Seoul-Based Investors Bank on Big Data

Kevin Kane, Jumpgates chief operating officer, on the left, and Kristof Olesch, the firms chairman and CEO.

 

Steve Cha

The traditional Wall Street stock picker tries to keep an eye on a handful of factors that could affect his or her investments, from bond yields and the yen exchange rate to oil prices and monthly consumer spending figures.

But a new class of hedge-fund firms are betting that it can beat these money managers by gathering as much as the world’s data as it can — from satellite imagery of Wal-Mart parking lot occupancy to the heat signatures emanating from oil refineries — and placing rapid-fire bets to take advantage of the hidden relationships between all those data sets.

The approach is part of a recent shift towards relying more on big data and algorithms to gain an edge on rival investors. One Seoul-based company, called Jumpgate Technologies, says it tries to eliminate human involvement altogether, giving its machine-learning technology free rein to discover — and take advantage of — the world’s growing trove of data.

The role for the hedge fund’s human founders? Designing a good system that can make use of the reams of data points, and collecting more streams of data to feed through the program.

Jumpgate, while small, is part of a class of so-called fintech companies that attempts to marry the technological innovations of Silicon Valley with the market savvy of Wall Street.

For instance, Kristof Olesch, Jumpgate’s chairman and chief executive, says he has been programming since he was 13 years old, and investing in the stock market since he was 16 years old. The firm has recruited a number of doctoral graduates in engineering.

One much bigger firm, New York-based Two Sigma Investments LLC, the subject of a page one Wall Street Journal story this week, programs its machines to cull torrents of information from earnings reports, weather bulletins and Twitter.

To allocate its $24 billion in assets under management, Two Sigma’s strategy will generate different investment models based on this data, and then have an algorithm pit the models against one another before executing a trade.

For the first time, these investors say, the world’s computers are capable of storing and learning from the information that machines around the world are gathering, from supercomputers to smartphones and small processors embedded in everyday household items.

Most of those data points may be of little help to a stock investor; sometimes, an arcane data point is just an arcane data point.

But Mr. Olesch is betting that there is more to it than meets the eye. By harnessing the power of computers, he hopes to find on a large scale the informational edge that traditional investors seek out through meetings with company management, meticulous reading of financial statements and channel checks.

“Kodak came to an end because of a technological change, and so will the way of working for asset managers now,” says Mr. Olesch, who says he has about 3,000 streams of data now, which he soon hopes to be able to ramp up to about 10,000 streams.

One example: commercial satellite imagery of store parking lots, which can offer hints about store traffic, but may also have implications for driving habits, weather patterns and a host of other metrics that may not occur to a human fund manager.

Similarly, watching heat signatures at an oil refinery can offer clues about whether the facility is running at full capacity or not.

“The world only processes about 1% of the data out there in the world,” Mr. Olesch says. “We want to get the data that’s out there, instead of waiting for someone to tell us that full capacity has been reached.”

For that reason, much of the company’s focus right now is on what Mr. Olesch calls “technology scouting”: seeking ways to tap into streams of data, either for free or in partnership with companies and institutions that may be sitting on reams of potentially useful data.

So far, Jumpgate, which was founded in Seoul but which is registered in Singapore, says the strategy is working. While his fund is tiny next to larger firms like Two Sigma, Mr. Olesch says the fund has recorded positive returns in each of its first three months, even as its benchmark, the S&P 500-stock index, has faltered