MiesOnCharts·BTC / USD · daily · 2012 → todayloading…

Bitcoin Value Score

One number, 0–100, that says whether Bitcoin is historically cheap or expensive. It blends seven metrics — MVRV, NUPL, AVIV, STH- and LTH-MVRV, RHODL and power-law deviation — each normalized by an envelope that decays as Bitcoin ages, so the shrinking cycles of a maturing asset still register as full-blown tops and bottoms. Computed walk-forward: every day's score uses only data available on that day.

Components, all computed walk-forward with expanding windows, in three families. Cost basis: ln(price / realized price) — the log MVRV ratio; NUPL = 1 − realized price / price; ln(AVIV) — MVRV restricted to active coins (lost/ancient supply excluded); ln(STH-MVRV) and ln(LTH-MVRV) — price against short- and long-term holders' cost bases. Behavior: ln(RHODL) — the 1-week vs 1–2-year realized-HODL ratio, which spikes when fresh coins dominate (tops) and collapses in dormancy (bottoms). Trend: the residual of log10(price) = a + b·log10(days since genesis) — deviation from the power-law trend. Each component is centered on its expanding mean and divided by a decaying envelope σ(t) = √(π/2)·(c + d·log10 age), fitted by regressing the component's absolute deviation on log-age — so a +1.5σ stretch in 2025 counts the same as a +1.5σ stretch in 2013 even though the raw move is far smaller. The enabled z-scores are averaged within each family, then across families — so the five cost-basis siblings can't outvote the other signals — and mapped through the normal CDF: Score = 100·Φ(z̄). The toggles above the score chart pick which components enter the average — showing how much each metric drives the zones. The score starts where every enabled metric has data (all seven: 2014); disabling the newer on-chain metrics extends it back, and power law alone reaches 2011. Zones: ≤ 25 historically cheap (bottoms cluster under 5), ≥ 80 historically stretched (tops cluster above 90). Fit today: power-law exponent –, envelope σ today – (log10). This is a descriptive rear-view model, not advice — a regime break (or a different normalization choice) would change the zones.