Interesting, what assumptions would you say are being made with this model? E.g., Why should price be modeled as the log of the rank?
This is the
King effect. While the
majority of data points fit onto the line, the top few are above it. This suggests that we are unlikely to have the top 10 days having a similar price, i.e. there will be a significant reduction after the top is reached. This has been particularly applicable to Bitcoin so far.
They clearly don't though. The line overestimates price around 20 and underestimates around 14.