HSBC BANK PLC – Form 8.5 (EPT / RI)


Robots wrestling on Wall Street with the confusing world of Reddit

(Bloomberg) – Rocket emojis for stock gains. “Tendies” as slang for profit. GIFs with Ticker.Reddit forum WallStreetBets are difficult for people to follow at the best time. But spare yourself thinking about machines. After the frenzy of retail inventories last month caused unprecedented devastation, hedge funds gambling platform algorithms have a renewed sense of purpose in their mission to grasp the next craze in the market. Still, it proves massive pain. It’s not easy to train computers to extract amateur chatter on bulletin boards into data that is ubiquitous for real-world trading. Even for the very basic task of identifying securities, the algo must learn how to match millennial speeches, memes, and spelling errors with the intended subject. And that’s just the beginning. Just ask the people behind Reddit’s robots like Stefan Nann. “You can’t just apply a standard English word library,” said the CEO of Stockpulse, a German social media analytics firm. “We read these comments and decided whether this comment was positive or negative – that’s how we train the machine.” WSB mood analysis is the latest thing in the world of alternative data, which is projected to rise from last year’s 1.64 to 17 , $ 35 billion in supplier revenue by 2027. NN Investment Partners and PanAgora Asset Management are among the systematic investors who scrape social media to trade signals, while several brokerage houses offer clients the tools to do so. For good reason in theory. A herd tracking strategy would yield retrospectively large profits. According to Stockpulse, the indicator measuring the GameStop bush on Reddit peaked in early December, a solid month before its price began to rise. The strategy of simply buying the five companies most discussed at the WSB in the previous week could return 61% in 2020, the backtest of data provider Quiver Quantitative has shown. That roughly works. A Reddit user is slowly talking about why BlackBerry Ltd. worth more than four times the share price. Algo, trained by a vendor like MarketPsych, then records the ticker and moves the post to trade sentiment signals. He tries to determine the intensity of whip calls from signs like “market leader”, emojis, use of future tense, even disturbing ones. Then the procedure is repeated in all securities. “Some call us because they’re trying to take advantage of the herd,” said Richard Peterson, a board-authorized psychiatrist who founded MarketPsych. “Some are just trying to find ways to protect themselves.” But for a form of artificial intelligence known as natural language processing, bulletin boards are not as simple as, say, detecting a company’s CEO signal on a call for earnings. If new slang appears, people need to update the machine vocabulary. It’s even harder to do around the world. It took a while for MarketPsych’s team to understand what different emojis mean in different cultures or for British stock-traders to use more subtle insults. In one example cited by StockPulse, a computer must learn the difference between “hold” as a verb versus a stock traded on a stock exchange with a tag containing the same characters. These efforts are in high demand. “At one point, it starts to look like one hedge fund because they all behave the same way at the same time,” said Francesco Filia, executive director of hedge fund Fasanara Capital, which began supervising Reddit internally last month. “You have to be on top.” Thinknum and Social Market Analytics are among the growing number of mood data providers that have recently introduced stock-consolidating products on Reddit in one form or another. Bloomberg LP, the parent of Bloomberg News, provides access to alternative data sources on the terminal and through the Bloomberg Data License. With individuals now making up about a quarter of U.S. stock trading volume, knowing where retail cash should theoretically prove lucrative. Such insights would help the long-running crowd figure out where the bombs are landing in the market. The retail military is the busiest in the record industry in January, as they crammed into some of the most hated names to hurt retailers in shorts or “shorts” in the native language of Reddit. signals in the news and social media have had a better effect than merely chasing price swings. But certain forums like Reddit don’t have a long enough history of picking stocks to shred data that quanta like Lin have. Despite her anecdotal evidence, small investors do not have consistently good results, she says. “It’s not a simple linear relationship between mood and stock returns,” Lin said. “You need a lot more research to filter out noise to really make it useful.” Read more: Quants Sound Alarm as everyone chases the same alternative information Emmanuel Hauptmann says his quantum team from RAM Active Investments recently began monitoring stock chatter on Twitter – which he considers a high correlation with Reddit – but his expectations are low. “Right now we see this as a pure risk reduction,” the fund manager said. “We don’t use it as an alpha signal.” Yet, even with all the data in the world to track low tide and retail sentiment movements, timing the market is everything. “When people completely fall for you – saying ‘I love this, death to shorts!’ – it’s too late, “said Peterson of MarketPsych. (Adds story on alternative data) For more articles like this, visit us at Subscribe now to take advantage of the most trusted business news source. © 2021 Bloomberg LP