Inexact Science: When, If Ever, Should Intuition Trump AI?

The importance of data scientists in finance seems to grow every day with more and more firms adopting the use of big data in their trading strategy.

For many, the days of incorporating any type of emotion or intuition into trading is long gone. Algorithms and analytics rule how some shops choose to work.

And with this evolution in how the markets now work, a question has arisen: Is finance no longer a business problem so much as a science one?

That debate was brought up at a panel discussing how to optimize data strategy to enhance trading performance at the Chicago Trading and Technology Summit 2015, which was held on the 80th floor of the Aon Center. Two panelists had particularly strong takes on the subject.

Lauren Crossett is a firm believer that finance comes down to science. It's not surprising: Crossett is the director of business development for Rebellion Research ─ a registered investment advisor (RIA) that relies on artificial intelligence (AI) for its investment decisions.

She joked that due to the nature of the firm, people have emailed her asking if she's a real person. For Crossett, finance is a science game.

"We think that systematic is the way to go. Every study says that emotions are bad, I think that's pretty much accepted," Crossett said. "Everything is down to a science problem and finding out if your numbers are good and things match up."

A+B=C?

But not all in attendance agreed. Andrew Kurmiega, adjunct professor at the School of Applied Technology at Illinois Institute of Technology, argued instead that finance is a more of a softer, social science.

Kurmiega said chemistry, physics and mathematics all represent true, hard sciences. Those areas, unlike finance, have definitive answers to problems.

"If you do chemistry and put A with B it's always going to equal C. In this world, you put A with B and you're 80 percent sure that maybe you'll make some money. ... People make the mistake that it's a hard science. I don't think it is."

For Kurmiega, a huge part of the process is engineering. While the structure for an idea is based around science, the network is built through engineering.

As an example, Kurmiega said he's supervised several PhD mathematicians in the field. Some have been good, while others have been "absolutely horrible." The problem, according to Kurmiega, was that many of them believed there was some type of equation that would explain how things work.

"There is nothing in Black-Scholes model that definitively says, 'this is the price of an option,'" Kurmiega said.

Little Bit of Both

In the middle of the debate between Crossett and Kurmiega was Drew Wade, senior managing partner of Chicago-based hedge fund AIA Group. Wade said the rise of data has led to data scientists having increased importance.

While masters of mergers and acquisitions or traders have always held important roles in finance, Wade said, the profile of those who can measure, assess and analyze data has risen significantly.

But Wade still seemed to side with Kurmiega. "I think it has a scientific piece to it, but ultimately it's still that business side that matters," Wade said.

Kurmiega added that there has never been a system that produces definitive answers all the time.

"They're building machines that are very clever, but at the end of the day there is still some error term," Kurmiega said. "Every model I've ever used has had an error term. And the question is around whether they like the error term, or do they not."

Not Perfect

Crossett admitted that the systematic approach isn't perfect. There are errors, but the key is to stick with the AI's decisions.

Crossett said within the Rebellion Research offices, there is an ongoing joke whenever a stock it owns is falling about whether they should override the AI and sell it. And while Crossett said that might help the drawdown number of the firm, deviating from what the system says isn't part of the plan.

"We might look at all the other funds that sold out when something starts to go down and say, ‘Ok, well we know where the fork was in the road. Let's see how it comes out," Crossett said. "We pretty much depend on what the AI said. It's not to say if we think there is something wrong with the data we're not going to rerun the system. But we're not going to say, ‘Hey, I have a good feeling about this, so let's overweigh something.' We're not going to do that."

The Bottom Line

• As the importance of data has risen, some question whether finance is more about science than business anymore.

• Some say studies have shown emotions negatively impact trading and being successful in the capital markets is all about science.

• Others say it's more of a social science due to the fact that there are no definitive solutions like in typical hard sciences.