Quantopian, a community where around 11,000 or so quants collaborate on algorithmic trading strategies, has raised $6.7 million from Khosla Ventures and Spark Capital. The new round brings the company's total funding to $8.8 million. Quantopian says it is the world's first browser-based algorithmic trading platform. Developers can build and test algorithmic trading strategies on the platform. Then, they can share their work and collaborate with others. Over the past 10 years, quantitative and high-frequency trading strategies have risen in power as daily turnover has exploded. Human speed and judgment hasn't been able to keep up with the ever-growing complexity of global financial markets. "People are realizing that anything that is subject to human judgement can be improved by automation and machine learning," said CEO John Fawcett. Fawcett, who used to be a chief technology officer at Tamale Software, a investment research software company that sold to Advent back in 2008, said that he saw an opportunity to support a community of independent traders who wanted to refine algorithmic strategies on their own outside of big funds. "It's directed at the individual who wants to strike out on their own," said Fawcett said. "I saw that there were people practicing as quants today, who wanted to operate on their own but couldn't given compensation, lifestyle and the nature of funds." He added, "The other thing I wanted to do was be part of a movement for an increasingly systematic approach to investing." The company has racked up about a decade of pricing data and history that developers can backtest their ideas against. Right now, it's just U.S. equity data, but it wouldn't be hard to imagine added historical data on currencies, commodities or other asset classes. "We built this cloud infrastructure that has historical market data for U.S. equities and an integrated development environment where you implement two functions," he said. "We feed all the data through those two functions, which can decide when to buy or sell any security." They launched earlier this year and recently poached Jessica Stauth, who used to run Thomson Reuters' quant strategy, as a vice president. So far, the site has about 50,000 algorithms. But they don't have a revenue model yet. They're merely building the community first. Fawcett is critical of the typical fee model where fund managers pick up a set percentage of assets under management and a percentage of how much they return. He says that model incentivizes firms to break up as star investors accumulate resources and start to think about going off on their own. "Having an umbrella organization over many managers never really worked because it's a star driven business," he said. "With the fee or following model, you're taking a little bit of the compensation out of what the fund manager is earning. This basically penalizes success." They're thinking about charging a monthly fee for access to a particular algorithm, but haven't settled on that. "We're still in the phase of building a community, in the same way that AngelList spent a couple years getting access to startups before they built syndicates," Fawcett said. As for investors who want to harness the strategies that quants have built on the platform, Fawcett is still thinking about how much Quantopian should open the kimono and help prospective investors understand what they're buying into. "We envision having progressive disclosures about the way these strategies work for people that are backing them," he said.
and 5 more articles