How alts behave when Bitcoin rallies

The pattern is often described as high-beta assets outperforming in an advance. That is a statement about sensitivity to a shared factor, and it carries a symmetric implication that is usually left out: an asset with a higher sensitivity to the factor also falls further when the factor moves the other way. The companion page on declines describes the same mechanism from the other side.

It is important not to read the dispersion as a ranking of assets. The observation is that outcomes spread out, not that the assets with the largest moves are better. A wide spread means the outcome is uncertain in both directions, and the same thin liquidity that produces a large advance produces a large decline. The dispersion is a measure of uncertainty, not of quality.

There is also a selection problem in how the pattern is usually reported. An advance is identified after the fact, and the assets that rose most are the ones that get discussed. Assets that failed during the same period are absent from the current large-cap set, so a retrospective account of "what alts did in the rally" is drawn from the survivors. The dispersion among all assets that existed at the time was wider than the dispersion among those still listed today.

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. Aggregated closes across venues.

Period. The advances in the record from 2017 onward, the period over which the large alts have continuous daily quotes and a comparable venue landscape. Earlier advances are excluded because the alt market was too small and too differently structured.

Method. An advance is defined by a trough-to-peak move in bitcoin over a stated window, and the alts' behaviour is described over the same dates. The comparison uses a common window rather than each asset's own extremes, which understates the dispersion rather than overstating it.

Limitations. Survivorship is the dominant limitation: assets that failed are absent from a current large-cap set, so the measured dispersion is narrower than the dispersion actually experienced. The asset set changes over time, so a long comparison is computed across a changing basket. And the tier boundaries used above are descriptive rather than fixed, so a different cut would produce a different grouping without changing the underlying pattern.