How Artificial Intelligence Can Give Equities Analysts an Edge

If you’re an equities analyst, it sometimes can be hard to see the forest for the trees when you’re doing your job. You might be deep into Excel spreadsheets, trying to make a column add up, when you lose sight of the bigger picture. To keep on track and gain the upper hand, you need to act strategically, especially in situations of great uncertainty, which is what we’re all faced with when doing any kind of public and private equities analysis.

The Air Force vs. Equities Analyses

The OODA (Observe, Orient, Decide, Act) loop, developed by military strategist and United States Air Force Col. John Boyd, is a system for winning dog fights that has many applications across the business world. The concept has four components: observing the environment, orienting yourself with respect to what you see, deciding what to do, then acting. Boyd’s conjecture was the faster and better you can go through this cycle, the better you are as a pilot.

The OODA loop has been adapted for business analysis into three essential steps: understand what’s going on, and once you understand it, wargame out your options and act. The idea is to do this faster and better than competitors. You want to have a deeper level of insight than your competitors and have more confidence that the outcomes you’re predicting will actually come to pass. Automatic execution at great speeds is the recipe for winning.

The Future of Equities

But when people predict the future of equities, they tend toward optimism. Sometimes consensus effects drive this: individuals are often in competitive environments where the people who are advanced are the most optimistic. That shows there are human biases that affect predictions, meaning there’s no objectivity.

Billionaire investor Ray Dalio, describing what you should be doing whenever you’re analyzing a stock or a portfolio, wrote, “Think for yourself to decide 1) what you want, 2) what is true, and 3) what you should do to achieve #1 in light of #2.” This means that when you say something’s true, also understand that there’s a high probability that it may not be.

How do you challenge your assertions? Seek out the most credible people who disagree with you and ask them to challenge your thinking. This is where artificial intelligence (AI) plays a fantastic role because it can be objective and also challenge your thinking. More than that, it can supply reasoned justifications behind its predictions.

AI in Analysis

That points to the major constraint of using AI in analysis: people still believe that people are best at forecasting the future. This is a serious impediment to the use of AI because it leads to a lack of confidence in deploying technology that could improve forecasts immeasurably.

Most people think about AI as a branch of machine learning, which functions by scooping up data and finding the mathematical function that best fits the data. Our model is a data-centric paradigm because the essential idea that’s embedded in it is that the future will be some version of the past, and the reason we’re looking at past data is because we think the future will repeat that pattern in some sense.

This is different from the idea of hypothesis-driven forecasting, in which you don’t need any data at all. Instead, you create a hypothesis about the causal chain. My favorite example of this is an experiment that astronaut Commander David Scott Apollo 15 conducted on the moon. Holding a feather and a hammer, he said, Galileo predicted that in an environment where there’s no air friction, the feather and the hammer will hit the ground at the same time.” When Scott attempted this on the airless moon, both hit the surface at exactly the same time. That’s an example of a prediction that did not rely on past data. The only past data available had the bias of being collected on Earth, where there is air resistance.

Predicting the Future

If we all knew the future, we’d all be great equities analysts. We’d know where to place all the chips on the roulette table. The world is driven by causes and effects. If we only understand the causes that create the effects, then we can predict the effects by knowing the causes.

Predicting the future depends heavily on understanding the agenda of the most powerful actors in the system. So, for example, understanding the actions of central banks and the logic that drives their decision-making process to the point that you can predict their future actions is essential to comprehending the future behavior of markets.

This methodology (known as Epistemology’s “Five Rings,” implemented using the SKAI platform) proceeds as follows: we start with what we know and gain a deeper insight about what’s true using causal science. After we understand what’s true, we wargame scenarios to understand what the expected future is, then position ourselves optimally with respect to that future. Finally, we take the optimal action.

Normally, this cycle is done by most investment houses within a three- to six-month window. The idea behind using AI is to speed this up to a week, a day, an hour, a minute, and this is where your ability to get super competitive and gain a trading edge occurs.

---

About James Ward

James Ward works with senior corporate executives to develop stronger approaches to strategy, innovation, marketing, sales, and profit through scientific, evidence-based decision making to better meet the targets set by boards and shareholders. He’s a specialist in the application of causal inference and Bayesian networks (including Bayesian machine learning) to drive excellence in corporate strategy formulation and to improve the profit and operational performance of cross-functional enterprise teams, especially in the areas of customer engagement and customer experience.

---

This technology industry article is adapted from a GLG teleconference. If you would like access to events like this or would like to speak with technology industry experts like James Ward or any of our more than 900,000 industry experts, contact us.