The social network Twitter has long been considered a key place for discussion of cryptocurrency topics both among experts and among ordinary cryptocurrency market players. Moreover, based on the mood of messages on Twitter, you can build a trading strategy that allows you to effectively earn on cryptocurrencies.
So, two companies – the eToro trading platform and the TIE analytic company – linked trading with Twitter analytics, resulting in the powerful TheTIE-LongOnly CopyPortfolio tool. How the new tool works, what role Twitter plays for the cryptocurrency industry, and what other ways traders have an analysis of the cryptocurrency community’s moods, we understand in detail the material.
A cryptocurrency portfolio that depends on tweets
Those who manage the cryptocurrency portfolio have the opportunity to put into practice what many had previously suspected: Twitter has become not only a storehouse of information and analytics for the crypto community, but also a pulse that cannot but correlate with what is happening with cryptocurrency prices.
EToro and TIE have launched TheTIE-LongOnly CopyPortfolio cryptocurrency portfolio for all types of investors. This tool is managed by eToro and is based on Twitter posts. This is especially interesting because, as stated in eToro, “unlike traditional assets, cryptocurrencies do not have fundamental factors – they do not have a stream of income, dividends or an indicator of debt burden.”
The minimum threshold for using heTIE-LongOnly CopyPortfolio is an investment of $ 2,000. At the moment, the initial assets in the portfolio include the following assets:
The fact that almost half of the portfolio is occupied by DASH suggests that the maximum high volume of positive tweets was observed on this altcoin in early October. At the same time, the composition of the portfolio is reviewed monthly, automatically and without human intervention, and eight more cryptocurrencies can be included in the portfolio – Bitcoin, Ethereum, Litecoin, Bitcoin Cash, NEO, Stellar, Zcash and Cardano (ADA).
The main criterion for the revision is the dynamics of sentiments on Twitter: if it shows an increase in positive moods on one of the coins, then it is included in the portfolio, and the other cryptocurrency, losing optimism on Twitter, may disappear altogether. The share of a digital asset is also determined solely on the basis of sentiments expressed in cryptocurrency tweets.
Twitter post analysis technique
Obviously, when analyzing the information flow on Twitter, messages must be carefully filtered using Artificial Intelligence (AI) and word processing using Natural Language Processing and messages from various scammers and bots. There are also about 80 popular words in the cryptosphere that are actively used, but have several interpretations. So, for example, the name of the DASH altcoin is similar to the pronunciation of the special Dash browser API.
As a result, companies have developed 30 different metrics that measure a particular tweet. TIE co-founder Joshua Frank noted that if there are a lot of tweets from Twitter accounts in a certain period of time, in which messages usually rarely appear, then we can talk about a custom campaign to promote some kind of cryptocurrency, and such tweets are not included in the analysis .
Of the more than 850 million tweets published daily, about 50,000 posts are about cryptocurrencies, which are grouped by certain types of digital assets. It should be borne in mind that half of them are messages about bitcoin. At the same time, TheTIE-LongOnly CopyPortfolio filters out more than 90% of tweets, which are an obsessive promotion of any cryptocurrencies, or scam projects, as well as message bots.
There is also a certain pattern in the crypto community, according to which the adherents of a certain asset often write (and almost always positively) about their token. Accordingly, for the analysis of moods, it is not the volume of such tweets in relation to the total number of messages about other digital assets that is used, but the change in moods relative to some cryptocurrency in dynamics over the past seven days.
The analysis method was developed for seven years by Social Market Analytics (SMA), a co-owner of TIE. The corporation has compiled a dictionary with 100,000 linguistic units, and based on the analysis of the content of tweets that were passed through the cull filter, Twitter messages are rated based on the Raw Sentiment Score system.
This process takes place almost in real time: only 300 milliseconds pass between the time a tweet is posted on a social network and the end of its analysis.
New tool shows excellent results
This tool has already been tested by TIE for a limited number of private clients, as well as hedge funds, but now thanks to eToro it has become available to the general public. Testing of the algorithm began in October 2017 and showed that for every dollar invested, the investor received a net income (i.e. minus commissions) of 213.7%. The return on invested funds in bitcoin over the same time turned out to be noticeably more modest – 41.1%.
At the same time, retrospective dynamics, for example, for the current year, shows that such a portfolio in January “lost weight” by 17%, and in February, on the contrary, rose sharply to 410%, once again showing that cryptocurrency trading remains a highly risky type of trading operations, especially at short time distances. It is no coincidence that eToro and TIE emphasize that they have a portfolio that focuses on a long-term strategy for working in the cryptocurrency market.
Research on the correlation of Twitter sentiment and cryptocurrency market
Of course, studying the dynamics of moods on Twitter can be built not only on the SMA technique. Research data can also be found on other sites, for example, on the sites Cryptoswarmanalytics.com and Bitinfocharts.com. However, the fact that crypto traders should study the mood changes on Twitter, indicates a number of studies.
