Friday, January 16, 2009

Iterative wagging of the dog

(Traders often say that the news breaks with cycles... now will the cycles break with news? Hmmmmm... how's this for a new trade, long the squint and short the print? - AM)

By Anuj Gangahar
Published: January 15 2009 19:59
Financial Times

Thomson Reuters, the information provider, will on Friday launch a system designed to predict market volatility using real-time news feeds, as traders focus on future price movements in preference to historical data as the main basis for their automated trading systems.

It comes after a period of 18 months during which sudden spikes in volatility have become the norm and the quantitative and automated models used by many traders, the bulk of which rely largely on historical data, have proved unreliable at dealing with it.

The Thomson Reuters system uses news alerts that feed directly into an algorithm, or computer model, to try to predict future market volatility which can in turn be used to better inform a trading decision or risk management process. A series of real-time indices have been developed to measure when abnormally large amounts of news happen in various categories. When the level of news reaches a certain threshold, a signal alerts customers to potential market movements.

The new system differs with past efforts that have tended to focus on historical data as the main inputs to risk management and trading models rather than real- time news.

Sang Lee, managing partner at the Aite Group, the US consultant, said: “When you look at algorithmic trading overall, most participants are increasingly hungry for data. But it used to be the case that most of the data used by algorithms was traditional tick data and historical data.”

Efforts to improve algorithmic trading come as its use continues to spread. Recent estimates suggest that almost 50 per cent of all equity trading is now done algorithmically.

No comments: