The math, briefly
Power Law regression fits the equation price = a × days_since_genesis^b to Bitcoin's full historical price series, solving for the constants a and b via log-log linear regression (taking the log of both price and days converts the power relationship into a straight line, which ordinary linear regression can fit). The result is plotted as a corridor — a central fitted line plus upper and lower bands representing historical deviation around it.
Why it looked compelling
Unlike models pegged to fixed calendar dates or specific price levels, Power Law's log-log framing means it naturally accommodates orders-of-magnitude price growth over a long timescale without needing constant rescaling. For much of Bitcoin's history, price has stayed roughly within a wide band around the fitted line, which is part of why the model attracted a following.
The core objection
A power law relationship between price and time isn't derived from any economic mechanism — it's a curve chosen because it fits reasonably well after the fact, out of many possible curves that could be tried. When price has broken outside the corridor (as it has, historically), the standard response has been to refit the corridor around the new data. The breach itself never gets treated as evidence the model was wrong. A model that gets redrawn every time it fails can never actually be falsified, which is the opposite of a scientific forecast.
How this site handles the refit problem
Rather than pretending the model is static, this site's Power Law page refits the regression live, on every page load, from the complete available price history — and says so explicitly. That doesn't fix the underlying methodological issue (it's still a curve fit, not a causal model), but it avoids the worse practice of quietly redrawing a "fixed" model behind the scenes without disclosure.
Today's price against the power law corridor — refit live from the full price history, every refit disclosed — sits on the Power Law chart.