chadwickpy

Retrosheet play-by-play files to data tables. Pure Python, one pip install.

Retrosheet publishes every play of every MLB game as text files. The Chadwick tools turn those files into tables. chadwickpy is those same six tools, rewritten in pure Python: no compiler, no C library, no dependencies. Output is byte-identical to Chadwick on every season from 1910 to 2025 (for inputs where the C behaves in a defined way).

pip install chadwickpy            # or: uvx --from chadwickpy cwevent -h
curl -O https://www.retrosheet.org/events/2010seve.zip
unzip -q 2010seve.zip -d retro2010 && cd retro2010
cwevent -y 2010 -n -f 0,2,3,4,10,14,29,34 2010NYA.EVA > yankees_events.csv

"GAME_ID","INN_CT","BAT_HOME_ID","OUTS_CT","BAT_ID","PIT_ID","EVENT_TX","EVENT_CD"
"NYA201004130",1,0,0,"aybae001","petta001","S9/F9S-",20
"NYA201004130",1,0,0,"abreb001","petta001","K",3
"NYA201004130",1,0,1,"huntt001","petta001","8/F8LXD",2

The six tools

Same command names and options as Chadwick. Also usable from Python (from chadwickpy.tools.events import event_rows).

Good to know

• Speed: Single-core processing is ~0.66 s per team-season. When processing multiple files (such as a full season), chadwickpy automatically parallelizes across available CPU cores, processing an entire 2,430-game season in ~2.3 seconds (~80,000 plays/s), matching native C throughput.
• Override parallelism anytime via -j <workers> (e.g. -j 1 for sequential) or the CHADWICK_JOBS environment variable.
• It reads Retrosheet's files; it does not include or download them.
• Pure Python (zero dependencies). Python 3.11 or newer (also fully compatible with PyPy 3.10+ and Python 3.13+ JIT). Linux, macOS and Windows.

Documentation

https://cbwinslow.github.io/chadwickpy/: getting started, a guide to each tool, Python usage, recipes, the full option and field reference, how it was verified, and an FAQ. Machine-readable: llms.txt.

Verification

tests/ compares every tool with captured real-Chadwick output, and with the real programs when they are available (CHADWICK_BIN, CHADWICK_SRC). CI builds Chadwick at commit c685ab5 and runs the whole suite on Python 3.11-3.13, failing on any skipped parity test. Run it yourself: uv run --with pytest pytest.

Licence and credit

GPL-3.0-or-later (LICENSE). chadwickpy is a derivative work of Chadwick, Copyright (C) 2002-2023 Dr T L Turocy and the Chadwick Baseball Bureau, GPL-2.0-or-later (COPYING-chadwick, NOTICE); each ported module keeps that notice. It is independent: Chadwick and Retrosheet do not endorse or sponsor it, and it ships no Retrosheet data.

Retrosheet's data-use notice asks that anyone using its data say so. If you publish anything built on it, include:

The information used here was obtained free of charge from and is copyrighted by Retrosheet. Interested parties may contact Retrosheet at "www.retrosheet.org".

Contributing, questions, security

CONTRIBUTING.md · Discussions · Issues · SECURITY.md. The rule of the project: the port stays a translation of Chadwick's C; rules come from the C, never from tuning output.