Crypto correlation matrix
How closely the major crypto assets actually moved together, from real daily closes through 2026-08-05. If you hold six coins and they all sit above 0.8, you hold one position in six wrappers.
The short version
Over the last year these 10 assets correlated at an average of 0.80. The closest pair was ETH and SOL at 0.89, and the least alike were BTC and DOT at 0.67. A pair at 0.90 gives you almost no diversification; a pair at 0.40 gives you real separation β though see the crash caveat below before relying on it.
Last year
How these assets are behaving now. Average pairwise correlation 0.80 across 10 assets.
| Asset | BTC | ETH | XRP | BNB | SOL | DOGE | ADA | LINK | LTC | DOT |
|---|---|---|---|---|---|---|---|---|---|---|
| BTC | β | 0.88 | 0.85 | 0.76 | 0.85 | 0.77 | 0.79 | 0.80 | 0.73 | 0.67 |
| ETH | 0.88 | β | 0.83 | 0.78 | 0.89 | 0.84 | 0.84 | 0.88 | 0.75 | 0.74 |
| XRP | 0.85 | 0.83 | β | 0.72 | 0.83 | 0.84 | 0.83 | 0.83 | 0.74 | 0.74 |
| BNB | 0.76 | 0.78 | 0.72 | β | 0.78 | 0.73 | 0.74 | 0.75 | 0.70 | 0.67 |
| SOL | 0.85 | 0.89 | 0.83 | 0.78 | β | 0.84 | 0.84 | 0.87 | 0.79 | 0.77 |
| DOGE | 0.77 | 0.84 | 0.84 | 0.73 | 0.84 | β | 0.87 | 0.86 | 0.80 | 0.81 |
| ADA | 0.79 | 0.84 | 0.83 | 0.74 | 0.84 | 0.87 | β | 0.86 | 0.78 | 0.84 |
| LINK | 0.80 | 0.88 | 0.83 | 0.75 | 0.87 | 0.86 | 0.86 | β | 0.78 | 0.81 |
| LTC | 0.73 | 0.75 | 0.74 | 0.70 | 0.79 | 0.80 | 0.78 | 0.78 | β | 0.81 |
| DOT | 0.67 | 0.74 | 0.74 | 0.67 | 0.77 | 0.81 | 0.84 | 0.81 | 0.81 | β |
Last 3 years
Long enough to cover a full swing in sentiment. Average pairwise correlation 0.66 across 10 assets.
| Asset | BTC | ETH | XRP | BNB | SOL | DOGE | ADA | LINK | LTC | DOT |
|---|---|---|---|---|---|---|---|---|---|---|
| BTC | β | 0.81 | 0.63 | 0.66 | 0.74 | 0.76 | 0.68 | 0.68 | 0.61 | 0.65 |
| ETH | 0.81 | β | 0.63 | 0.67 | 0.72 | 0.74 | 0.71 | 0.75 | 0.66 | 0.71 |
| XRP | 0.63 | 0.63 | β | 0.48 | 0.59 | 0.62 | 0.74 | 0.65 | 0.65 | 0.64 |
| BNB | 0.66 | 0.67 | 0.48 | β | 0.60 | 0.59 | 0.56 | 0.58 | 0.55 | 0.60 |
| SOL | 0.74 | 0.72 | 0.59 | 0.60 | β | 0.68 | 0.69 | 0.69 | 0.57 | 0.68 |
| DOGE | 0.76 | 0.74 | 0.62 | 0.59 | 0.68 | β | 0.71 | 0.67 | 0.64 | 0.72 |
| ADA | 0.68 | 0.71 | 0.74 | 0.56 | 0.69 | 0.71 | β | 0.72 | 0.62 | 0.75 |
| LINK | 0.68 | 0.75 | 0.65 | 0.58 | 0.69 | 0.67 | 0.72 | β | 0.66 | 0.74 |
| LTC | 0.61 | 0.66 | 0.65 | 0.55 | 0.57 | 0.64 | 0.62 | 0.66 | β | 0.71 |
| DOT | 0.65 | 0.71 | 0.64 | 0.60 | 0.68 | 0.72 | 0.75 | 0.74 | 0.71 | β |
Last 5 years
A whole cycle, for the assets old enough to have one. Average pairwise correlation 0.68 across 10 assets.
