Throughout a typical trading day, Derwentâs technology will scan Twitter and select about 10 per cent of the available tweets at random. Those messages will then be sorted in to one of a dozen mood states: calm, alert, sure, vital, kind, and happy. Based on how many users fall into each state, the software will make predictions about where stocks are headed three or four days laterâand then make trades.
However, Dr John Bates, CTO of Progress Software, who provides market surveillance technologies to financial institutions such as the FSA, is skeptical about trading via Twitter.
âPredicting stock market moves via Twitter is highly unlikely. If an event such as a war or financial crisis occurs in real-time, information in 140 characters is not going to help organisations make real-time business decisions.
âAny information source that can track human emotion requires real-time surveillance technology to monitor exactly how the public mood induces changes in financial markets. For example, unusual peaks several days after as a misinformed tweet could lead to market manipulation or trading errors.â
Bates went onto say: âWhile from a humanitarian standpoint, global consciousness could help us understand world sentiment better, organisations must understand that Twitter is just an interesting economic indicator by the time global sentiment has "caught up", but not suitable as a real-time trigger. It might be useful to see when a recession is ending or the public is fed up with a world leader or other such momentum-based events.
âIn the future, analysing sentiment on the Bloomberg network might provide a greater real-time trading indicator for financial markets.â