Correlation with bitcoin: your own coin, or bitcoin under another ticker

What this number answers

You picked a coin because of a good chart, a piece of news and rising volume. If its correlation with bitcoin is 0.9, none of that matters: you opened a bitcoin position, only with worse liquidity and a wider spread. The number answers exactly that question — does the instrument have a life of its own. It answers nothing else, and the second half of this lesson is about what not to expect from it.

Computed on returns, not on prices

We take the last 25 hours, fold minutes into hourly candles and compute HOURLY RETURNS — not the prices themselves but their changes. That gives up to 24 points; with fewer than 8 the number is not shown at all. The difference between prices and returns is fundamental, not technical: any two rising charts produce a high price correlation simply because both are rising. That way gold and a meme token would come out related when they have nothing to do with each other. Returns show co-movement itself.

A measurement that argues with the common claim

Altcoins correlate with bitcoin at 0.6–0.9 is a phrase repeated everywhere, and we repeated it ourselves in the risk course until we computed it on our own data. A measurement across 677 coins at one moment: the market median is 0.07, and 57.6% of coins fall within ±0.2. The common claim turned out to hold only for the largest: DOGE 0.79, XRP 0.67, SOL 0.66, ETH 0.64.

Correlation with bitcoin is a property of liquid coins

The picture is smooth and explainable. Large coins hold institutional and algorithmic money that moves the whole market at once. A small coin lives by its own pump, its own listing and its own developer — bitcoin is simply not its boss. The practical conclusion runs against habit: the smaller the coin, the less sense there is in explaining its move by bitcoin.

The limit of meaning: where the number ends and noise begins

This is the core of the lesson. A correlation over 24 points is a short sample and carries a spread of its own. A check on synthetic data: take two series of PURELY RANDOM numbers, 24 values each, and the measured correlation between them lands between −0.34 and +0.34 in 90% of cases. Not because a link exists, but because there are few points. On 168 points (a week of hourly data) the same spread shrinks to ±0.13.

Hence the practical rule for reading our column: distinguishing 0.05 from 0.15 is pointless — both sit inside the noise. A value like 0.65 differs from zero with confidence. And the band from roughly −0.35 to +0.35 on a daily window reads as no link visible rather than no link. Those are different statements, and the second does not follow from our data.

What correlation does not tell you: neither size nor cause

Correlation measures agreement of DIRECTION, not the size of the move. A coin correlating at 0.9 may travel three times as far as bitcoin: same direction, its own amplitude, and position risk does not follow from it — that is computed from the instrument's own volatility. The second silence: correlation has no direction of causation. It cannot tell the coin follows bitcoin from both follow the general mood of the market, and for small coins the latter is more often true.

The only real diversification in our table

A measurement across the other asset kinds at the same moment: 507 stocks have a median correlation with bitcoin of +0.02, 17 commodities −0.18, and 14 currency pairs −0.03. That is genuine independence — not because the numbers are small but because there is mechanics behind them: gold and EURUSD move on rates, reports and geopolitics, not on crypto sentiment. Five altcoins are one bet spread over five rows; crypto, gold, oil and a currency pair are four different causes.