Technology Review, a publication of MIT, published an article today with the headline, "AI that picks stocks better than the pros."The story describes a computer program designed by college computer science professors which analyzes the text in news articles found in Yahoo! Finance about various stocks and then buys them or shorts them based on the news it finds*. When they tested it out, according to the author, they found that "it appears to be able to beat Wall Street's best quantitative mutual funds at their own game."
Before investors run out and buy a computer that can "beat Wall Street" and stop researching good companies themselves, a few important points are worth considering.
First, according to the story, the maximum period of time this black-box strategy holds a long position or a short position is . . . twenty minutes! Talk about having a short-term focus! Trading at that frequency would be extremely expensive in terms of both commissions and taxes and impractical for a real investor.
More importantly, we believe that the criticisms we have raised about "black box" investment strategies in the past are reinforced by this new story.
Note well the line in the story in which it is revealed that, in order to test their computer investment strategy, the college professors selected "five non-consecutive weeks in 2005, a period chosen for its lack of unusual stock market activity."
We have pointed out before that backwards-looking, mathematics-driven investment strategies, such as the "quant" or "black box" strategies of the computer age or the "various systems" of market prediction that T. Rowe Price wrote about all the way back in the 1930s, usually fail (in the words of Rowe Price) "at crucial turning points in the market."
It's great to have a sophisticated investment strategy that "beats the pros" if you can manage to only use it during "periods chosen for their lack of unusual stock market activity." This is exactly why so many quant strategies were wiped out by the "unusual activity" of 2008 - 2009.
We would point investors to our post from last year entitled "The Dismal Science" in which we discuss the real problem with this kind of thinking, and the real way out for investors. The problem is that "no mathematical model can predict the next entrepreneurial development that will create new markets and drive economic activity in a completely different direction."
It is these unexpected innovations, and the companies which develop them or take advantage of them, which create real wealth and growth. This is the message to which we constantly return in our posts on various topics, and we believe it is the real investment truth which our readers should never forget.
Subscribe (no cost) to receive new posts from the Taylor Frigon Advisor via email -- click here.
For later posts on this same subject, see also:
For later posts on this same subject, see also: