Getting Started

Quick Setup

1. Install uv if you haven't already:

   curl -LsSf https://astral.sh/uv/install.sh | sh

2. Clone and sync dependencies:

   git clone <repository>
   cd <project>
   uv sync

3. Activate the virtual environment:

   # macOS/Linux
   source .venv/bin/activate
   
   # Windows
   .venv\Scripts\activate

That's it—uv sync creates the venv automatically and installs all dependencies from pyproject.toml and uv.lock.

Fix your system path Alex: C:\Users\alexn\AppData\Local\Python\pythoncore-3.14-64\Scripts\uv.exe sync

Project structure

• algorithms.py — sample strategy library and utilities.
• stat_utils.py — statistics helpers used by analyzers.
• backtests/ — stores log files from the official IMC backtester and logs from prosperity4btest.
• montecarlo_backtester/ — modular Monte Carlo backtesting framework for Prosperity 4.
• rounds — stores all code and data for each round
  • 0-tutorial — example for how all future rounds should be setup
    • data — contains Prosperity 4 data packages
    • traders — has python scripts for backtesting and submission
    • explore.ipynb — jupiter notebook to call helper functions from stats and algorithms to visualise and manipulate data for that round

Built-in backtesters and visualisers

1. log_visualiser

   • Visualises official Prosperity 4 logs and log files from prosperity4btest prosperity4btest a fork of jmerle's backtester for Prosperity 3
   • To use just run log_visualiser/dashapp.py

2. montecarlo_backtester/imc-prosperity-4

   • Uses 1st Place USA Chris Roberts backtester
   • Full Monte Carlo backtester + data model (prosperity4mcbt/) for stress-testing strategies.
   • Can implement models for assets for new rounds or clone from original repo if Cris Roberts keep updating
   • Videos on the topic
     • Backtester: https://www.youtube.com/watch?v=Mi-vVCZ0Vo4
     • Prosperity Breakdown: https://www.youtube.com/watch?v=PI2lJ063sJ8

Monte Carlo quick start (Rust + npm)

1. Install Rust toolchain (Cargo):

   • Windows: winget install --id Rustlang.Rustup -e --source winget or https://www.rust-lang.org/tools/install
   • macOS/Linux: curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh

2. Install Node/npm:

   • Windows: winget install --id OpenJS.NodeJS or https://nodejs.org/
   • macOS/Linux: brew install node or package manager of choice

3. Run the Monte Carlo runner with a trader sample:

   prosperity4mcbt example_trader.py --quick --vis --out /tmp/mc_run/dashboard.json

4. Read through backtester README for more info