When world chess champion Garry Kasparov lost to Deep Blue, many felt it was the first sign of the last days for human intelligence:
…losing to a computer wasn’t as harsh a blow to me as many at the time thought it was for humanity as a whole. The cover of Newsweek called the match ‘The Brain’s Last Stand.” Those six games in 1997 gave a dark cast to the narrative of “man versus machine” in the digital age, much as the legend of John Henry did for the era of steam and steel.
Nowhere is this anxiety more prevalent than in finance, where headline after headline have crystallized a narrative of a mechanistic overthrow. If Business Insider is to be believed, it is only a matter of time before finance is entirely soulless automatons, all buying and selling and evaluating heaps of data with nary an analyst in sight. “But,” as Gary Kasparov notes,
it’s possible to draw a very different lesson from my encounter with Deep Blue. Twenty years later, after learning much more about the subject, I am convinced that we must stop seeing intelligent machines as our rivals. Disruptive as they may be, they are not a threat to humankind but a great boon, providing us with endless opportunities to extend our capabilities and improve our lives.
To that point, there is more to the story of artificial intelligence and chess. In 2005, two chess enthusiasts, Steven Cramton and Zackary Stephen, entered a freestyle chess tournament that allowed for computer-aided play. The two had spent months honing their cyborg strategy, and their diligence worked:
…they won the tournament, leaving grandmasters and some well-known programs in their wake. It was quite a shock but it proved the theory worked: certain human skills were still unmatched by machines when it came to chess and using those skills cleverly and co-operatively could make a team unbeatable. Humans playing alongside machines are thought of as the strongest chess-playing entities possible.
What’s true for chess is also true for the equally strategic world of finance: the future isn’t artificial intelligence eliminating human analysis; it’s artificial intelligence augmenting human analysis. The firms that will thrive in the coming years will be those that are able to identify and synthesize the best of artificial and human intelligence.
It’s obvious that machines outperform humans at a number of tasks. Machines can, for instance, quickly and comprehensively collect and sort vast quantities of numerical and linguistic information. Machines can also produce quantitative evaluations of qualitative data and can do so without bias. That said, it’s also obvious that humans maintain our fair share of advantages: curiosity, creativity, assigning value, and assessing other humans are just few.
A synthesized approach to financial decision making is precisely what Prattle’s analytics allow for. And, perhaps unsurprisingly, this approach yields compelling results. For instance, over the last year Prattle’s monetary policy calls for the G-10 currency central banks were right 97% of the time, outperforming market consensus (90%).
Correct & with Consensus
Correct & against Consensus
Incorrect & with Consensus
Incorrect & against Consensus
Prattle’s forecasting success is a testament to the effectiveness of human analysis informed by automated evaluation.
The automated portion of Prattle’s approach includes an algorithm that quickly and comprehensively collects and sorts vast quantities of linguistic information. Currently, that set includes every publicly available communication from the 15 most important central banks in the world and earnings calls for almost 4,000 publicly traded US equities.* Our algorithm also produces unbiased, quantitative evaluations (scores) for each and every one of these market-moving communications in real-time. These scores represent, in the case of central bank communications, the communication’s likely impact on the market or, in the case of corporate communications, the call’s likely impact on the company’s stock price.
This novel data set equips human analysts with a real-time feed of structured, quantitative data where there once was a mountain of unstructured, qualitative information. The trends that emerge in the data—and the outliers thereof—help analysts quickly find and focus in on important patterns in a specific security or in the market as a whole. This process dramatically expands the coverage capabilities of an individual analyst, whose average capacity, according a recent Integrity Research Survey, is just 15 stocks. With the grunt operations taken care of by our algorithm, human experts, with a deeper understanding of the underlying business factors, are now free to perform the detailed analysis that only they can.
With the AI boom on the horizon, it’s easy to get uneasy about how automation is going to reshape finance. But it’s important to remember that much of the history of progress is actually the history of machines allowing humans to move away from the repetitive and the menial and towards the creative and the meaningful.
The story is no different for the financial space. Technologies like Prattle aren’t going to make human analysis a thing of the past—they’re going to make human analysis better than ever. Analysts who could once only cover 15 stocks will be now be able to cover 150. Analysts who would once spend hours tediously sifting through documents will now be able to immediately identify what merits their attention. These are just a few of the ways our automated research platform is enhancing financial professionals. At the moment, only cutting-edge firms are integrating machine learning and AI technology into their processes. Soon this approach will become the status quo.
The Prattle Team
* Soon to be over 10,000 global companies.
Copyright © 2017 Prattle Analytics, LLC.