Correlation and causation
When two prices move together, does one actually push the other, or are they both responding to something else?
In plain words
Correlation measures how closely two series move in the same direction over a period, from minus one to plus one. Causation means a change in one series produces the change in the other. High correlation alone never establishes that link.
Why it matters
Traders often treat a strong co-movement as a lead-lag signal: one asset is assumed to predict or drive the other. But two assets can move together because they share the same underlying flows, sentiment or funding conditions, so the apparent driver may carry no information at all.
In a real week: ETH, week of 28 August 2023
In the week of 28 August 2023, BTC and ETH hourly prices had a correlation of 0.96, labelled very close (0.9 or above), and both assets moved in the same direction, with ETH the larger weekly move. Yet ETH fell more: eth_minus_btc_return was -0.86 percentage points and the ETH/BTC ratio changed -0.86%. ETH's -1.36% weekly return was therefore not explained by BTC alone.
A common mistake
Reading a correlation of nearly one as proof that one asset drives the other. A shared cause can produce the same tight co-movement without any causal link between the two prices.