About

CorrGrid provides access to ~100 billion precomputed, survivorship bias-free, sophisticated pairwise correlation and cointegration metrics, powering CorrGrid’s cutting-edge selection tools, which identify pairs, triplets, or quadruplets of stocks, revealing hidden market topology and providing associated hedge ratios, betas, and weights.

CorrGrid is a service designed for portfolio managers, quantitative analysts (quants), traders, and analysts seeking insights for pairs trading, statistical arbitrage, correlation and cointegration analysis, mean-reversion strategies, and exploring hidden market structure via topology analysis.

• Top and Bottom Correlations: The 10,000 most positively and 10,000 most negatively correlated pairs exhibiting strong mean-reversion properties and sufficient¹ residual volatility for real-world strategy implementation, selected and filtered² from over 10 million pairs.
• Kendall’s tau: Robust correlation measure converted to the Pearson scale for improved interpretability.
• Betas & Hedge ratios: Calculated and provided for all pairs. These can be used to construct market-neutral portfolios, implement mean-reversion strategies, and for hedging.

• Johansen’s method: The Johansen’s eigenvalue method offers a way to quantify the strength of cointegration.
• Top cointegrations: The 20,000 most cointegrated pairs, triplets, and quadruplets, selected from over 10 million computations for a specific date, based on both Johansen’s method and the ADF test.
• Betas & Hedge ratios: Calculated and provided for all cointegrated pairs to facilitate the construction of mean-reversion and spread strategies.
• Spurious Cointegration Filtering: Leverages the revealed hidden topology (MST) to prioritize cointegration between variously connected instruments (e.g., same-sector, same-cluster, etc.).
  
  
  

Provides deep insights into the structure of financial instrument correlations and market topology.

• MST (Minimum Spanning Tree): A connectivity tree constructed using Kruskal’s algorithm.
• Variants: Based on raw, ex-SPY, or ex-ETF cointegrations and correlations .

• Bias-Free: Correlation and cointegration data are free of survivorship bias (including delisted stocks), enabling historical interdependency analysis for specific past dates.
• Mid-Price: Utilizes quote prices (mid-prices) rather than transaction prices, avoiding side effects such as bid-ask bounce and delayed trade prices.
• Each product is available in three variants:
        - raw
        - ex-SPY, which excludes broad market movements using SPY as a proxy
        - ex-ETF, which excludes sector-wide movements using the largest sector ETFs as proxies.

By removing these common external factors, the ex-SPY and ex-ETF variants reveal hidden interdependencies.