I don't know why, but I now feel the urge to post another one:
Source:
Bitcoin’s natural long-term power-law corridor of growthThis is an uberbullish log-log graph (log axis on values, log axis on time):
It predicts that the price will not reach $100 000 before 2021, but it also predicts that the price will not be lower than $100 000 by 2028. It predicts that the price will not reach $1 000 000 before 2028, but also that the price will not be lower than that after 2037. The model predicts ever increasing prices, although at a slower and slower rate.
Hope you like, Mic
Looking at the graph the maximum price before 2021 is well under 100k (90k).
Other than that: cool graph. Seems to be a bit more on the conservative side when compared to the past.
The concept of putting time on a log scale is... interesting.
Generally, you would use a log scale for something that accumulates, grows or shrinks. Such things as stock and commodity prices, the mass of a pile of rocks, and absolute temperatures Kelvin don't exhibit negative values; they have a well defined zero point, and positive values are the only values to consider. It makes sense to use logarithms on scales where zero is fixed, accounting for changes
in proportion.
Time doesn't have this proportional quality. One doesn't say there was twice as much time in the year 2000 as there was in the year 1000. And to use a log scale at all, you have to set a point for time zero.
Coinmonks didn't state what point he uses as time zero for his log axis, which led me to do a little mathematical forensics on his chart to find out.
You'll notice that the distance from 2011 to 2012 is the same as the distance from 2013 to 2015. so:
(2012-t
z)/(2011-t
z) = (2015-t
z)/(2013-t
z)
and
time zero = 2009
edit: InTheLoop made a comment on Coinmonk's Medium post that bears directly on the matter of defining bitcoin's time zero for the price model
Thanks Harold. You are using time in days since 2009, likely based on bitcoin’s genesis block on 3 Jan 2009. However, in the beginning there wasn’t a great deal of trading and thus price discovery,
so it’s natural to introduce the starting day as an additional parameter in the model. Together with slope and intercept this leads to a three-parameter power-law model, exactly what I have described in my article.
That shift parameter then turns out to be 312 days before 17 July 2010 (the origin of your and my price data), so based on current full historic data, bitcoin’s starting date of the power-law is estimated at 8 September 2009.
Using such 3-parameter fit will increase the quality of the fit even further. See:
https://medium.com/@intheloop/when-moon-rational-growth-ranges-for-bitcoin-ffaa94c9d484