Many years ago, there was a brilliant mathematician named Jim Simons. He taught math at MIT and Harvard and spent some years working for the National Security Agency (NSA) using math to break codes. Then he started using math to figure out price moves in the stock market. He created a hedge fund, Renaissance Technologies, and hired a bunch of mathematicians and data scientists. They didn't know anything about the stock market, but they were really good at pattern recognition and finding anomalies in data. His hedge firm was unbelievably successful, and Simons became a multi-billionaire.
According to Simons in his Ted Talk, hedge funds had a big advantage over individual day traders who tried to trade stocks rapidly. "The real thing," he said, "was to gather a tremendous amount of data." Once you had all the data, you could try to spot trends and outperform the market. "We take in terabytes of data a day."
Image source: Getty Images.
One of the interesting things about Simons and his work is that the original anomalies his data scientists found disappeared. You continually have to find new anomalies. And your success in finding them depends on the quality of your scientists. Jim Simons retired in 2010. And that dramatic outperformance started to dissipate.
Over the last few years, billions of dollars have been removed from the hedge funds and their computer models, specifically from Renaissance Technologies. In fact, over the last year, a simple ETF of the stock market has trounced an ETF of the hedge fund industry.
Can peer-to-peer networks outperform Renaissance Technologies?
Now an interesting phenomenon is happening in the world of crypto. A small start-up, Numerai, is trying to recreate the success of Simons and Renaissance in its early days. Founded by Richard Craib, a mathematician trained at Cornell, Numerai uses peer-to-peer networking to challenge and inspire math nerds around the world. Numerai is providing massive amounts of stock trading data free of charge, just like what the hedge funds use. Data scientists are invited to compete and spot anomalies in this data.
Everything is blinded (kind of like a clinical trial in the world of medicine) so that nobody can steal your model, and you can't steal anybody else's model. And none of the data scientists are aware of the trades that are made. To data scientists, it's simply a competition to spot and predict anomalies in a data set. Once you build your model and are confident in it, you stake your crypto, called Numeraire (NMR 0.61%). If your predictions are right, you are rewarded with more Numeraire. If your predictions are wrong, your staked crypto is destroyed. Craib has a white paper explaining his thesis. (Numerai has a help page for newbies who are confused).
You might think the possibility of losing your crypto would keep people from staking it, but you'd be wrong. It's like investing in the stock market. There's a possibility that you will lose money, so there's a risk. But if you're right, you stand to make a lot of money. In the case of Numeraire, right now 4,675 computer models are controlling the hedge fund. The number of data scientists competing is small, in the thousands. So far, Numerai has paid out $64 million in crypto to the data scientists. The more accurate your model, the higher your rewards.
While Numerai is most exciting for data scientists, ordinary investors can profit, too. You simply buy the crypto and hold on to it. Rewarding the models that actually predict stock market movements -- and destroying the staked crypto that fails to deliver -- makes the remaining crypto more and more valuable over time.
The numbers suggest Numerai is on to something
Despite all the people abandoning the hedge fund industry, it's still a very large industry, worth over $100 billion. Numerai is a tiny operation right now. Its crypto has a market value of $175 million. (Fully diluted, it's $325 million).
Can this anonymous network of data scientists outperform the big money of Wall Street? So far, that's exactly what Numerai is doing. In 2020, many of the quant hedge funds were crushed. Renaissance's RIDGE fund fell 31%, and its RIDA fund dropped 32%. But Numerai did fine, rising 6%, despite all the volatility from COVID-19.
From September 2019 (when Numerai launched its new fund) up until December 2021, Numerai's peer-to-peer fund has returned 21%, versus net losses from traditional quant funds. The model, which rewards winning and originality, seems to be successfully finding new anomalies and thus creating outperformance. Of course, this is just a two-year track record. But if Numerai can keep up this outperformance, the big money might start noticing and join in. If that happens, Numeraire's crypto will become a lot more valuable. A small investment now might reap big returns in the future.