Tweets Culled for Investor Sentiment
Gauging investor sentiment from tweets and other social media is receiving support from the trading and academic communities.
“Twitter is full of alpha-generating potential,” said Emmett Kilduff, CEO and founder of Eagle Alpha, whose product Social Sonar, channels Twitter-based intelligence on stocks and global macro topics. “Currently this potential is masked by excessive white noise, inefficiencies in accessing relevant tweets and compliance concerns. Investors leverage our information and data to help give them more conviction on an investment but also for unique insight they wouldn’t ordinarily find.”
Social Sonar is available to market participants seeking an investment edge, including the buy-side, sell-side, corporates and retail investors.
With 500 million tweets, 200 million tweeters and 135,000 new users each day, it is becoming increasingly difficult to unearth relevant data. Eagle Alpha rectifies this by enabling users to access validated pre-built and bespoke lists.
“From game-changing CEO opinions to instant reaction around key financial news and company reporting, Social Sonar is an essential addition to the run of the mill industry newswire services and gagged sell-side analysts,” said James MacLachlan, senior trader at CF Global Trading, a provider of institutional trading services in the global equity and credit markets.
A number of recent examples highlight Twitter’s potential in delivering alpha. On July 11, 2013 a tweet stated that Grayling was calling in the U.K.’s Serious Fraud Office to look into G4S because it had refused to cooperate with a voluntary forensic audit. This was published at 12:16 pm. A half-hour later, Bloomberg published the news and the share price dropped by 7%.
At 8.05am on July 30, 2013 a tweet from a South African TV station reported the shooting of a union member at mining company, Lonmin. Seven minutes later Bloomberg published the story. The stock fell 2% inter-day while other mining stocks remained flat.
Of course, Twitter-based investing is far from foolproof, as the AP ‘hash crash’ made clear. Eagle Alpha guards against phony tweets by augmenting its service with human input. “We knew the AP tweet was fake because if the Syrian Electronic Army had written it properly, it would have followed the AP style guide,” said Kilduff. For example, “breaking” should have been in upper case, “Barack” should have had the word “President” before it.
Capital Markets Cooperative Research Centre (CMCRC), an Australian think tank, has found that models that aim to predict daily stock returns perform better when they combine text data, such as company announcements, media news, and social media, with financial data.
“By analyzing announcement and financial quantitative data, the combination of these two different types of data gives the research far more variables to analyze, which seems to have led to more accurate predictions,” said researcher Tony Zhao.
To examine the performance of these combinations, the research used 19,282 ASX announcements from the first half of 2010. The research used 80% of the announcements to train the algorithm and 20% to test the different combinations.
Combining quantitative and text data rather than treating them separately reducing errors by almost 3% compared to results when only quantitative data is examined.
The new offering sports 11 liquidity-seeking algorithms with more to come.
GMAG says vacation season is a good time for design, implementation and testing of machine-based trading.
A bank will be able to recommend specific changes to clients to improve their access to liquidity.
Tech vendor will support Canadian equities trading and interlisted securities trading via its AMS.
A target-close model allows traders to mitigate market impact, Deutsche Bank's Andrew Royal writes.