I wanted to chime in with a couple more comments relating to algorithmic news feeds.
Even if an algo news feed is providing sentiment tags, it has to remain the responsibility of the algo engine to own the trading decision, and most news is still a long way from being predictable enough to base a buy or sell on.
Most algo designers still use news only as secondary validation, not as the primary source of trading decisions. A while back I was working very closely with one of the sharpest and most successful algo traders on Wall Street. I proposed a news-based strategy which he reviewed and built a model around. His conclusion was that news was useful as a secondary data source that would strengthen other trading models. For example, a long-short pairs trader might benefit from news by removing a symbol from his trading universe if there are a bunch of news stories with the word ‘merger’ or ‘shareholder lawsuit’ in them.
Most news-based trading today is not occurring as a result of out-of-the-blue corporate news, but on pre-scheduled events like the announcement of unemployment data or housing starts. The primary responsibility of the algo news feed, then, is just to introduce as few milliseconds as possible between the release of the data and the algo engine receiving it.