The idea behind cryptocurrency price regression.
This is a simplified regression, where the price depends entirely on time. In our case, the time equals the number of days since the first trading date in an array of historical prices.
We've assumed that cryptocurrency prices won't rise exponentially forever. Eventually, the crypto market will enter the maturity stage, so the average growth rate will slow down to the average growth rate of financial or IT markets.
Such assumption can be approximated by a function, similar to sigmoid.
The regression has been drawn with the help of sigmoid-like functions with 3, 4 and 5 free coefficients.
The three different functions provide distinct results, which may be treated as best-, neutral- and worst-case scenarios.