The Tinder for Markets Is Run on Crypto
(Bloomberg) -- Richard Craib built his quant-driven hedge fund Numerai on the idea that if one data scientist is good, 40,000 are better. Now he’s inviting competing funds to draw from the same pool of brains for hire.
San Francisco-based Numerai crowdsources its strategy from thousands of data scientists around the world who compete in a weekly tournament. Participants use Numerai’s encrypted data to develop predictions about financial markets. Numerai announced today it’s opening a marketplace that any hedge fund can use. The platform, called Erasure, operates with Numeraire, a digital token created by the company.
“If someone has really good insight into something -- say a consumer company -- the best thing they can do right now is buy that stock,” said Craib. “Instead we want them to be able to go and sell that information to a hedge fund. It’s a whole new way of drawing information into the market.”
Erasure is like Tinder for markets, bringing together math wizards willing to bet on their predictions and hedge funds ready to pay for hot investment ideas. More participants will produce better data and push up payouts, said Craib, 31, who was a quant trader before starting Numerai. It will also increase demand for Numeraire and should drive up the value of the tokens.
The premise -- that more data scientists equal better predictions and fund performance -- makes sense, said Lawrence Mosley, who holds a Ph.D. in industrial engineering and has participated in Numerai’s tournaments.
“The more data scientists they get, the better,” Mosley said. “There’s so many different techniques, so that if you can aggregate them in some meaningful way, you gain power.”
Numerai’s tournaments bring the gig economy -- where Uber drivers and Airbnb landlords operate -- to the world of finance. Though firms such as Two Sigma Investments, Quantopian and WorldQuant have commissioned coding competitions of their own, the movement toward freelance work has yet to fully take hold in the hedge fund industry.
Hedge funds’ siloed approach breeds inefficiency and limits the mobility of quantitative analysts, Craib said.
“The reputation of the people is trapped inside of the firm and isn’t portable. And that’s true of the data too,” Craib said. “We don’t need more hedge funds. We just need data aggregation.”
Numeraire, the digital token, has dropped by about 90 percent this year, according to CoinMarketCap.com, as digital currencies declined since peaking in 2017.
To submit a prediction, data scientists have to put up a stake in Numeraire. If their model fails, they lose it.
It’s important that “data scientists have some skin in the game,” said Mosley, 34. Otherwise, one person could flood the marketplace with models in the hope at least one would be successful. Once multiple participants join, the transactions will be strictly peer-to-peer with no gatekeeper.
“No one’s going to be in the middle of it, including Numerai,” Craib said. “No data on Erasure will ever reach Numerai’s servers -- ever.”
Rewarding data scientists with digital tokens “sets off some alarm bells,” said Michael Wellman, a professor of computer science and engineering at the University of Michigan. “You can set up a marketplace without setting up your own economy, which when you make up a new currency, you’re sort of doing,” said Wellman, who said he is a skeptic of cryptocurrencies.
Given the limited amount of information hedge funds typically share publicly, there’s little evidence crowdsourcing has worked for actual investing, Wellman said. Numerai declined to provide its returns or its assets under management, only disclosing that it oversees less than $150 million.
The company has raised about $7.5 million from backers including Howard Morgan, a co-founder of Renaissance Technologies, and Coinbase’s co-founder Fred Ehrsam. Paul Tudor Jones, the founder of Tudor Investment Corp., has also invested, Bloomberg has reported.
So the success or failure of Erasure boils down to whether crowdsourcing actually results in the best market predictions. Craib and Numerai provided the platform. Now it’s up to individual hedge funds to filter through it.
“There’s a quote in statistics: ‘All models are wrong, some are useful,’” Mosley said. “That’s the catch.”
©2018 Bloomberg L.P.