A recent study from two Yale University academics reveals that crypto projects with a higher number of posts from Twitter bots tend to have lower returns in the long-term. The researchers analyzed over one million tweets about 48 different cryptocurrencies created between 2019 and 2021, focusing on projects such as Axie Infinity (AXS), CowSwap (COW), and Polygon (MATIC). Excluded from the study were Bitcoin (BTC) and Ether (ETH).
To measure engagement, the researchers created an engagement coefficient system that quantified the interest and interaction displayed by social media users. According to Prof. Tauhid Zaman and PhD student Khizar Qureshi, a high engagement coefficient signals artificial activity most likely generated by automated bots.
Their research showed that projects with low engagement coefficients generally experienced low returns over time. The study also discovered that engagement coefficients were correlated with the future returns of cryptocurrencies.
An investment strategy devised by the researchers, which weighted portfolios based on bot activity and engagement coefficients, showed that the latter appeared to have a more significant influence. The engagement coefficient was found to be especially important during a token’s first month.
While this can lead to pump-and-dump schemes, the study suggested that long-term returns are less dependent on social media activity than short-term returns.
It is essential to note that, in some cases, bots can be programmed to like and retweet tweets about a certain cryptocurrency, making it appear more popular than it is in reality. For instance, Krypto, a token for in-game purchases in the NFT battle game Kryptobellion, demonstrated solid initial returns before plummeting and turning negative.
The researchers concluded that social media activity could be used to predict cryptocurrency performance over months rather than just hours or days. Traders can potentially use engagement coefficients and daily social media feeds for cryptocurrencies to make short-term bets on the most viral tokens.
Though the data may be affected by certain factors like bull markets, the researchers believe their social media engagement model is adaptable enough to work across various situations. Future research could explore whether the model can predict other outcomes, such as performance in movies, TV shows, fashion sales, political campaigns, and even larger market cap cryptocurrencies.
Source: Blockworks
