Tool_premodel

Polymarket 15-Minute Binary Option Pricing Model

Reverse engineering market maker behavior in Polymarket's ultra-short-term cryptocurrency prediction markets.

Project Summary

This project analyzes 15-minute binary option markets on Polymarket (e.g., "Will Bitcoin be above $95,000 in 15 minutes?"). Through automated data pipelines and financial mathematics, we reverse-engineer market maker pricing logic.

Core Research Questions

How do these three variables non-linearly determine Yes/No prices?

1. Time to Expiry (Theta): Time decay effects on price
2. Distance to Strike (Delta/Gamma): Price sensitivity to underlying movement
3. Volatility (Vega): Pricing premium during market panic or calm

Key Scenarios

• Linear Decay Zone: ~3 min remaining, small price dip -> minor correction (-3¢)
• Lock-in Effect Zone: ~1 min remaining, price far from strike -> minimal movement (-1¢)
• Extreme Reversal Zone (Gamma Risk): Near expiry, price crosses strike -> collapse (-60¢)

Technical Roadmap

Data Engineering (Python)

• Exchange API integration (Binance/Bybit) for high-frequency spot prices
• Polymarket API for order book and trade data
• Time-series alignment at second-level precision

Mathematical Modeling

• Binary options pricing model
• Implied volatility calculation
• Greeks sensitivity analysis (Delta, Gamma, Theta, Vega)

Visualization

• 3D pricing surface: Time × Price × Contract Value

Expected Deliverables

1. Automated Python data cleaning pipeline
2. Pricing model: Input (current price, time remaining) -> Output (theoretical Yes price)
3. Analysis report on market maker efficiency and arbitrage opportunities

Quick Start

# Read CLAUDE.md first for development rules
# All source code goes in src/
python src/main.py

Project Structure

Tool_premodel/
├── CLAUDE.md          # Development rules for Claude Code
├── README.md          # This file
├── .gitignore         # Git ignore patterns
├── src/               # Source code
├── tests/             # Test files
├── docs/              # Documentation
└── output/            # Generated output files

Development Guidelines

• Always search before creating new files
• Extend existing functionality rather than duplicating
• Commit after each completed task
• Push to GitHub for backup