Honey Badger Hedge Fund: Hackers Predict Stock Market With Open Source Mojo

Somewhere, out on the web, there’s a secret Twitter account that’s one hell of a stock picker. So far, the thing has recommended seven stock trades — all to a private group of traders — and each one of them has turned a profit.
From left to right Andrew Delikat TaeHwan Jo and Brian Weiden creators of the Honey Badger algorithm. Photo Hack Reactor
Andrew Delikat, Tae-Hwan Jo, and Brian Weidenbaum, the coders behind the Honey Badger algorithm.Photo: Hack Reactor

Somewhere, out on the web, there's a secret Twitter account that's one hell of a stock picker. So far, the thing has recommended seven stock trades -- all to a private group of traders -- and each one of them has turned a profit.

The account isn't run by a bunch of Ivy League MBAs scheming their way up the Wall Street ladder in lower Manhattan. It's driven by a computer algorithm cooked up by three guys at a hacker school in San Francisco. They call it the Honey Badger algorithm, after that ongoing internet meme that espouses going your own way. And it shows that, thanks to the rise of open source software and cloud computing, anyone can compete on Wall Street. At least for a time.

Andrew Delikat, Tae-Hwan Jo, and Brian Weidenbaum built their Honey Badger stock picker at Hack Reactor, a 12-week bootcamp for programmers. All three had backgrounds in economics and technology, and they joined the program to shore up their programming skills. Hack Reactor students work 11 hours a day, six days a week on projects meant to further their understanding of computer science -- such as massive peer-to-peer systems that set new records for solving weird math puzzles.

It was Jo who suggested the trio combine their common interest in finance and programming to build Honey Badger. The original idea, explains Delikat, was to give anyone access to the sort of market analysis software available to large financial firms. But their algorithm worked so well, Delikat and Jo decided they should use it as the basis for their own hedge fund (Weidenbaum is leaving the team, after accepting a job at Google).

That's why the account is secret. We've seen the thing in action, but agreed not to reveal its location. The hackers' new fund -- known as Honey Badger Capital Management -- has already lined up soft commitments from several investors.

Yes, seven out of seven of the trades the team recommended on Twitter were profitable, providing a return of 9.38 percent. That's an impressive streak, but it's an awfully small sample set, and Delikat and Jo admit that the algorithm isn't perfect. "It's not something we can ever hope to calculate with 100 percent accuracy -- just because there's so many variables involved with human behavior," Jo says. But they're confident they can consistently beat the market. They say they've made many more trades than just those shared on Twitter, and almost all have made money.

This sort of thing isn't exactly unique. As Jo points out, there are many tech-driven methods -- such as high frequency trading and momentum trading -- that reliably beat the market. It's entirely possible, Jo concedes, that big Wall Street firms are already using a similar algorithm as part of their strategy. But something along the lines of the Honey Badger method, he says, has never been available to everyday investors.

Delikat explains that the algorithm crunches many terabytes of data in an effort to make comparisons across historical performance on Wall Street. This would have been prohibitively expensive a decade or two ago, but today, with the rise of cloud computing, it's relatively cheap.

Most of the Honey Badger platform is written in Python, an open source programming language popular with mathematicians and web programmers. And the team stores and processes its data with a combination of Hadoop -- an open source clone of Google's big data crunching system -- and the tried and true open source database MySQL. The team pays Amazon and Microsoft Azure a few thousand dollars a month for cloud hosting -- a bargain compared to what they would have had to pay upfront for supercomputers ten years ago.

The team hasn't completely abandoned its desire to bring high end analysis to the masses. "One of the things we're trying to do is not be an extremely exclusive hedge fund," Delikat says. "And we think we may eventually release the consumer product that we intended to."

Ultimately, the hedge fund is just a means to an end, a way of financing the team's other startup ideas. "We love working on [Honey Badger]," Jo says, "But we have other ideas." He says they're considering a move to China or Singapore, or possibly joining an incubator in San Francisco.

One idea is for a new type of dating site, but frankly, that sounds kind of lame. Bill Gates is using his fortune to try to cure malaria. Jeff Bezos bought one of the country's largest and most respected newspapers out of his own pocket. Elon Musk is trying to send people to space, reinvent the automotive industry and build a high speed rail between LA and San Francisco. Can't these Honey Badger guys think of anything more ambitious than a dating site?

Well, they do have another idea, but it's a bit cryptic. "It's called Antecedent, and it's a crowd thinking platform for knowledge synthesis," Delikat says. "But we might have to change the name of that one." Understood. But let's hope the name of their hedge fund never changes. Whether it succeeds or fails, Honey Badger Capital Management is a great name.