Making sense of the chaos? Algos scour social media for clues, for computer that moves

LONDON (Reuters) – After months of relative calm in the cryptocurrency markets, bitcoin has exploded back into life in April with its best price and jump in for more than a year, but very few people were able to convincingly explain why.

FILE PHOTO: A copy of the bitcoin is on the PC motherboard can be seen in this image, October 26, 2017. REUTERS/dado Ruvic/File Photo

The 20% jump, addressed to the attention of the investors is one of the enduring mysteries of cryptocurrencies: what is it that the price of a rising asset in a non-transparent, largely unregulated market.

For some, the answer lies in social media. Hedge funds and money managers looking for an edge to his training in computers in order to scrape social media sites for the triggers, which could make the price of the digital currency.

Their goal is to have the material on algorithms is the selection of the “signals” of the level of noise from sites ranging from Reddit, and WeChat to Twitter and Telegram.

A lot of investors have been using computer models to identify the market, the price disparities among the hundreds of cryptocurrency trading exchange.

But with the potential for arbitrage is narrowing as the emerging sector, the major players are increasingly looking to build or buy one of the more sophisticated robots, and to find the market-moving signals, online, and, according to interviews with six of the hedge funds and asset managers and three software developers.

However, while the use of algorithms, or algos, to the parsing of social media to help them grow, some of the respondents said the main challenges and risks in the broader implementation of the cost and the complexity.

“It’s an arms race for money managers,” he said, Bin Ren, the chief executive of Elwood, in Asset Management, specializing in digital assets, and is the property of the Brevan Howard founder Alan Howard.

“Very few players are able to deploy and supply them, but I don’t think it’s very, very rewarding.”

These “sentiment analysis” as a computer-based reading, social media voting will be used as a tool in the traditional markets such as equities and foreign-exchange market, the consumer’s feelings toward a business or the assets.

However, it would be of greater interest in the cryptocurrency markets, where there are few authoritative sources of information, such as a central bank, can hardly provide reliable data to measure the intrinsic value of economic indicators in the financial statements, and a high proportion of individual investors.

It is early days for the technology in the computer industry, with very little industry-wide data regarding the performance of, and the many, many questions about its effectiveness. None of the other options, Reuters said it would give information about the performance of their algorithms, citing commercial confidentiality.


To be sure, the digital currency’s share some of the drivers in the traditional market as well as the comments made by the decision-makers. Bitcoin is subject to the comments made by the supervisory authorities in particular and The has fallen sharply over the past week, after U.S. Federal Reserve chief called for a halt to Facebook’s planned to Scale cryptocurrency project.

But seeing as cryptocurrencies are linked to the internet, as of the start of a decade or so ago, when the word has been spread in the forums and chat rooms, it seems to make sense to look at the price triggers that will be online.

However, it is far from cheap or easy to design an algorithm that allows to find market-moving signal in the cacophonous world of social media, the analysis of a large number of messages in dozens of languages, while the selection of unreliable information.

Andrew Leccese, president of Bluesky Capital, an investment firm in New York, said the upfront cost for a robot that is capable of only reading from the Twitter account of the English, were in the range of $500,000 to $1 million, with the bulk of the funds to qualified developers. This has been put off, Bluesky, the use of the technology, ” he said.

A major challenge is the large number of social media channels. Over on Twitter, the sites are often used by cryptocurrency enthusiasts have been a Telegram, a messaging app with a free tv, and Reddit, and a messaging board.

In east Asia, and is home to many retail stores, of apps, such as Line in Japan and Kakao in South Korea, are in high demand.

Tens of thousands of comments on cryptocurrencies can be pumped out around the clock, both at the national as well as international ones.

Reddit’s main forums, or the subreddit for bitcoin only has 1.1 million members. And Twitter see tens of thousands of posts with a mention of bitcoin, each day, with between 14,000 and 32,000 per day for the past three months, according to the BitInfoCharts web site.

In an attempt to extract meaning out of all this chaos, the algorithms will make use of a so-called natural language processing to identify key words and emotions that indicate a change in the way in which a social media user with a specific digital currency.

Investors are using algorithms, that is to say, that they are to identify patterns of information that gains traction online.

The information is distributed, not at random, but by a very well-defined structure, it shall be like a tree,” said Elwood, with the Dash, which is used in sentiment analysis, for nearly two years, following the development of a software application.

“It’s very similar to the one the modelling of the spread of the virus.”


Other investors are focusing on the challenges of teaching machines to be able to spot biased or inaccurate information.

A Reuters report (here) in November of last year, it became apparent that a lot of users are on social media, participate in the money in the positive reviews about the digital currency.

BitSpread, a cryptocurrency investment manager based in London and Singapore, which uses its own capital to trade with the help of an algorithm, it is started to develop about a year ago, the chief executive officer of Cedric Jeanson, told Reuters.

It is a rather narrowly-focused software. The aggregation of the Twitter feeds, it looks for messages in the winding-up or closure of positions in stock markets.

“It’s just a matter of gathering all the info, will try to get to know who is trading where, and what type of card it may appear,” he said. “It’s a strategy that makes sense.”

However, he recognised the disadvantages as well.

“The sense of self, and what we’re seeing on Twitter, you can really focus on is the fake news. We have to be very careful with what we read in the news, because most of the time, we have seen that there is a bias.”

A lot of algorithms of machine learning, where they are thought to be improved by the experience and a better understanding of how social media posts translate to the market movements.

Developers can often identify people with outsized voices, and a large number of his followers, to weight more heavily in their algorithm, said Bijan Farsijani, from Augmento, a Berlin-based startup, which launched a algorithm to sentiment analysis in the last month.

He said a number of hedge funds had bought the software from the company since its launch.


Bitcoin is the largest cryptocurrency, and is a bellwether for the sector has increased by more than 180% this year, driving up the interest of the larger investors in the trading companies to hedge funds.

Bitcoin is in the latest rally, just last month, was seen by analysts as driven by the expectations of the wider use of cryptocurrencies is driven by Facebook’s Scale.

This movement was mirrored by an increase in the rate of interest is online. A Google search for cryptocurrencies to hit their highest level in three months, on the 18th of June, when Facebook made its announcement.

However, it is difficult to determine the chicken and the egg: on-line chatter, if the price is moving.

“There may be a single value of sentiment analysis in the computer, but most of the time what people are tweeting can be a lagging indicator of price, go for it,” said Leccese by Bluesky Capital.

“However, there is the potential,” he added. “People are going to go and have a look at this more in the next five to 10 years, because there’s diminishing returns as a result of the increased competition in the traditional areas.”

While there is a general lack of data specific to this technique, the “quantitative” cryptocurrency, funds that make use of the methods of arbitration, for the question of sentiment analysis, which is considerably better than the funds with longer-term efforts in the first quarter of this year, the PwC report shows.

They are saying that they are in increasing demand.

Slideshow (3 Images)

Taiwan-based, Marc, and Howard, along with more than 500 machine-learning experts and crowd-sourcing sentiment analysis algorithms to the data sources, including Google Trends, Reddit, and development of the platform on GitHub.

Howard said that his bitcoin investment, and with the help of an algorithm to beat, and the funds just keep track of the price of the cryptocurrency, with 54% in the year to the 24th of June, adding that the funds are in New York city and new Taipei had to beat him in order to help them to develop their own analysis.

“It’s pretty hot right now,” he said. “A pension fund that is worth their salt, they will be working on a number of resources and their allocation for sentiment analysis.”

Report by Tom Wilson and Simon Jessop; Editing by Pravin Char

Recent posts

Follow us

Don't be shy, get in touch. We love meeting interesting people and making new friends.

Most popular