Baseball Fun (bbfun, for short)

About

Baseball Fun is an open-source baseball simulator.

🚧 This simulator is under active development 🚧

Goals currently working on (in order)

• Simulate a simplistic baseball game (packages/utils/src/entities/game-sim/GameSim.ts is where this lives)
• Save the simulated game results
• Build a view to show the statistics/logs of a game
• Train data from the 2011 MLB season to build models for simulation (2011, because Cardinals!)
• Website to simulate 2011 MLB season and display league leaders, standings, etc.

How to run

• Install packages with bun install at the root folder
• In VSCode, run tasks by pressing Ctrl+Shift+P and typing Tasks: Run Task
• Run task dev-server to start the dev server

General Project goals

• Use the Bun runtime because it's awesome.
• Use Typescript/Javascript as much as possible
• Use R language for any data science/analysis not possible in the Javascript ecosystem
• Use a Bun workspace

Personal goals

• Learn machine learning to build models for simulation
• Make fully typed (no Typescript any!)

Libraries currently used

• biomejs for linting and formatting code
• Bun runtime to run Typescript code
• csv-string to parse CSV files
• dayjs to parse CSV files
• faker to generate fake data
• MessagePack to serialize data
• lodash for utility functions
• TensorFlowJS for machine learning
• zod for validating data

Data used

• baseballr for historical ID player mappings
• Out of the Park Baseball for player ratings
• PitchFx for pitch data

Conventions

• Camel case everything

• The utils package hosts the constants, entities, functions, and types that are shared across the project.

The file names are prefixed with an abbreviation of the type of file for easier searchability within VSCode.

• cFILENAME = constants
• eFILENAME = entities
• fFILENAME = functions
• tFILENAME = types

Current Bun limitations

Haven't been able to find a Node.js machine learning library that works in Bun. I'm currently using TensorFlowJS, but it's not ideal. Training is slow and need to save/access the models via upload and download the models to/from the server. TensorFlowJS Node is working in Bun!

Feedback

The main reason I'm creating this simulator is to learn. If you have any feedback, comments, or questions please reach out. I'm open to any suggestions and ideas.

Click the Discussions at the top of this repo to reach out.