Two Sigma Taking Crowdsourcing Approach To Trading Algorithms

Quants' job security may not be so secure after all.
Author:
Updated:
Original:

The efficient markets hypothesis has been around longer than the average hedge fund trader has been alive. As the theory goes, swarms of market participants immediately incorporate economic data and business news into asset prices, making it tough for your average trader to edge out average returns. To outperform, you have to be remarkably savvy, uncommonly well-connected, or – as successful funds like Two Sigma, Citadel and Renaissance prove – a computer.

two-sigma-gif

Bollocks to all that, says Two Sigma. Bloomberg has the story:

“There’s a shortage of folks looking at economic data,” Kartik Raghavan, who runs Two Sigma’s business innovation and growth team, said Tuesday at the firm’s New York headquarters. “So the question we’re posing is, can we get more people interested in this?”

The solution: Through a partnership with a company called Kaggle, which pits users against each other in data engineering contests, Two Sigma is offering $100,000 in prizes to a pool of 730,000 coders to craft the best predictive trading algorithm. Because evidently there aren’t enough computers out there trading financial assets at the moment.

David Siegel, the Two Sigma co-founder who helped build his firm on learning algorithms, has lamented the limitations of AI. “Artificial intelligence today doesn’t have anything that resembles common sense, and common sense is a key feature of intelligence,” he said earlier this year.

It’s only appropriate, then, that Two Sigma is addressing one rising tech trend, AI, with another: crowdsourcing. The economics might just work out pretty well, too. The $100,000 in prizes would be shared between seven winners. If any one of those algos generates significant revenue, the $37 billion fund has plenty of capital to devote to it.

The only question is how all of this makes Two Sigma’s current quants feel. What happens when some Starbucks-dwelling hacker sends your boss a few lines of code that outshine the algo you get paid a fortune to write?

The dynamic is like a recapitulation of the last round of disruption in financial speculation. Human traders have seen themselves increasingly replaced by quants, who whip up codes that do their jobs for a fraction of the price. Could quants soon see themselves replaced by crowdsourcing networks similarly capable of doing their jobs for fractions of the price?

Two Sigma’s current coders will be the first to find out.

CORRECTION: Two Sigma will not be putting the winning programs into practice – this is more of a recruitment thing. “The algos created through this competition will not generate revenue as they are not going to become part of Two Sigma’s current models,” a spokesperson clarified.

Related