Geek magazine hacker daily blog

1 year ago
Social networks, finance and robots: As the trader earned $2,4 million in 28 minutes by means of one tweet

A few years ago one London hedge fund started the new project so-called "Twitter fund". The special computer system read 100 million tweets a week and defined a situation with the current economic trends on their basis.

If the general tonality of the message was positive, the fund took shares and if negative — that the probability of falling of stock market increased and it was necessary to think of sale of shares.

Invention it was awful — the fund warmed up in two years. However, perhaps, the idea was not so bad — managed to earn by spring of 2015 of one their traders in only 28 minutes more than $2,4 million, using a microblog service as a source of news.

Tweet force

It was talked of negotiations between Intel and manufacturing company of Altera chips. Information on them leaked out into the press, however official confirmation of this information was not yet when the unnamed trader took options for about 300 thousand shares Altera. These options had the price of execution of $36, and the stock price at that time of time made $34. Thus, these options belonged to so-called options "out of money" — everyone who would like to perform them, would have to pay $36 for the event costing $34. The transaction was performed in March, and the term of expiration of options fell already on the middle of April so they did not cost very much — $0,35 for everyone, the total amount of purchase made about $110 thousand.

20 seconds later after commission of this transaction trading in shares of Altera was suspended. The reason — the reporter of Wall Street Journal of Dan Mattiolli published a tweet with reference to Dow Jones information service that Intel and Altera really conduct negotiations on merge:

Other traders began to doubt that their colleague could manage to make the transaction on purchase of options in 20 seconds — by their calculations on it would leave about a minute. More plausible explanation — transaction was made by the computer. On optsionnoy the market high-frequency robotic trade gained distribution later, than on other sections of stock markets, however now and here the show is run by robots. So there is nothing surprising that the computer could overtake competitors, using as information source the tape Twitter.

Note: Afterwards the Reuters agency published information that the transaction on purchase of options by the trader happened before the publication of a tweet of the reporter of Wall Street Journal. After this incident negotiations between Intel and Altera were suspended. As a result the declaration of purchase of Altera for $16,7 billion was made in the summer of 2015.

It in general is legal?

This incident generated the next wave of discussions of a question of whether high-frequency trading is a fair method of trade, or it should be equated to insider trade? To insider trade traders get advantage before competitors due to access to information unknown to other players in a case — it is evidently dishonest. But on the other hand most of players has no market also access to superfast computers and software which can make the transaction on purchase of options for 300 thousand events less than in a minute.

In conversation with the Fortune edition the strategist on work with derivatives of the MKM Partners company Jim Straggler (Jim Struggler) does not agree with such setting of a question. The expert says that HFT can be a problem if robots use data from the sources unavailable to other investors or will get access to these data earlier. However, if the transaction is performed on the basis of publicly available information, for example from a social network tape, then here everything is fair.

To create the superfast algorithm analyzing data from the most different sources not only large hedge funds, but also individual traders can at home (moreover, online trading becomes a popular hobby for technically grounded people).

At the same time the existing algorithms are not ideal therefore to say that the system works nekorretno it is impossible, Straggler is sure.

"To us people who want to sell to our company analytical computer systems for making decisions on transactions on the basis of these news or tweets constantly come. We always refuse to them — ten errors are the share of one truly predicted transaction in such systems now".

The picture on a teaser: Emmanuel Dunand — AFP/Getty Images

This article is a translation of the original post at
If you have any questions regarding the material covered in the article above, please, contact the original author of the post.
If you have any complaints about this article or you want this article to be deleted, please, drop an email here:

We believe that the knowledge, which is available at the most popular Russian IT blog, should be accessed by everyone, even though it is poorly translated.
Shared knowledge makes the world better.
Best wishes.