AI's Latest and Greatest

Hedge Fund Driven by Anonymous Ideas Gets New Investments


Photo

Numerai received $6 million in new investments from a group including Union Square Ventures, headed by Fred Wilson.

Credit
Lucas Jackson/Reuters

SAN FRANCISCO — A hedge fund powered by ideas from anonymous contributors is attracting some prominent investors.

The San Francisco-based hedge fund Numerai is to announce on Monday $6 million in new investments from a group including Union Square Ventures, the New York-based venture capital firm led by Fred Wilson.

Union Square Ventures is joining Numerai’s previous partners, which include Peter Diamandis, a founder of Singularity University, and Howard Morgan, a founder of the unusually successful hedge fund firm Renaissance Technologies.

Numerai, created by a 29-year old former quantitative trader, Richard Craib, makes its investments based on models submitted by data scientists, who are allowed to remain anonymous and who are paid in Bitcoin based on the success of their models. The models all employ machine learning, which allows algorithms to integrate and adapt to new information over time.

“This is likely the largest ensemble of stock market machine learning models in the world,” said Andy Weissman, a partner at Union Square Ventures who is working with Numerai.

Since starting last year, Numerai says it has given $150,000 to its contributors and is aiming to eventually pay out $1 million a month.

Numerai is the latest investment fund that has won backers with the promise of sourcing investment algorithms from the crowd.

Perhaps the most successful of these new ventures, Quantopian, is already onto its Series C fund-raising round. Quantopian has attracted hundreds of millions of dollars from the venture capital firm Andreessen Horowitz and the new investing firm run by Steven A. Cohen, Point72 Asset Management.

With Quantopian, data scientists submit investing algorithms that asset managers around the world can license and use to trade. The algorithm’s authors get a cut of any profits. The investment ideas are not predictions of specific price patterns, but instead are mathematically based models that can be applied to different investments.

Numerai is different from similar funds because its contributors remain anonymous and prove themselves through weekly “tournaments” held by the fund manager.

Contributors build their mathematical models using historical stock pricing data put together by Numerai. Their models are then tested on live trading data.

Numerai builds its own investment model by aggregating ideas from contributors and weighting them according to their success. Unlike many so-called quant funds, Numerai generally buys and hold stocks for months rather than for minutes or seconds.

“We can have very high conviction that a model will generalize to live data because of how we have framed the problem as a pure machine learning competition,” Mr. Craib said.

Because it pays in Bitcoin, Numerai can send money to contributors anywhere in the world without collecting personal information, though this raises the risk that Numerai ends up doing business with unsavory people.

Mr. Craib seeded the fund with $1 million of his own money — he was previously a so-called quant at a large asset manager — alongside outside investors. He would not say how large the fund is at this point. But he did say that 7,500 data scientists have submitted prediction models so far, leading to 2,675 payouts to contributors.

Numerai maintains a leader board of its contributors. Someone with the screen name WSW is at the top, with $10,434.47 in earnings so far, while PIOTREK has pulled in $9,259.66.

Mr. Craib said that if the fund reached $1 billion in size, it should be paying contributors $1 million a month.

Continue reading the main story

By NATHANIEL POPPER