Acquire Media Gives News Traders an Edge

Roseland, NJ-based news aggregator Acquire Media has incorporated a factor into its news displays and feeds that shows the anticipated direction of price movements as a result of specific news items.

The new value, dubbed Edge (Expected Directional Grade), analyzes words and phrases within news items to determine the sentiment of a story, then analyzes trading activity in the run-up to the news, and compares it with how prices have historically performed in similar circumstances in response to similar news. For example, if a company consistently beats earnings expectations, the system would adapt over time to recognize that better-than-expected results may not produce a rise in stock price, whereas a failure to meet earnings might signal a drop.

"Edge doesn't care whether it is good or bad news, only what impact it will have on a stock in relation to how it has behaved historically and where it is now - it's more about pattern recognition of how companies react over time than it is about natural language processing," says Larry Rafsky, chief executive of Acquire Media.

Edge then delivers a value of between -9 and +9 to indicate how likely the price of a stock is to fall or rise based on the news, which can be applied as a tag to its news feeds or represented graphically in its front-end displays using green or red to denote an expected upwards or downwards movement.

Acquire Media previously had an Impact rating that also looked at volatility prior to a news event, but did not predict price direction, and a Relevance rating that determines how relevant the main news is to each company mentioned in a story.

However, the vendor does not apply an Edge value to every story. Rafsky says that if Acquire cannot compute an Edge from past price behavior, it won't include one, but adds that the vendor is developing models for every industry sector and company size, since these can all react differently to news, and is also ensuring its filters adapt to the different writing styles between a news story, a press release, and a piece of research or commentary.

The Edge ratings have been in development for two years, though Rafsky says much has changed within that timeframe - especially as the markets have begun adopting algorithms that depend on low-latency news services.

"Two years ago, we said we would allow 200 milliseconds of latency [to process an Edge]... but now the game has changed. Our Tiger [Traders Integrated Execution Ready] feed is in the microsecond latency range, and our Relevance, Impact and Edge ratings could add a total of 50 milliseconds of latency... so, since it does add a little latency, most people would probably choose to turn it off in the Tiger feed," he says.