all ‘‘bubbles’’ and going long on all stealth companies, but there are a number of near-
term reversals (hours to days) that can be seen with certain types of news. A downward
correction after a positive news-induced run-up or an upward bounce after negative
news are two observations. Whether these signals induce a programmatic hedge or
simply provide a confidence indicator for humans overseeing trading activity, one
can implement these types of strategies with relative ease.
In addition to refining some of the more widely published signals on volatility, trading
volume, momentum, and reversals that may be present in news, some of the more
interesting capabilities lie in the ability to predict the direction, magnitude, and duration
of pricing movements. Analyzing sentiment and other text characteristics is growing in
popularity, but simple ‘‘buy on the good news’’ and ‘‘sell on the bad news’’ strategies
won’t likely generate significant alpha as news analytics become more widely adopted.
Studies such as those written by Nitish Sinha who looks at the ability to use news volume
and sentiment to predict medium-term alpha (one to five months out) or those by David
Leinweber who looks at portfolio strategies indicate some of the promise from emerging
datasets like news analytics. These techniques can be further refined with more com-
prehensive selection criteria, including weighting various types of news differently and
applying certain filters to the news. These techniques may involve underweighting
market commentary, repetitive news items, less important sources, and alerts vs. articles
depending on the signal timeframe. One might also overweight other types of news such
as exclusives, interviews, strategic news, broker research, and more intense news days.
Additionally, adjustments for market and sector sentiment can highlight relative senti-
ment performance and enhance the signal, particularly during market regime shifts as
seen during the downturn in 2008 and upturn in 2009.
While we are in the early stages of the market’s adoption of news-based signals for
trading and investing, one can clearly see how some simple strategies for enhancing
algorithmic trading with news signals may offer a great deal of promise. For those
thought leaders already implementing such strategies, it’s only a matter of time before
these techniques become mainstream and the ‘‘easy’’ alpha will disappear. For those
who have yet to explore the power of unstructured text, there is a new frontier for alpha-
bearing content.
310 Industry insights, technology, products, and service providers