifted review copy from Packt Publishing — honest, unpaid opinion below.
Just finished reviewing Jason Strimpel's Python for Algorithmic Trading Cookbook, 2nd Edition (Packt, 2026) ahead of publication.
Strengths
This is a genuinely useful reference for anyone building trading systems in Python. The recipe format is the book's greatest strength — plenty of self-contained recipes spanning data acquisition, data wrangling, and visualization; building fundamental data screeners; factor trading; backtesting (both vector-based and event-based) and its evaluation; and using the IB API for historical and streaming data in portfolio building and management. Everything is backed by extras: clear boxes of notes and tips, and useful contextual links.
What's new in the 2nd edition
Beyond that, the second-edition additions, themselves enriched with recipes and extras, include a chapter on accelerating financial market data analysis with Polars and DuckDB, another zooming in on building AI-powered research tools with LangChain and LlamaIndex to augment the trading workflow, and a chapter of GPU-accelerated recipes for trading research with RAPIDS.
Where it falls short
It's a bit of a pity that missing-data imputation — a problem that's widely troublesome in acquired data and often requires fairly complex methods to solve — is handled rather narrowly on the hands-on side, limited in practice to simple back- or forward-filling. Beyond these, there's a far more extensive taxonomy of missingness types, and correspondingly more complex methods to handle them, that goes untouched here. Still, it's fine that the problem is at least addressed a bit more broadly and suggestively at the descriptive level.
Strategy coverage
As for the methods combined into algo trading strategies, there are effectively infinitely many of them, so it's no surprise the author couldn't cover even a fraction — but such methods, being the crux of algo trading, its spiritus movens, are here largely curbed to factor trading — with just a couple of standalone strategies stepping outside that frame. To counterbalance, and worth stressing, there is a genuinely useful, generic methodology for converting a trading strategy embedded in a research paper into code with LlamaIndex.
Backtesting
As for backtesting, deeply worth emphasizing is the essential and seemingly widely overlooked distributional view of performance metrics — achieved in the book via bootstrapping, obviously implemented in Python, and already present in the first edition too — which usefully expands on the common one-number view.
Verdict
To summarize, this book supports both paper and live algo trading through genuinely comprehensive, hands-on content, allowing for considerable upskilling. It's well structured, with well-sized sections that never feel overloaded, and it reads well — not boring, not headache-inducing, but easy to soak up — with, I'd say, a high ROL (Return on Learning) ratio.
Disclosure
I received a complimentary review copy of this book from Packt Publishing (via Shruthi Shetty) ahead of publication. The opinions above are my own.
You can also see this review on LinkedIn.
