Q50 is the fitted power law. Other bands are historical rich/cheap offsets around that line, compressed after the last data day — not separate forecasts.
Tap and drag on the chart for full details.
Where Bitcoin sits versus the long-term trend
Q50 is the fitted power law. Other bands are historical rich/cheap offsets around that line, compressed after the last data day — not separate forecasts.
Tap and drag on the chart for full details.
Pick a year-end and a corridor line to see the implied price and CAGR from today's actual BTC price.
—
| Period | CAGR | Start Price | Start Date |
|---|---|---|---|
| Loading Bitcoin CAGR… | |||
| Regime | n | +3 months | +6 months | +1 year | +2 years |
|---|---|---|---|---|---|
| Loading historical returns… | |||||
| Loading crossover projections… |
| Asset Class | 30d | 90d | 180d | 1y |
|---|---|---|---|---|
| Loading rolling correlations… | ||||
/correlations. Property proxy = VNQ (US REITs). Updates with scripts/update_data.py.
log₁₀(price) ~ log₁₀(days since genesis) using all data up to that date.
/stats.
| ID | Test | Status | Live metric | |
|---|---|---|---|---|
| Loading falsifiability tests… | ||||
/stats.
F5 uses expanding-window OLS R² (paper §7); naive trailing 3-year refits are cycle-dominated and shown only
as a diagnostic. F2 (address-growth βA) and F4 (Metcalfe R²)
need on-chain address series and are not monitored here.
Every path starts at today's price. The residual r = log₁₀(price) − log₁₀(Q50) then follows a daily Ornstein–Uhlenbeck step r ← r − κ(r − 0)Δt + σ√Δt · Z with κ = ln(2) / half-life. The deterministic power-law trend is projected forward and the residual path is added back.
Floor reflects residuals below the historical mean − 2σ. Student's t uses 5 degrees of freedom, scaled to unit variance.
/monte-carlo/calibration.
Paths are generated in the browser so sliders stay responsive; the same OU engine is also available at
POST /monte-carlo/simulate. Residual-bootstrap is a planned alternative to the OU shocks.