Debunking Bitcoin’s natural long-term power-law corridor of growth

Some models are useful, some fail to meet required underlying assumptions.

What’s this all about?

After PlanB wrote his (by now) famous piece on the relation between Bitcoin’s stock-to-flow ratio, a lot of people tried to debunk his model, including the author of this piece. It also inspired a lot of people to develop a competing model. But to my knowledge until today no-one succeeded in either a successful academically valid rejection of the model, or to come with a better model. One person in particular keeps coming back saying his model does a better job. Harold Christopher Burger’s idea was to build a comparable model, but instead of taking the natural logarithm of the stock to flow ratio as an independent variable, he thought the natural logarithm of time would be a better input. He reasoned that stock-to-flow is a function of time, so time itself must perform better. Even though Nick has demonstrated repeatedly (here for instance) how PlanB’s model is just better, Harold keeps claiming his work is superior. Time to take a closer look at his model and either applaud him for doing a fantastic job, or just send it (=his model) to the graveyard where it can take a rest with other failed attempts.

Regression again

Just like PlanB’s work, Harold’s work is also based on Ordinary Least Squares Regression. So, we check to see whether all the required assumptions are met. I have mentioned those…