Interesting paper just out from physics dept. Univ. of Manchester researcher, notes that if Bitcoin return under the power law falls inversely with age, and he hypothesizes and partially measures a power law relation for falling variance, suggesting that also falls inversely with age. https://t.co/MBnonpZKu7 Under such conditions one naturally gets a flat optimal Continous Kelly μ/σ^2 going forward (ignoring risk free rate which is ok to first order). His data series is CoinGecko which leaves out several early years of data, and he ends up with a too low price of age power law β, but it does not impact the overall argument about the Kelly fraction. In June I wrote a Substack arguing for a rising Continuous Kelly based on a fit with vol going as ~ 1/(t+c), reciprocal with offset, in that case variance falls faster and Continuous Kelly allocation rises (and has been rising for the past 6 or 7 years). https://t.co/xOoe23JFk3 And then I recently wrote this article on the Empirical Kelly method. In that I fit a -0.64 power law to volatility which means a faster than reciprocal fall in variance that implies also a rising Kelly fraction with time. I also correction for autocorrelation, including r and ran a block bootstrap (10,000 instances) that agreed nicely. https://t.co/KIU4RQgB78 Bitcoin is getting safer to invest in, according to these two models. Sharpe is falling but looks to be on an asymptote to about 0.53, but Kelly appears to be rising. @ScientificBTC