| Asset | BTC | ETH | XRP | BNB | SOL | DOGE | ADA | LINK | LTC | DOT |
|---|---|---|---|---|---|---|---|---|---|---|
| BTC | β | 0.84 | 0.62 | 0.72 | 0.70 | 0.70 | 0.70 | 0.70 | 0.69 | 0.70 |
| ETH | 0.84 | β | 0.63 | 0.74 | 0.71 | 0.71 | 0.72 | 0.76 | 0.72 | 0.75 |
| XRP | 0.62 | 0.63 | β | 0.54 | 0.57 | 0.58 | 0.71 | 0.63 | 0.62 | 0.63 |
| BNB | 0.72 | 0.74 | 0.54 | β | 0.63 | 0.61 | 0.62 | 0.65 | 0.64 | 0.68 |
| SOL | 0.70 | 0.71 | 0.57 | 0.63 | β | 0.60 | 0.67 | 0.66 | 0.60 | 0.68 |
| DOGE | 0.70 | 0.71 | 0.58 | 0.61 | 0.60 | β | 0.67 | 0.64 | 0.63 | 0.67 |
| ADA | 0.70 | 0.72 | 0.71 | 0.62 | 0.67 | 0.67 | β | 0.72 | 0.66 | 0.75 |
| LINK | 0.70 | 0.76 | 0.63 | 0.65 | 0.66 | 0.64 | 0.72 | β | 0.69 | 0.77 |
| LTC | 0.69 | 0.72 | 0.62 | 0.64 | 0.60 | 0.63 | 0.66 | 0.69 | β | 0.72 |
| DOT | 0.70 | 0.75 | 0.63 | 0.68 | 0.68 | 0.67 | 0.75 | 0.77 | 0.72 | β |
Build your own
5 of 10 pickedHow to use this
Count your real positions, not your tickers
A portfolio of eight assets that all correlate above 0.8 behaves roughly like one or two independent positions. That is not automatically wrong β plenty of people are deliberately long the whole asset class β but it is worth knowing, because the risk is being taken whether or not it was chosen. The portfolio analyzer turns this into the number that matters: how much calmer the combination actually was than its parts.
Low correlation is a fair-weather property
Correlations rise in a sell-off. Two assets that spent a quiet year at 0.4 will often spend a bad week at 0.9, because in a liquidation everything is sold to raise the same dollars. Any diversification you are counting on is weakest exactly when it is needed, so size positions for the crash correlation rather than the calm one.
Stablecoins and cash are the honest diversifier
Nothing in this table is a hedge against crypto β every asset on it is crypto. If the goal is to reduce the swing rather than to change which coin drives it, the only lever with a reliable effect is holding less of the whole thing. That is unglamorous, and it works where a rearranged altcoin mix does not.
Frequently asked questions
οΌWhat counts as a high correlation?
Above roughly 0.8 the two assets are close to interchangeable for risk purposes β holding both is holding one position in a larger size. Between 0.5 and 0.8 there is some separation but not much. Below 0.3 you have genuinely different exposures. Most large-cap crypto pairs sit in the 0.6 to 0.9 band, which is why crypto portfolios are usually far less diversified than the number of tickers in them suggests.
οΌWhy is it computed on returns rather than prices?
Because two assets in the same multi-year uptrend have correlated price levels almost by definition, which makes a correlation quoted off price charts close to meaningless. What a holder needs to know is whether the two fall on the same days, and that is a question about daily returns.
οΌWhy do some assets disappear from the longer windows?
Because they are younger than the window. Measuring a two-year-old asset over a five-year table would put a different period in the same grid, and the pair that then looked least correlated would just be the one measured over different years. Anything dropped is named under the table.
οΌDo these numbers hold in a crash?
No, and this is the most important caveat on the page. Correlations rise sharply in a sell-off: assets that drift apart in calm markets tend to fall together in a bad week, which is exactly when the diversification was supposed to help. Treat a low correlation as a fair-weather property, not a guarantee.
οΌHow often is this updated?
It is recomputed from the price snapshot every time the site is rebuilt, and that snapshot currently runs to 2026-08-05. The date on each table is the real end of the data, not the date you happen to be reading it.