This paper identifies nine methodological concerns across two prominent Bitcoin price power law analyses: a widely circulated retail Monte Carlo simulation and Santostasi and Perrenod (2026), “A Mechanistic Derivation of the Bitcoin Price Power Law.” Both analyses demonstrate that Bitcoin’s price history from 2010 to 2026 is well described by a power law with exponent approximately 5.69. Neither establishes that this relationship constitutes a forward‑binding structural constraint.

The retail variant fails on five grounds: residual stationarity demonstrates historical consistency but not causal necessity; the Monte Carlo boundary condition is circular; the volatility decay narrative is a log‑scale coordinate artefact; the halving‑cycle regime segmentation is entirely in‑sample; and no generative mechanism is provided for the relationship’s expected persistence.

The academic variant fails on four additional grounds: the composition identity \(\beta = \beta_{1} \times \beta_{2}\) is algebraic necessity, not independent evidence; the epidemic spreading derivation is post‑hoc rationalisation fitted to the measured exponent rather than an independent prediction; the Bayesian stability analysis applies a conjugate update to autocorrelated rolling estimates as though they were independent draws, artificially compressing posterior uncertainty by a factor of approximately 5.6; and the paper omits a Granger causality test between price and address count, leaving the direction of causation unestablished.

The paper concludes with a five‑stage pattern analysis and a diagnostic checklist for identifying when quantitative models have crossed from empirical description into unfalsifiable advocacy.

Retail Monte Carlo simulation
1. Residual stationarity establishes historical consistency, not causal necessity.
2. The boundary condition (price cannot sustainably trade below the floor) is circular – imposed by construction.
3. The volatility decay narrative is a log‑scale coordinate artefact, not an observed property of the price series.
4. Halving‑cycle regime segmentation is entirely in‑sample; no out‑of‑sample validation is provided.
5. No generative mechanism explains why the power law should persist.

Santostasi & Perrenod (2026) – academic variant
6. The composition identity \(\beta = \beta_{1} \times \beta_{2}\) is algebraic necessity, not independent evidence; all three quantities are estimated from the same 5,696 daily observations.
7. The epidemic spreading derivation is reverse‑engineered from the fitted exponent; the adoption network degree distribution and transmission parameters are not independently measured.
8. The Bayesian stability analysis applies a Normal‑Normal conjugate update to 1,899 overlapping rolling estimates as though they were independent. First‑order autocorrelation ≈0.95 reduces effective sample size from 1,899 to approximately 40–100, inflating precision by a factor of ~5.6. The reported 95% CI of [5.703, 5.754] is an artefact of the estimation method.
9. Omitted Granger causality test: the paper assumes addresses drive price, but the reverse arrow (price drives address creation) is at least equally plausible and remains untested.

The power law is not the first Bitcoin model to follow this trajectory. The pattern occurs repeatedly:

1. The Observation – Genuine empirical regularity identified (e.g., log‑linear trend).
2. The Promotion to Law – “Follows” becomes “obeys”; curve fit acquires physical interpretation.
3. The Unfalsifiability Architecture – Wide uncertainty bands, repricing floor, distant falsification criteria, outcome‑attribution asymmetry.
4. The Social Infrastructure – Community forms; dissent is reframed as analytical failure; technical apparatus raises barrier to critique.
5. The Collapse – Market delivers outcome sufficiently extreme that the architecture cannot contain it (e.g., stock‑to‑flow).

The power law is currently at Stage Four. Its uncertainty bands (±0.30 dex, factor of 2) mean any price between ~$45,000 and $180,000 at current levels falls within the prediction interval. The floor reprices upward continuously. Falsification criteria are set for conditions that have not been triggered and are unlikely to be triggered within any practical investment horizon.

• Does the model claim to derive a future price path from historical data alone, without reference to macro conditions, regulatory environment, or market structure?
• Are the uncertainty bands wide enough that the model cannot be directionally wrong over a 12‑month period?
• Is the floor or support level defined relative to the model itself, such that it reprices upward if price rises?
• Are the falsification criteria set at price levels or time horizons that are practically untestable within a retail investor’s horizon?
• When challenged on methodology, does the model’s advocate claim that critics have not understood the model, rather than engaging with the specific methodological objection?
• Does the model produce a specific price target that is dramatically higher than current levels — sufficiently large to be motivating, but sufficiently distant to be unverifiable in the short term?
• Has the model been validated out‑of‑sample, on data withheld from the fitting process, with pre‑specified criteria for success and failure?

A model that fails any two of these tests should be treated as a narrative rather than an analytical tool. A model that fails four or more should be treated as advocacy dressed as analysis. The retail Monte Carlo analysis fails five of seven criteria. The Santostasi‑Perrenod paper fails six.

Stripping away the apparatus, the analyses examined in this paper establish the following: Bitcoin’s price history from 2010 to 2026 is well described by a power law with exponent approximately 5.69. The same dataset shows address growth following a cubic trend and price scaling superlinearly with address count. The power law residuals are stationary over the historical observation period. These are legitimate empirical findings. What the analyses do not establish from this evidence:

• That β = 5.69 will remain the correct exponent in future market cycles.
• That the adoption network has the structural properties the epidemic analogy requires.
• That the floor is a genuine causal constraint rather than a regression line through a trending dataset.
• That the Bayesian stability result is informative about future stability rather than historical consistency.
• That price at any given future date will respect the power law’s confidence interval.

This paper relies on publicly available datasets. Bitcoin price and address count data were sourced from Quantodian Publications (Burger & Vijin, 2024). Rolling β₁ estimates were sourced from btcpowerlaw.nl. S&P 500 and Nasdaq Composite historical price data are available from standard financial data providers. No proprietary or restricted datasets were used. Replication code and supplementary materials are available to qualified researchers upon request to consulting@paulfaulkner.com.