Bitcoin Rainbow Chart

How to use it (and how not to)

The rainbow chart is best understood as a slow-moving context gauge. It will not tell you anything about next week. What it does tell you is whether price is historically stretched or historically compressed relative to its own long-run path.

• Lower bands have historically coincided with periods where drawdowns were deep and sentiment was at its worst — check that against the Fear & Greed history.
• Upper bands have historically coincided with cycle tops, but price has spent very little total time there, so "we are in the orange band" is not a timing signal.
• The middle is where price spends most of its life. A mid-band reading is genuinely uninformative, and treating it as a signal in either direction is the most common misuse.

Above all: the bands move. Every year of new data re-fits the curve and drags the whole rainbow. A price that sat in the red band in 2017 would sit several bands lower today at the same dollar level.

The maths

We fit a logarithmic regression across the full weekly history:

log10(price) = a × ln(days since 2009-01-03) + b

Coefficients a and b come from ordinary least squares over every weekly close since August 2010. The nine bands are drawn at fixed offsets in log10 space either side of that fitted line, which means each band is a constant percentage width rather than a constant dollar width — appropriate for an asset that has moved over six orders of magnitude.

Why you should be sceptical

• It is fitted to its own output. A curve drawn through past prices will always describe past prices well. That is not evidence of anything.
• Time is the only input. No supply schedule, no flows, no adoption, no rates. The model cannot know about anything that has not already shown up in price.
• The fit drifts. Each new data point changes the coefficients, so band boundaries quoted today will not be the same next year.
• Log growth cannot continue forever. Extrapolated far enough, the model implies prices no market could support.

Treat it as a way to keep perspective during both panics and manias — and pair it with data that reflects real positioning, such as open interest or the liquidation map.