I'm exploring with Claude a DCA and small leverage strat with Claude using the deviation and daily slopes from the power law.
This chart shows the growth in Bitcoin terms achieved by a strategy based on a Hidden Markov Model (HMM) of Bitcoin’s local power-law slopes. The model identifies different dynamic market regimes and assigns an actionable exposure to each state, ranging from a cash/Bitcoin combination to full Bitcoin and modest leverage (~1.5×).
Importantly, the results are based not only on backtesting but on a walk-forward test from 2014 to the present, where each decision uses only information available at that time.
The vertical axis shows how many BTC-equivalent units are obtained starting from 1 BTC, while the horizontal axis shows the maximum drawdown exposure to achieve that result. The turquoise curve represents the dynamic HMM strategy; the orange curve shows constant leverage as a benchmark.
The key result is that the regime strategy can generate substantially greater Bitcoin accumulation for a given level of historical drawdown than simply maintaining constant leverage. This suggests that Bitcoin’s changing local slopes may contain useful information for dynamically adjusting exposure.
An BTC balance of 70 % and 30 % cash is optimal.
These remain historical walk-forward results, not a guarantee of future performance, and real trading would also be affected by costs, execution, financing, and future regime changes.
This indicator will be accessible in the Scientific Bitcoin Observatory (with clearl indication on the current regime) in the Benefactor donation tier.