A new hedge fund called Numerai is redefining investment strategies by paying data scientists in Bitcoin to build algorithmic models. The fund, which describes itself as a “super intelligent stock market,” relies on machine learning and collective intelligence to enhance stock market predictions.
Bitcoin Incentives Drive Global Data Science Collaboration
Numerai gathers insights from over 7,000 data scientists worldwide, leveraging their machine learning models to improve prediction accuracy. According to founder Richard Craib, the company has already distributed $150,000 worth of Bitcoin to 100 scientists for their contributions. These models have collectively processed more than 29 billion equity predictions. The platform uses a leaderboard to aggregate the best models, but the goal is not competition—it's an invisible collaboration to build a meta-model. “No data scientist on Numerai has a machine learning model that is better than all the other models combined,” explains the company’s blog. “Numerai is not a search for the ‘best’ model; it is a platform to synthesize many different models with many different characteristics.”
Cryptocurrency and Prediction Markets: A Symbiotic Future
Numerai is not alone in using cryptocurrency to fuel machine learning and prediction markets. Projects like Bitcoin Hivemind (Truthcoin) and Augur are also leveraging blockchain technology to create decentralized prediction platforms. Hivemind, developed by Paul Sztorc, is a peer-to-peer oracle system tethered to the Bitcoin network, allowing users to speculate on global events. Augur, built on Ethereum, uses game theory and cryptocurrency incentives to crowdsource predictions. Both projects aim to outperform traditional markets by harnessing collective intelligence. “Prediction markets are ‘rule of the experts,’” Sztorc previously stated. Augur, currently in beta, is expected to launch live in 2017, with its native token REP already trading.
Voluntary Data Collection and the Future of Machine Learning
Numerai's approach highlights a shift toward voluntary data collection as a more ethical and effective alternative to scraping. By combining machine learning with portfolio management, the fund aims to achieve lower error rates and higher returns. “We are building the largest ensemble of stock market machine learning models in the world,” Craib said. As cryptocurrency and prediction markets evolve, such collaborative models may pave the way for global learning systems that transcend traditional financial boundaries.