Abstract:

Despite the illusory nature of their prospects according to a series of analytic studies, cryptocurrencies remain one of the most popular objects of investment in modern markets. A comparison with a pilot portfolio composed of shares of the comparable promising areas of business will reveal both the temporary competitiveness of cryptocurrencies and the presence of more sound prospects for competitors. The use of Fama – French analysis allows identifying the institutional separation of cryptocurrencies and the unsuitability of traditional tools for their analysis in the long run. At the same time, in the short run, the methods of machine learning and linear regression allow to reveal the relative predictability of cryptocurrencies markets. The Logit model on the base of data from previous periods also reveals additional features that far outweigh random selection.