Competitions

Discover novel alpha signals from obfuscated market data. Participants must build predictive models that generate uncorrelated returns across multiple market regimes. The dataset spans 5 years of multi-asset returns with engineered features — column names and asset identities are hidden to prevent data leakage.

Predict realized volatility for a basket of obfuscated instruments over 1-day, 5-day, and 21-day horizons. This competition tests your ability to capture volatility clustering, mean reversion, and regime shifts using engineered features derived from high-frequency data.

Design an optimal portfolio allocation strategy that maximizes risk-adjusted returns while adhering to realistic constraints including transaction costs, position limits, and turnover restrictions. The challenge uses a universe of 50 obfuscated assets with 10 years of daily data.

Build a cross-sectional momentum factor that predicts next-month relative asset performance. The dataset includes 200 obfuscated assets with 15 years of returns and fundamental signals. This classic quant challenge rewards robust factor design and risk management.

Classify market regimes (bull, bear, high-vol, low-vol, crisis) from obfuscated multi-asset data. This competition focuses on unsupervised and semi-supervised learning approaches to regime identification, with scoring based on out-of-sample regime prediction accuracy and transition detection.