By Efi Pylarinou

Quantopian, founded in 2012 in Boston, started in the Algorithmic trading space. It offered a web-based platform to write algorithms for US equities trading. In addition, Quantopian offered the capability to paper trade against live stock market data and – most importantly – to back-test the algorithm against historical data. So far, that describes most competent algo trading systems. Quantopian differentiates by adding a crowd-sourcing dimension. The coder of the strategy can invite other collaborators that can use or enhance the code. This is the open source aspect that Quantopian brought to an otherwise highly secretive market (proprietary algorithmic trading).

Quantopian is democratizing the secretive quant trading algorithms. It is making it easy to code (using Python), back-test and tweak the strategy. Back-testing is a serious pain point for quants. Quantopian mitigates this challenge and also creates a community to share (at each trader’s discretion) algorithms, strategies and ideas.

Recently, Quantopian enhanced their historical database (prices originally starting from 2002) with 12 years of historical data from Morningstar Corporate Fundamentals. This includes over 670 financial metrics across more than 8,000 stocks (raw values of revenue and earnings, as well as convenient pre-computed values and ratios such as market cap, earnings per share (EPS), price-to-earnings ratio (PE) and more).

Typical quant strategies include mean reversion, momentum, valuation, sentiment and seasonality. Under each of these broad strategies, traders can include their own rules that aim to produce alpha. At the end of the day, of course, traders want to produce consistently superior risk/adjusted returns compared to a comparable passive strategy.

As a Quantopian user, one can see the open source algorithm library, copy any one tradable coded strategy and maybe tweak it.

For example, think of a simple strategy like; BUY when the price is above the 50 day SMA and SELL when it is below. The original coder could have back-tested this strategy for a few ETFs (subsectors of S&P500) and I would like to back-test it for Russell 1000 and depending on the results, tweak the SELL order with a stricter condition (price is below the 50day SMA and the 100day SMA). In addition, there is a community to chat with about the nuances of the strategy.

Recently, Quantopian created their own hedge fund. This is where it gets interesting and disruptive.

Quantopian is a crowd-sourced hedge fund.

They pick the top performing managers from their quant community and subsequently connect them with investor capital. There is a trading contest for the selection process that has its guidelines and restrictions (e.g. leverage limited to x3 times, max drawdown less than 10%). Each month Quantopian picks a trader/coder winner who gets rewarded with $100,000 of investment capital to be managed free of charge (i.e. manager keeps 100% of profits) for 6 months.

Quantopian recently organized a disruptive quant trading event in NY called Quantcon. It included innovative talks and workshops, geared to improving investment performance by exploring algorithmic trading strategies and applying open-sourcing to investment ideas. Speakers included entrepreneurs, tech specialists, academics, quants, and investment managers.

This past week, Barry Ritholtz on his “Masters in Business” radio podcast, featured Meb Faber, chief investment officer of Cambria Investments. Faber explained how his investing approach has evolved from his days as a biotech analyst to becoming a quant “lite.” Faber is a disruptive quant, who this past December created the first no-fee global allocation ETF. There are a few Quantopian tradeable strategies based on Meb Faber strategies.

Quantopian has raised total capital of $23.8 million in 3 rounds. Investors are Bessemer Venture Partners, Khosla Ventures, Spark Capital, and Wicklow Capital