The words “systematic” and “discretionary” have, broadly speaking, defined hedge funds and their strategies for the last three decades. Point72 Asset Management’s Matthew Granade believes that these lines are rapidly blurring.

Technology and the ability to analyze data—in combination with human intelligence—will be the great differentiators of the future. Many would say this has always been the case, but as more data becomes available, and as the ability to store, process and analyze that data becomes less cost restrictive, and as open-source tools and machine-learning techniques become more ubiquitous and democratize analytics, it will force funds to change their strategies and the way they market themselves.

The Netflix prediction engine takes in tons of data about all its users. … It feels like you’re in charge of what you want to watch, but it turns out, you’re not.

Matthew Granade

Simply put, a boom in data, algorithms, and computing power is reshaping the way businesses operate. At this year’s Waters USA conference, Granade, chief market intelligence officer of hedge fund Point72, described the firm’s evolving strategy, calling it part of a new movement toward the “model-driven world.”

In this case, a model is a framework for making a decision, Granade said. It is composed of logic, derived by algorithms and data rather than explicitly programmed rules, and is used to make increasingly more accurate predictions through a feedback loop.

The principle should be familiar—just look at Netflix.

“The Netflix prediction engine takes in tons of data about all its users. It’s constantly recommending things for you to watch. And the amazing statistic we have is that 75% of the content consumption in Netflix comes via the recommendation engine,” he said. “It feels like you’re in charge of what you want to watch, but it turns out, you’re not.”

That’s a powerful concept, Granade added. These types of engines have long prevailed at tech giants like Netflix, Spotify and Amazon, each of which has powerful recommendation engines. Other industries—venture capital, debt collection, health care, and telecommunications, to name a few—want in because the reward is the same: having the ability to collect and interpret data via a model and make the best predictions.

“For a long time, [when] hedge funds started out, they were [for the most part] all people and little machine—[there were] very few models and they were very heavy on art and very light on science. That was certainly true at Point72, which was considered, and is still considered, a discretionary shop; we relied very heavily on the person power,” he said.

But that gap between traditional discretionary investing and systematic is rapidly narrowing, and Granade predicted that “what we’ve been seeing the last two or three years … we’ll probably see for the next year or two years,” which will “really break down that wall” between those binary strategies.

This is also not to say that humans are destined for extinction on the hedge fund trading floor—at least not at most hedge funds. As has been written about quite often here at WatersTechnology, while technological advancements will fully replace some jobs, at the savvier firms, technology will be used to augment what humans can do, making them more efficient and freeing them up to focus on more complex problems. That is how hedge fund managers will need to differentiate themselves in the future.

“Model-driven businesses are the powerful business models of the future,” he said. “[But] humans have very large roles to play given their power and idea generation. Both of those things are going to come together to shape what the future of the hedge fund industry looks like.”

The Way Forward

Granade has been on both ends of the investing spectrum, having worked as a quant at Bridgewater Associates prior to Point72. When the quant funds entered the scene, their strategies diverged from the discretionary shops, each with a bias and narrowness to their approaches. There is another reason, though, why these worlds are colliding.

Hedge funds have not escaped the vicissitudes facing the rest of the buy side—regulation, shrinking margins, and the rise of passive investing. To compensate, firms ranging from systematic/quant funds to more traditional discretionary shops are adjusting their trading strategies and this shift will change how hedge funds are evaluated and what separates one shop from another.

Granade said that while Point72 has relied mostly on a team of seasoned portfolio managers and analysts, it is increasingly incorporating “machines and algorithms to trade stocks algorithmically and keep their books mathematically constructed at all times.”

The thinking—“people plus machines”— is new, even if obvious. This approach combines the staff’s greatest talents—working with sometimes scarce data, defining and redefining goals, and asking the right questions—with the strengths of the data science world, which include automating repetitive tasks, processing vast datasets, and spotting patterns consistently.

As examples, the firm is moving its policies and systems around proprietary data, training, and tooling. The finance and accounting elements are table stakes, Granade said, but the differentiators are—and increasingly will be—data, programming, and information.

Specifically, the fund is building out its capabilities for handling all the new sources of alternative data by using machine learning to tag relevant data points, the hoped-for results being better signals, surer bets, and less data-flooded traders. The shop is also exploring the idea of running a system or model on top of idea input from portfolio managers and analysts to correct common human errors such as leaving a long or short position on for too long, trading too aggressively, or not balancing the portfolio for factor considerations.