So, scientists at Stanford University, Stuart Colianni, Stephanie Rosales and Michael Signorotti decided to test the hypothesis that tweets can correlate not only with changes in stock market indices, but also with a decrease or decrease in the price of bitcoin. It turned out that the accuracy of predicting the dynamics of the price of bitcoin by the clock in this case is 59%.
A group of scientists from a number of scientific schools in the United States also thought about the impact of social media on changing the price of bitcoin. As a result of the study, they came to the conclusion that social media is an important indicator of bitcoin profitability in the future. Meanwhile, experts from Sweden, Linus Roxberg and Simon Shadman, pay attention to the fact that when analyzing the moods on social networks, it should be borne in mind that most of the cryptocurrency trading takes place in Asian countries, which means that it is worth giving more weight to social media messages. this region.
At the same time, a study conducted by another Swedish scientist, Evita Stenquist and Jacob Lenne, rightly indicates that one of the key factors in the accuracy of predictive patterns in the behavior of bitcoin prices is an analysis of the relevance of tweets that fall into the forecast. At the same time, their test model showed high accuracy in predicting the behavior of the price of bitcoin depending on the mood of messages on Twitter – 83% of the forecasts were correct. And across the cryptocurrency market, the test model worked with a slightly lower probability of a successful forecast – 79%.
All this suggests that difficulties in building accurate predictive models arise even when analyzing the oldest cryptocurrency – bitcoin, and when you turn on various other cryptocurrencies, they can increase. In addition, along with an increase in the flow of funds from institutional investors, the correlation of tweets and the dynamics of the cryptocurrency market may weaken.
But for now, one can agree with what eToro experts say:
“People’s moods are the most important driver for cryptocurrency price movements.”
Meanwhile, there is a nuance in using the assessment of such sentiments on Twitter: despite the system of rejecting irrelevant tweets, the head of marketing at Rice University Utpal Dolakia fears that Twitter messages can be used quite easily by traders to “raise prices” of cryptocurrencies.
Google Trends and Bitcoin
We cannot but pay attention to the indicators of the frequency of requests for a particular cryptocurrency in Google. For example, the lack of bitcoin in the new eToro and TIE product looks logical, given the fact that recently Google Trends shows a low level of requests for the word “bitcoin”. However, here traders need to consider two nuances:
- First of all, you need to remember that the cost of altcoins is moving significantly in the same direction as the price of bitcoin, which means that even if you do not include the first cryptocurrency in the cryptocurrency portfolio, you still need to make a forecast of its value.
- Secondly, the dynamics of Google Trends shows that while on June 26, for the first time in a year and a half, the price of bitcoin reached $ 13,000, this day turned out to be the most record one in terms of the number of requests to Google for the word “bitcoin”. However, an increase in the number of such requests could be observed as early as June 8.
Index of fear and greed
In addition to the presented portfolio with built-in tweet analysis tools, as well as Google Trends, in your investment strategies you can focus on the cryptocurrency indicator of the fear / greed index (Crypto Fear & Greed Index). When this index shows a high level of “fear”, this means that there is an opportunity to enter the cryptocurrency market. If “greed” dominates, then the market is likely to go down. Of course, this is a common tool for an investor, without advice for buying or selling a particular cryptocurrency. But given that with the growth of the cryptocurrency market, the cost of bitcoin also goes up, this indicator can be used as one of the tools for building an investment strategy based on the oldest cryptocurrency.
It is curious that the “fear / greed” indicator relies only on moods on social media for 15%, as it also has other interesting components, namely:
- Google Trends (15%)
- Bitcoin dominance changes (10%)
- Opinion Opinion Survey on Cryptocurrencies (15%)
- Market volatility (25%)
- As well as cryptocurrency trading volumes (25%)
Significance of the Bitcointalk and Reddit Platforms
It is worth noting that, in addition to Twitter, there are other social platforms where there is an active discussion of cryptocurrency topics, in particular, Bitcointalk and Reddit. Despite being inferior to Twitter in influencing the mood of the crypto community, the correlation of messages on these platforms would also be worth analyzing. Attempts to do this have already been made by Augmento experts, who noted that FOMO (fear of missed opportunities) is most pronounced on Twitter and the Bitcointalk forum, but to a lesser extent on the Reddit website. Moreover, the rise in the cost of bitcoin in the first half of this year, analysts associated with an unusually large manifestation of FOMO on Bitcointalk.
The main conclusion from the discussion of cryptocurrencies
The study of what they wrote in social media and in the media, including classic ones, about cryptocurrencies was devoted to a large-scale study conducted by Comparitech and covering the period from January 2017 to July 2019. The main conclusion that can be drawn from this analysis is that at all sites where they talk about cryptocurrencies, most of the messages and articles carry a positive tone.
And this allows us to make the following important conclusion: a consensus has already been formed in the world that cryptocurrencies are a phenomenon that will develop and become one of the main factors shaping the future of mankind. This means that in the long run, those associated with this market are doomed to success.