Crypto Correlation Matrix

Each cell is the Pearson correlation coefficient between two assets' bar-over-bar log returns. A value of +1.00 means the assets moved in lockstep across the lookback window, -1.00 means perfectly opposite, and 0.00 means no linear relationship. Most crypto majors cluster between +0.6 and +0.9 on daily returns because the market trades as one risk-on bloc.

r_i  = ln(P_t / P_t-1)   (log return per bar)
ρ_AB = cov(r_A, r_B) / (σ_A · σ_B)

Why use log returns? Log returns are additive across bars and symmetric for moves of equal magnitude in either direction, which keeps the correlation estimate stable across volatility regimes. Linear percent returns would over-weight the largest bars.

How to read it for portfolio construction. Two assets at correlation 0.9 give you almost zero diversification benefit: a portfolio of BTC and ETH behaves a lot like BTC alone. Pairs with correlations near 0 (or negative) reduce portfolio variance. In crypto, true diversification usually requires stepping outside the asset class.

Caveats. Correlations are time-varying — a 30-day correlation and a 180-day correlation between the same pair can differ by 0.2 or more. The matrix is also sensitive to the bar interval; intraday correlations are noisier and tend to overestimate co-movement during news events. Use the matrix as a starting point, not as an input to a position-sizing rule.

Data is pulled live from Binance's public klines API on page load. No account required; nothing is sent to our servers.