Reddit Posts: Which cryptocurrencies are viewed most favorably?

14 Aug 2019 by svet

I've noticed that some crypto-medias have started lately to actively refer its readers to comparitech.com post "Which cryptocurrencies are viewed most favorably? Analysis of 48K Reddit posts". Authors claim to be " ... using machine learning .. (and) ... scraped 48,469 posts on cryptocurrency -related subreddits and 7,500 articles from 48 national and international news outlets". Their conclusion is: "While only 14.5% of subreddits about cryptocurrency were scored as having a negative sentiment, a whopping 85.5% were deemed positive."

I guess we have here a typical situation of the sampling bias. We can argue that 87% of authors' statistical base (subbreddits) constitute articles written by people, which, most probably, hold cryptocurrencies or, at least, can be associated with the group of cryptocurrencies supporters (not to mention the professional coin marketers). Of course, as we all know, enthusiasts can also spread fud and negative information but in a long-term they tend to be more positive, specially, because most authors still hold tight to their portfolios built during the high-market epoch.

Additionally, authors make the following speculative assertion: " ... public sentiment surrounding each of the thousand-plus cryptocurrencies can influence their trading prices. The success of your trade, therefore, may ultimately depend on the opinions of others, both experts and enthusiasts".

What exactly can or can't "influence trading prices" is, apparently, the long standing academic debate, which we won't be digging into. However, mentioning "opinions of others" authors, eventually, forget to take into consideration the opinion of professional players (or "whales"), which historically tend to act just in opposite to "public opinion" direction.

Nonetheless, as to compensate for at least part of those discrepancies, this piece contains curious data on the total number of posts on crypto-currencies subreddits for the period of about 30 months. It shows that its drastically increased in November 2018, coinciding with the drop of prices on the market. It kept growing until January 2019 to be halved in February and then renewed its growth reaching July's climax.

There can be a number of different explanations for that. One, which seems to me more or less plausible connects this curve to smaller companies' advertisements expenses increase (after their managers realized that the market won't do their job for them) as well as to the growing interests of corporate employees to the DLT space (Libra).