IMC Trading Competition Analysis

Overview

This repository contains my analysis and trading approach for the IMC Trading Competition. I focused on understanding relationships between products and developing effective statistical trading strategies across multiple rounds.

Round 2 Analysis

Products Analyzed

• Squid Ink
• Resin
• Kelp

Key Findings

I performed extensive analysis with various plots and visualizations that can be found in the notebook section of this repository. Some notable findings include:

• Discovered that Resin exhibits mean-reverting properties
• When plotting the difference between Squid Ink and Resin signals, we found a pattern that was symmetrical to the Squid Ink pattern with respect to a horizontal line
• This relationship provided key insights that informed our trading strategy for Squid Ink using Resin as a reference

Analysis Approach

The repository includes Jupyter notebooks with detailed plots tracking the day-to-day movements of Squid Ink and related analysis that guided our trading decisions.

Round 3 Analysis

Products Analyzed

• Jams
• Croissants
• Djembes
• Baskets 1
• Baskets 2

Key Findings and Strategies

For this round, I expanded my statistical analysis approach:

• Jams Analysis:

  • Applied stationarity testing (ACF test) which revealed Jams was not stationary
  • Used differentiation to transform the signal
  • Implemented EWMA (Exponentially Weighted Moving Average) models on the differentiated signal
  • Traded based on forecasts generated from these models

• Croissants Analysis:

  • Found that standard EWMA was insufficient for effective predictions
  • Developed a modified "moving alpha" approach to improve the EWMA model
  • The moving alpha represented the impact of the most recent observation (X_i) on the next price
  • Made alpha a function of trailing volatility, which significantly improved performance

• Basket Trading Strategy:

  • Applied these forecasting models to trade basket combinations
  • Specifically traded baskets consisting of 4 croissants and 2 jams
  • Used the forecasts to determine optimal trading positions

Round 4 Analysis

Products Analyzed

• Volcanic Rock
• Options on Volcanic Rock (various strike prices)

Options Trading Strategy

For the final round, I developed an options trading approach:

• Volatility Estimation:

  • Used trailing volatility as an estimate for Volcanic Rock's future volatility
  • Applied this volatility estimate to inform options pricing and trading decisions

• Directional Options Trading:

  • Implemented strategies to trade options based on directional predictions of the underlying Volcanic Rock

• Butterfly Spread Strategy:

  • Created butterfly spread positions by simultaneously buying and selling call options at different strike prices
  • This approach allowed for profiting from specific price movement scenarios while managing risk
  • Strike price selection was informed by volatility estimates and directional predictions

Learning Journey and Reflections

Knowledge Gained

I entered this competition with no prior trading experience and gained valuable insights into multiple trading strategies and concepts:

• Mean reversion trading strategies
• Momentum trading approaches
• Volatility trading techniques
• Options trading fundamentals
• Statistical models for price prediction (EWMA, ARIMA)

Areas for Improvement

Upon reflection, several aspects could have been improved:

• Risk Management: Should have implemented proper risk management tactics such as stop-loss and take-profit orders
• Hedging: Failed to adequately hedge positions during options trading, exposing the strategy to unnecessary risk
• Volatility Modeling: Could have used the Black-Scholes model to extract implied volatility rather than relying solely on trailing volatility as an estimate
• Position Sizing: Better position sizing techniques could have optimized returns while minimizing risk

This competition provided an invaluable learning experience that significantly enhanced my understanding of algorithmic trading and financial markets.