Cryptocurrency hedge funds and asset managers are spending up to a million dollars on algorithms that track community sentiment, according to a Reuters report. The report, published Wednesday, states that the huge fees are being paid so that big players in the crypto world can identify sentiment and trade accordingly, with between $500,000 and $1 million being spent on bots that are capable of only reading Twitter in English, such is the demand for them.
Sentiment Arms Race
The sentiment arms race has come about because very little of the price action that occurs in cryptocurrency is fundamental in nature. Unlike in traditional markets where most of the players are hardened traders with the stomach to hold through tough times, the majority of crypto traders are amateurs who react to the slightest change in price by buying or selling. This makes the market very easy to manipulate for those that can read sentiment and act accordingly, which is what these bots can do, and with regular double-digit swings playing out on a daily basis in the crypto world, there is tremendous money to be made from it.
Demand to Increase
The bots attempt to extract meaning from the millions of Twitter, Reddit, and Telegram posts made every day by using ‘natural language processing’, the act of identifying keywords and emotions that indicate the changing views of social media users with regard to particular digital currencies. They can then use this information to work out when the sentiment on a token is overwhelmingly bearish or bullish, and trade accordingly.
One issue they encounter is where paid ‘shilling’ of a token is undertaken by a person or a group of people on social media. This has the ability to skew the data, but, like training a seeing-eye dog to ignore a hot dog stand, bots can be trained to ignore such posts, with varying amounts of success. In a world where sentiment has the ability to drive price over substance, developers that can tap into such a rich seam of information will be in increasing demand as the space evolves.
