Quantopian Brings Algorithmic Trading To The Masses - Forbes

High frequency trading has been scrutinized in recent years because of its links to financial scares like the Flash Crash. But the actual algorithms used to power much of high frequency trading are not well understood, and locked up in hedge funds and large financial firms.

Startup Quantopian, which is coming out of beta today, aims to make these algorithms available to a much larger audience. The company, which has raised an unspecified seed financing from Spark Capital and market maker GETCO, provides browser-based tools for anyone with basic programming knowledge to create algorithms and simulate actual trading with them. John “Fawce” Fawcett started the company in 2011 and brought in cofounder and CTO Jean Bredeche, who he previously worked with at Tamale Software and Advent Software.

Quantopian provides market and company data for people to create their algorithms. They write the code on Quantopian’s site, then can run the algorithm on the site’s back-testing simulation to see how they would’ve performed historically. All algorithms are private by default. Quantopian does not have connections to live trading yet, but eventually plans to add that so traders can use their algorithms to trade through third party brokers. Quantopian has made its back tester tool open source to enable others to assist in developing the tool.

People who trade stocks and other securities based on algorithms and other quantitative methods–the “quants”–are a relatively small number of about 10,000 people worldwide, says Fawcett, founder and CEO of Quantopian. Most of the large firms have built their own proprietary systems to manage their own algorithms. As a result, up until now, only those large firms have had access to the ability to quickly make and test these types of algorithms. “We think that’s inherently dangerous for the population at large,” he says. “Things like the flash crash are amplified by the small number of people involved.”

People who use Quantopian come from a broader variety of backgrounds. Some are not in finance but are just interested in testing a trading idea. These people could be a physics Ph.D.’s, other academics or math or engineering specialists who may not even know programming. The Quantopian platform is designed to do for algorithmic trading what Heroku did to make Ruby on Rails faster and easier to build web applications, says Jean Bredeche, cofounder and CTO at Quantopian.

The types of strategies people could develop on Quantopian could include technical strategies, a classic example of which would be when a 50-day moving average crosses a 20-day moving average. The other type of strategies are more statistical analyses in which people set some criteria and try to find certain securities that match that criteria. These are just basic examples, but people can create whatever algorithms they want on the platform.

The company now plans to develop the ability to integrate live trading and also develop its community features. The site’s current forum enables people to ask others for help in writing an algorithm and has become quite active, Fawcett says. The site also provides sample algorithms that can be “cloned” and changed. Quantopian currently doesn’t offer tools for consumers or average investors to use, but that could be a natural move in the future.  ”As Quantopian becomes successful it will blur the line between someone who’s a consumer and a professional for quantitative trading. That will allow the best ideas to emerge,” says Andrew Parker, principal at Spark Capital.