Live prices are currently unavailable — the exchange feed could not be reached and no recent cached reading is held.

Bitcoin Data GuideBitcoin Reference

Research · Altcoins

When alts decouple from Bitcoin

There are periods when the large alts stop tracking bitcoin closely. They are real, they are usually short, and each one has a specific explanation in what was happening in the market at the time rather than in the alt's own merits.

Last reviewed 2026-09-21Source: CoinGecko historical daily prices; Coin Metrics community network dataEpisodes are described qualitatively. No correlation coefficient is asserted without its window and asset set.

Documented episodes

Decoupling is not a single event with a single cause. The episodes below are the ones that recur in the literature and in market commentary, and each is described by the condition that produced it. The dates are given as periods rather than as precise turning points, because a rolling correlation crosses a threshold gradually and the exact date depends on the window chosen.

Periods in which the large alts stopped tracking Bitcoin closely, what was happening in the market, and why the divergence did not persist.
PeriodWhat was happeningHow the alts behavedWhy it did not persist
Late 2017 to early 2018A retail-driven issuance wave: new tokens listing and trading on venues that quoted them against bitcoin and etherNewly listed assets moved on their own listing and liquidity events, which were not synchronised with bitcoin's daily movesThe issuance wave ended; the surviving assets reverted to trading as part of the same risk complex as bitcoin
2018 through 2019A prolonged decline in which bitcoin's own liquidity concentrated while smaller assets lost depthThin order books produced idiosyncratic moves, including sharp rallies that bitcoin did not shareThin markets amplify both directions; the divergence was a liquidity artefact rather than a change in relationship
2021A period of rotation in which capital moved between bitcoin and the large alts on a scale not seen beforeThe large alts outperformed bitcoin for stretches while bitcoin consolidated, producing a visible gap in relative performanceRotation is a reallocation of the same capital, so the two remain exposed to the same marginal buyer
2022A credit event in the crypto lending and exchange sector, with failures specific to particular firmsAssets with direct exposure to the failing counterparties fell further and faster than bitcoinIdiosyncratic credit exposure is asset-specific by construction; once resolved, the shared factor reasserted itself

Last reviewed 2026-09-21Source: Market-history description; CoinGecko daily series for the underlying pricesPeriods are approximate. A rolling coefficient crosses any threshold gradually, so the boundary depends on the window.

Why decoupling is usually temporary

The channels that produce co-movement do not switch off during a decoupling episode. Alt pairs are still quoted against bitcoin, the same venues still make markets in both, and the same macro factor still prices both in dollars. What changes during an episode is that a second, asset-specific force becomes large enough to dominate the shared one for a while. When that force fades, the shared channels are still there and the co-movement returns.

This is why the episodes above share a shape. Each is driven by something that is by nature temporary: an issuance wave, a liquidity drought, a rotation, a credit event. None of them is a permanent change in how the assets are priced, and none of them removes the mechanical link between an alt's dollar price and bitcoin's dollar price.

The practical consequence is that a decoupling episode is weak evidence for a durable change in relationship. A reader who observes a falling rolling correlation should ask what asset-specific force is operating, how long it can plausibly last, and whether the window is short enough that a few unusual days are driving the result. In most cases the answer is that the episode is real, explicable, and bounded.

Dataset, period, method and limitations

Dataset. Daily closing prices for bitcoin and for the largest non-bitcoin assets by market capitalisation, from CoinGecko's historical price endpoint, with Coin Metrics community data as a cross-check. The asset set is the large-cap group, not the full universe of listed tokens.

Period. The episodes span 2017 to 2022, the period over which the large alts have continuous daily quotes and a comparable market structure. Earlier periods are excluded because the venue landscape and the asset set were different enough that a comparison would not be like-for-like.

Method. A decoupling episode is identified as a stretch in which a rolling correlation of daily log returns falls below a stated threshold. The threshold and the window must both be stated for the episode boundaries to be reproducible; this page describes the episodes qualitatively and does not fix a threshold.

Limitations. Episode boundaries are sensitive to the window, so a different convention would move the dates. The descriptions of what was happening are market history rather than measured quantities, and they are not independent of the price data they are used to explain. Survivorship is a real problem: assets that failed during an episode are absent from a current large-cap set, which biases any retrospective comparison toward the survivors.

Sources and references

The episode descriptions draw on the price record and on contemporaneous market reporting; the measurement convention is documented on the companion page.

Related reading