Q&A: Twitter's 'leaked' results

So what happened with Twitter on Tuesday night?

The social media site fell victim to the speed of its own distribution when Selerity, an independent data gathering company, found Twitter's earnings on Twitter's investor relations website and broadcast it to the world - on Twitter.

Embarrassing, but what does it show?

News from Twitter is increasingly moving markets.

Two years ago the Associated Press Twitter feed was hacked and falsely alleged an attack on the White House. Within minutes the Dow Jones had dropped 150 points and threw trading in the E-Mini S&P 500 futures contracts on CME. Gold and market volatility as measured by the CBOE's Vix index also rose.

Nevertheless, the financial world is well aware of the power of Twitter as a distribution tool. A few years ago Bloomberg incorporated select accounts into its terminal while the Securities and Exchange Commission, the US markets regulator, allowed companies to tweet corporate activity.

Companies such as Selerity and Gnip take data from well-known social media sites and feed them to capital markets participants.

In the past Carl Icahn, the US investor, has made a paper profit and moved the share price of Apple legally by taking to Twitter to announce he had bought a stake in the US technology group. As a result, any investor can react to unexpected events without even owning a Bloomberg or Reuters terminal.

It still relies on the human to act on the machine's findings though, does it not?

A human still plays a critical role and probably limited the potential damage from the fake AP tweet (people check other news sources for confirmation). Even so, the lines between human and machine are getting blurred.

Among the developments in recent years are services that can scan the news automatically, known as machine readable news. Exchanges, data providers and independent software makers have developed systems that can read messages for data or sentiment, which in turn can generate an automatic trade at high speed.

It is particularly useful for expected announcements such as major economic data, although software has been developed to include scanning for written phrases such as "rise", "fall" and "warn". The stories are then tagged as neutral, positive or negative.

Machine-readable news for economic data from reputable sources such as the US Treasury have become an accepted - if small - part of the market, in part as it is expensive to use.

Using free news sources on the internet has presented other problems. In recent years investors such as hedge funds have been exploring how to incorporate Twitter and other internet-based news sources to gather intelligence, but not necessarily turn it into an algorithm that can trade automatically. That is in part because the computers, for all their cleverness, cannot easily analyse sentiment from short tweets. It is equally challenging for the machine to understand the context within sarcasm, emotions and profanity.

It is unclear if it was humans or machines that reacted fastest to Selerity's tweets. It is worth noting that Selerity has actively courted the market's high-speed traders. Equally, it emerged at a quiet time of the day. A few big orders from a human reading the news on Twitter and reacting could equally have helped push Twitter's share price much lower.

Does this create an issue for markets?

There are two main areas that concern traders, exchanges, software vendors and global regulators alike. The first is one endemic to computer-based trading. One of the dangers within all automated systems lies in what is known as a positive feedback loop, in which a small change in computer trading feeds back on itself, triggering a bigger change, which in turn feeds back on itself, and so on. In effect, people see markets move and react to it. The process amplifies volatility.

Second, as markets are so interlinked it can be difficult in tracing the source of massive market dislocation. US authorities only last week arrested a UK trader for trading allegedly related to the "flash crash", which took place nearly five years ago. A hacked Twitter account of a reputable news source could potentially create some dislocation, as the fake AP tweet showed. Finding out who did what and when - and who benefited - is difficult even when traders have unique electronic trading identifications. Trying to discover culprits who may inject falsehoods into the market may be even harder.

© The Financial Times Limited 2015. All rights reserved.
FT and Financial Times are trademarks of the Financial Times Ltd.
Not to be redistributed, copied or modified in any way.
Euro2day.gr is solely responsible for providing this translation and the Financial Times Limited does not accept any liability for the accuracy or quality of the translation