Paper Finds 16 Years of Data Cannot Confirm Bitcoin’s Power-Law Trend

A new statistics paper questions how much weight investors should put on the bitcoin power-law price model. Carlos Baquero and Raquel Menezes submitted “Deceleration without identification” to arXiv on November 10, and Crypto Insiders reported on it on November 14.

The authors argue that a close fit on logarithmic charts is weak evidence, because persistent, random-walk-like deviations can create convincing trendlines and high R-squared values. They tested 820 weekly closing prices from January 2011 to September 2026, according to Crypto Insiders. In simulated price series with no power-law structure, a strong fit appeared often: about 14% of the artificial series fitted the line better than bitcoin’s actual price, the outlet reported.

The paper says constant drift in log price is rejected at the 5% level, while a log-age coefficient of about 5.7, close to the published exponent, is not rejected. That result does not hold if deviations are allowed to be more persistent than a random walk. A Gompertz curve, which assumes growth levels off toward a ceiling, fits about as well, and the authors say they cannot show that prices revert to the trendline.

The authors pre-registered forecasts for 2027 to 2030 so future prices can score the models. Crypto Insiders reported central end-2030 forecasts from five fitted models ranging from about $85,000 to $1 million, with the power-law model near $380,000. The paper itself is a preprint and has not been peer reviewed.

Track the market yourself with AiTraderView’s FREE tools: Crypto Cockpit for charts and 75+ indicators, Crypto Screener to find setups, and Crypto Portfolio to follow your holdings. Just create a FREE account and try them yourself!

Sources:

• arXiv: Deceleration without identification (Baquero, Menezes)

• Crypto Insiders: Onderzoek zet vraagtekens bij populair Bitcoin-koersmodel