Crypto correlation matrix: Pearson heatmap across BTC, ETH, and TradFi.

A correlation matrix is a table showing the Pearson correlation coefficient between every pair of selected assets. The coefficient ranges from +1.0 (assets move in perfect lockstep) through 0.0 (no linear relationship) to -1.0 (assets move in opposite directions). In crypto, correlations tend to be high during market stress (everything sells off together) and lower during normal conditions, making correlation monitoring a key input for portfolio risk management.

The Pearson correlation coefficient measures the linear relationship between two variables using their return series over a specified period. It is calculated as the covariance of two assets' returns divided by the product of their standard deviations. A value of +1.0 means returns move proportionally in the same direction, -1.0 means they move proportionally in opposite directions, and 0.0 means no linear relationship exists. Sharpe calculates Pearson coefficients using daily log returns.

The Correlation Matrix supports up to 10 assets simultaneously, producing an NxN grid of up to 45 unique pairwise correlations. You can mix top crypto assets with configured traditional finance assets like gold and the S&P 500. The 10-asset limit keeps the matrix readable and computationally efficient, and it still covers the most common portfolio analysis use cases.

Correlations are calculated over 30-day, 90-day, 1-year, and 3-year windows. The 30d window captures current market regime behavior and is useful for tactical risk checks. Longer windows like 1y and 3y reveal structural relationships and are more appropriate for strategic portfolio allocation decisions. Comparing short and long windows highlights when current correlations deviate from historical norms.

Including traditional finance assets like gold and the S&P 500 reveals whether crypto is currently behaving as a risk asset (high correlation with S&P 500), a store of value (high correlation with gold), or an uncorrelated alternative (low correlation with both). These cross-asset correlations shift during macro events: BTC-S&P 500 correlation spiked during the 2022 rate hiking cycle but decoupled in other periods.

Correlation coefficients are recalculated using daily closing prices, so the matrix updates once per day with each new daily close. The rolling time-series chart adds a new data point daily for each pairwise relationship. While intraday correlation changes are not captured, daily frequency provides the statistically meaningful sample sizes that Pearson correlation requires for reliable coefficient estimates.

A rolling correlation calculates the correlation coefficient over a fixed window (for example, 30 days) that moves forward one day at a time, producing a time-series of correlation values. This reveals how the relationship between two assets changes over time. Static correlations can be misleading: ETH and BTC may show 0.85 correlation over a year but experienced periods of 0.95 (lockstep) and 0.60 (decoupling). The rolling view exposes these regime shifts.

Yes. The Correlation Matrix is available free on Sharpe Terminal with no account required. All timeframes, up to 10 assets including TradFi benchmarks, rolling time-series, and the NxN heatmap are accessible immediately. The same data is available through Sharpe's REST API, MCP server, and CLI tool for programmatic access.

A cell is unavailable when one or both assets lack enough overlapping daily log-return observations for the selected window. This is common for newer tokens, recently added assets, or crypto-TradFi pairs where weekend and market-holiday calendars reduce overlap. Sharpe returns unavailable cells instead of forcing a low-confidence coefficient.

The matrix requires broad overlap across the selected lookback and uses daily log returns. If the overlap falls below the configured coverage threshold, the pair is treated as unavailable for that period. This keeps long-window correlations from being distorted by sparse or newly listed assets.

A correlation of 0.8 means the two assets' daily returns move in the same direction roughly 80% of the time with similar magnitudes, so they are strongly but not perfectly synchronized. In crypto, this level is common between major alts (ETH-SOL, BTC-ETH typically sit in the 0.7-0.9 range). A correlation of 0.8 provides limited diversification benefit; a portfolio of two assets at 0.8 correlation only modestly reduces volatility compared to holding either alone. True diversification requires correlations below 0.3, which typically means mixing crypto with TradFi assets like gold or bonds.

The BTC-ETH 30-day rolling correlation has historically sat between 0.7 and 0.95, tightening during broad-market rallies and selloffs and loosening during ETH-specific catalysts (Merge, upgrades, ETF decisions). The exact current value updates daily. Open Sharpe's Correlation Matrix or the `/correlation/bitcoin-vs-ethereum` pair page for the live number, the rolling chart, and the 1-year min/max/percentile context. In practice, readings above 0.9 indicate a macro-driven crypto regime where ETH offers little diversification versus BTC, while readings below 0.75 indicate ETH is trading on idiosyncratic flow worth a separate thesis.

Both relationships are regime-dependent. BTC-gold correlation flipped positive in 2020 to 2021 during the inflation narrative, decoupled through most of 2022, and re-correlated in 2023 to 2024 as macro-hedge demand returned. BTC-S&P 500 correlation has run higher (typically 0.4 to 0.7) since 2020 as crypto became institutionally allocated alongside risk assets. Whether BTC is behaving as a store of value (gold-like) or a risk asset (S&P-like) changes with the macro regime. Watch both the BTC-gold and BTC-S&P 500 rolling correlations in the matrix to identify which narrative is dominant right now.

A correlation regime is a sustained period during which a group of assets exhibits a characteristic correlation pattern, for example a "risk-on" regime where BTC, ETH, and the S&P 500 all move together, or a "decoupling" regime where crypto trades on idiosyncratic flow while equities move on macro. Regime changes matter because portfolio diversification assumptions break down at the regime boundary: historical correlations calibrated in one regime become unreliable in the next. Watching rolling correlations across multiple windows (30d, 90d, 1y, 3y) helps identify regime changes as they emerge, and a sharp divergence between 30d and 1y correlation is typically the first signal.

Pearson correlation on daily log returns is the standard metric for portfolio covariance, diversification, and pair-trade screening. It measures the linear relationship between two return series and maps cleanly into risk models. Crypto returns are fat-tailed, so individual outlier days still matter; that is why Sharpe exposes multiple lookback windows instead of relying on a single static coefficient.

Crypto correlations are systematically higher than stock correlations because crypto is a single asset class driven by shared liquidity, leverage, and narrative flows. BTC dominance alone often explains 60% to 80% of altcoin variance. Stock correlations within a sector (e.g., all banks) are high, but cross-sector correlations can be near zero. Crypto also has no earnings cycle, no sector diversification beyond L1/DeFi/memes/etc., and a much shorter history, so correlation estimates are noisier and more regime-sensitive. In practice, crypto portfolios concentrate risk more than equity investors assume, and genuine diversification almost always requires mixing crypto with TradFi assets like gold, bonds, or currencies.