Wednesday, January 23, 2013

Quantopian

Niche, vertical social networks are fascinating because they are such internet-native businesses.  The internet is the perfect place to have that "Wow, I thought I was alone in my obsession with ________, but there’s thousands others like me!" moment. 

Vertical networks can be great businesses, particularly when the members of the network are all highly-skilled people themselves. At StackExchange the best answerers in the community are all smart enough that they could command $800/hr consulting fees for their expertise. These highly talented technologists give their knowledge away for free on Stack in exchange for the social currency and cred they receive from other community members. And that social currency helps amplify their personal brand, which they can leverage into thought leadership in their community and even larger financial rewards (high paying jobs). Successful vertical networks like StackExchange have a wonderfully altruistic network effect that ties its users together.

It is through this lens that I felt a great connection when I first met John Fawcett (Fawce), the founder of Quantopian.  Fawce painted a very compelling picture of why there is not a definitive community for quantitative traders online today, and why Quantopian could become that community.

Fawce had built an open-source backtesting engine that made it very simple to create a quantitative trading strategy.  It was a Python IDE in the cloud that required zero setup.  A developer on the Quantopian platform need only use *two* functions to implement a trading strategy.  And then executing the strategy over years of historical equities trading data was just one button click. The simplicity was unprecedented. In my due diligence I could not find anyone else that had open sourced a high-quality backtesting engine, despite the fact that backtesting engines have been created from scratch by developers in nearly every hedge fund and prop trading firm that has ever existed.

A hedge fund has no incentive to open source the tools they develop, like a backtesting engine or an algorithm management and execution environment.  But if your goal is to build a community of people interested in quantitative trading, creating the easiest place to get started developing algorithms is a great way to start. We hope our open source backtesting engine, Zipline, becomes a standard template for all new emerging hedge funds in the future (16 forks already). But Quantopian’s real long-term goal is to build a community, and any community needs a means of communication and collaboration.

So the next stage of the business was to build out forums where users could post the results of their algorithms (and their source code), so other users could comment on their work, make suggestions for additions, or simply fork the code and build the next phase of the algorithm themselves.

Out of this simple combination, a merger of the creation process (a cloud-based Python IDE and backtesting engine) and the forums (a place for people to discuss their performance and code) a community of quantitative traders has emerged.  

The community is still young, but it’s developing a voice, and its members are surprisingly collaborative.  I’ve heard plenty of naysayers claim that quantitative traders would never discuss their work with each other online because they live in a cult of secrecy and punitive IP agreements.  I agree, a quant trader at Renaissance Technologies is an unlikely member of Quantopian.  We estimate that the “RenTech”s of the world employ about ten thousand active quants across all firms.  But, there are hundreds of thousands of people around the world that are interested in quantitative trading who simply don’t have access to the tools and the data. For every 1 RenTech trader, there are 10 IIT grads in India that are just as smart, if not smarter, than the RenTech trader, but don’t know where to begin.  The classic user is the Physics PhD candidate that is competitive by nature and interested in experimenting with algorithmic trading: the highly-talented, aspirational quant.  We’ve been amazed by their interest in collaborating with each other online. They finally now have a place online to have their Wow, I thought I was alone in my obsession with ________, but there’s thousands others like me!” moment.

Forbes has a great summary of the company that’s worth a read. Quantopian also discusses their emergence from beta on their blog. And best of all, here’s a picture from the second Quantopian meetup held in the Spark Capital NYC offices just last week. It was packed:

Quantopian Meetup

I have joined the board of the company and am delighted to be working with John Fawcett, his co-founder Jean Bredeche, and the rest of the Quantopian team. I’ve been impressed by advice and support from our co-investor in the seed round: GETCO.  Now that Quantopian has emerged from beta and is publicly announced their seed round, I’m glad I’ll feel free in the future to talk more about why I find this investment so exciting.  

(photo credit: Dan Dunn)

Notes

  1. thegongshow posted this