rmtf1111/imc-prosperity-4

This repository contains the trading code and strategy write-up for a team called rat_hunters that finished second place overall in IMC Prosperity 4, an algorithmic trading competition. The team earned a cumulative simulated profit of 1,459,764 units (called XIRECs), with the bulk coming from automated algorithmic strategies. The README is a detailed post-mortem explaining how each round was approached. The core recurring insight was mean reversion, the idea that when a price strays from its expected value, it tends to snap back, making it profitable to bet on that reversal. For each fictional product in the competition, the team identified a "fair value" price and entry thresholds, then programmed the bot to buy when the price dipped significantly below that value and sell when it rose above it. Later rounds introduced more sophisticated techniques: basket arbitrage (trading related products as a group when their combined price drifted), pairs trading (exploiting the price relationship between two correlated products), and lead-lag signals (one asset's price predicts another's future movement with a time delay). The team also ran basic market making, posting buy and sell quotes simultaneously to earn the spread, on products that didn't fit other strategies. The code is written in Python. This is a competition submission and strategy analysis document rather than a general-purpose tool, but it serves as a practical example of quantitative trading strategy design and backtesting within a constrained simulated market environment. The full README is longer than what was provided.