Two thirds of an altcoin's movement owes nothing to Bitcoin. But that independence vanishes when it falls: across 67 Bybit perpetuals, altcoins amplify its falls by about half, and merely follow its rises.
The two previous articles ran into the same question. The one on relative strength asked whether ranking contracts against each other makes sense, since they all more or less follow Bitcoin. The one on confluence recalled that a level's strength is supposed to come from the number of independent reasons to watch it. Both point to a single measurement: how much of an altcoin's movement truly belongs to it, and how much is simply Bitcoin's?
The average answer is reassuring. The answer at the moments that matter is much less so.
Measured on the 67 Bybit perpetuals that, among the hundred most traded, have a full year of history. For each one, its daily returns are compared with those of BTCUSDT, from 20 October 2025 to 19 October 2026, then over the whole available history, close to 1,000 days.
The median correlation to Bitcoin is 0.57 over one year. Half the contracts sit between 0.44 and 0.64. Bitcoin therefore explains only one third of a typical altcoin's movement (33 %). Two thirds belong to the altcoin itself.
| Correlation to Bitcoin | Contracts | Share |
|---|---|---|
| Below 0.5 | 25 of 67 | 37 % |
| Below 0.6 | 39 of 67 | 58 % |
| Below 0.8 | 62 of 67 | 93 % |
Only five contracts exceed 0.8. They are the large caps, which move almost like Bitcoin: ETH (0.91), SOL (0.87), XRP (0.86), LINK and BNB (0.82). At the other end, several small contracts are barely correlated to it at all, below 0.15. Their rallies and crashes follow a story of their own.
The median beta is 1.22, and 73 % of contracts have a beta above 1. A typical altcoin moves about a fifth more than Bitcoin.
That figure is not a constant. Over the first half of the year, the median correlation was 0.65. Over the second, it falls to 0.46. A preliminary measurement in early September, on 32 contracts, gave 0.63. Only 44 % of contracts see their correlation move by less than 0.15 from one half-year to the next.
Each contract's own share is real, but it varies from one period to the next. A correlation computed over the past year describes the recent past, not a lasting property of the asset.
Correlation sums up a whole year in one number, but it mixes very different days. Sorting the days by Bitcoin's move gives a different picture:
| Bitcoin moves | Days | Contracts moving the same way |
|---|---|---|
| −5 % or worse | 24 | 92.5 % |
| −5 to −2 % | 127 | 87.2 % |
| −2 to +2 % | 686 | 67.7 % |
| +2 to +5 % | 133 | 83.8 % |
| +5 % or better | 28 | 92.1 % |
On quiet days, one altcoin in three moves against Bitcoin. That is where the own share measured above lives. On days of heavy losses, more than nine contracts in ten fall with it. This is not down to a few contracts: with equal weight, a median contract falls on 92 % of the days Bitcoin loses 5 % or more.
Altcoins' independence is therefore real, but it is concentrated in the days when little happens.
The most useful result is about magnitude, not direction. The overall beta of 1.22 hides an asymmetry: altcoins do not amplify Bitcoin the same way when it rises and when it falls. To measure it, a beta is computed for each contract on Bitcoin's down days only, then another on its up days only:
| Window | Down beta | Up beta | Contracts more sensitive to falls |
|---|---|---|---|
| Last 365 days | 1.14 | 0.97 | 70 % |
| First half of the year | 1.13 | 1.01 | 64 % |
| Second half of the year | 1.12 | 0.96 | 61 % |
| Whole history | 1.37 | 0.99 | 87 % |
| First half of history | 1.41 | 1.05 | 73 % |
| Second half of history | 1.34 | 1.06 | 82 % |
In all six windows, the down beta exceeds the up beta, and this holds for a large majority of contracts. On days of marked moves, the gap is sharper still. When Bitcoin loses at least 2 %, a median contract loses 1.47 times as much, and 86 % of contracts amplify the fall. When it gains at least 2 %, a median contract gains 1.02 times as much, and only half of them amplify the rise.
In other words, a typical altcoin takes one and a half times Bitcoin's falls, and barely all of its rises.
This result settles the question left open by the article on relative strength. With two thirds of their movement their own, ranking altcoins against each other makes sense: they are not mere copies of Bitcoin.
But a large own share is not a predictable one. The same article showed that the top of the ranking does not reliably do better afterwards. And this measurement adds a reason: on the days when everything is decided, the own share fades, and contracts fall together whatever their rank.
It covers contracts that are among the most traded today. Those that lost their liquidity or were delisted are missing, and the sample leans towards survivors. It uses daily returns: intraday, the lags between Bitcoin and altcoins can be more pronounced. Finally, correlation only measures a linear link. Two assets can be weakly correlated from one day to the next and still fall together in a crash, which is exactly what the table of big days shows.
What it does establish is enough to correct a common reading. An altcoin is independent of Bitcoin when it costs nothing, and amplifies it when it costs.
This is not investment advice. This article sets out technical-analysis concepts for educational purposes. Past performance is no guide to future performance, and trading crypto-assets carries a risk of losing the entire capital committed. Legal information.