Earlier this year we saw the launch of two experiments in investment crowdsourcing. One was StockStream, an online multiplayer stock-trading game that lets anyone with an internet collection vote minute-by-minute on what names the fund should buy or sell. Seeded in late May with $50,000 by the Amazon employee who created it, the platform was bombarded in its early days by a mob of rabid Tesla longs but has since diversified into more than 200 stocks.
The other marriage of markets and crowds was the somewhat more serious startup Quantopian, which has received ten-figure infusions of cash from the likes of Steve Cohen and Andreessen Horowitz. It works by allowing hundreds of thousands of amateur quants to post their own algorithmic trading code to the system, which then whittles them down to the most promising strategies. In June, Quantopian's investment team assembled the best of the best into an honest-to-goodness hedge fund and put real money behind them.
The performance four months later? The mindlessly democratic online-sentiment wind-sock StockStream is up a solid 8 percent, beating the S&P's 6.6 percent. The glossy billionaire-backed startup Quantopian, meanwhile, is down around 3 percent, and things aren't looking good:
About two months ago, Quantopian’s chief investment officer, industry veteran Jonathan Larkin, left the firm, the people say...Late on Tuesday, the Boston-based firm sent an email to its community of algorithm authors informing them of Mr. Larkin’s departure and saying “the fund is being managed on an interim basis by other members of the investment team…the journey to creating something great is never easy.”
The comparison admittedly might not be all that fair. By design, StockStream encourages an endless variety of trading approaches and investment theses, resulting in what is essentially a tech-heavy index fund. Rather than reflecting any unique timing or underlying strategies, StockStream's market-beating performance probably just boils down to having an outsize tech exposure over a period when tech stocks outperformed. It remains to be seen whether StockStream's hordes of anonymous users will adjust profitably if and when tech starts underperforming.
Quantopian's challenges are more complex:
Quantopian is a young firm and could yet turn its fortunes around. But the challenges so far suggest that recruiting and training mechanical engineers, nuclear scientists and other amateurs around the globe to be quantitative traders may be more difficult that some had expected. While some of Quantopian’s traders generated profits, many of these algorithms had limits to how much money could be invested with them, the people said.
It's understandable how models developed in laboratory conditions by non-finance engineers might struggle in the wild, but market impact isn't exactly a niche problem for quants; it's one of the central limitations to the field. There's a reason RenTech has been closed to new investors since 1993: scalability. Here's a Quantopian blog post that lists the issue as one of the top mistakes a quant can make.
This doesn't mean Quantopian's model is bunk. Hedge funds are hard, quant funds even more so. Yet you have to wonder whether Stevie Cohen has any regrets about putting that $250 million into Quantopian rather than some Damien Hirst nonsense.