[podcast] Creating the ultimate algorithm: How Numerai gets experts to build successful trading strategies

automating Wall Street, fintech, and layoffs

We’ve spent the past couple of weeks of this show talking to leading early-stage investors in financial technologies — guys like Charles Moldow of Foundation Capital and Caribou Honig of QED Investors.

These are the same guys that have invested in some of the biggest names in the financial technology space. Companies like Lending Club, Credit Karma, Motif Investing, OnDeck Capital, Prosper, and Zopa. What makes many of these companies different from their incumbent competitors is that they’re bringing technology solutions to financial problems.

Richard Craib, Numerai
Richard Craib, Numerai

That’s not to say incumbent financial institutions don’t use technology — they certainly do, a whole lot of it. It’s just that this generation of fintech company is exactly that fin – tech: they’re technology companies at their core and their solutions to financial problems — whether lending to people without credit history or remitting money more cheaply and quicker across global banking borders. They’re technological in nature.

Some of the most interesting things we’ll see in the future is machine learning applied to the stock market. We know human beings’ biases when it comes to investing and even the world’s best investors fall prey to our own shortcomings, our own humanity. It’s entirely possible machines can do it better. Our next guest, Richard Craib, is working on a machine learning solution to the stock market much in the same way Netflix turned to top external algorithm developers to improve its own recommendation engine — by offering prizes.

Richard’s company, Numerai is an amazingly ambitious undertaking. He’s offering top investing systems developers cash prizes in return for using his firm’s encrypted data set to build alpha-producing strategies. Sitting on top of it all is Numerai with its own big data solution that’s investing its own nostro account, trying to optimize allocating capital between strategies. It’s a black box of black boxes but crowdsourced from some of the world’s smartest data scientists.

As I talked with Richard, I kept thinking to myself, what does this spell for the future of investing? People like Warren Buffett and Peter Lynch — two of my favorite all-time investors — seem so antiquated when you start thinking about AI’s potential to participate in investing. Will future episodes on CNBC profile — not today’s celebrity cowboy stock picker — but tomorrow’s PhD-touting data scientist who builds software for a living? Interesting.

Listen to the FULL episode

In this episode, you’ll hear about

  • how data science contests, like the Netflix Prize, are successful in getting scientists to work on innovative solutions to problems
  • how the stock market is also an interesting problem to solve for scientists
  • the encrypted data set Richard’s firm, Numerai has built as a toolset for trading system developers
  • Numerai’s challenge of optimizing its own allocation of capital across the models on the platform
  • the type of people, quants, students, and professors working on Numerai tournaments and what motivates them
  • Numerai’s business model and how Richard’s background in pure mathematics
  • How he used Reddit to get the word out to the machine learning community

More resources

EVEN MORE RESOURCES

 

Photo credit: spenceyc via Visualhunt.com / CC BY-ND

What is Tradestreaming: Screening 2.0

Top investment gurus like Benjamin Graham, Warren Buffett, Peter Lynch, and Joel Greenblatt didn’t only help investors enjoy huge market-beating returns in their funds.  They also left behind the keys to the (investing) castle: the methodologies they applied in their market-trouncing performance.  They’ve written books, complete with formulas and strategies, that propelled them to the top of their games and gains.  Tradestreaming aims to recreate these strategies as we pave our own way to outperformance.

Because a small number of expert investors wrote extensively about their investing techniques, we can now create complicated computer programs to reenact their strategies and apply them to today’s stock markets.  Screening 2.0 is all about using smart technology to bring history’s best investors back to life.

Technology-driven investing

Stock screens have been around for decades.  Using screens, we can filter through thousands of investment candidates on the prowl for the ideal investment.  Old screens merely searched databases of stocks using specific criteria (i.e. all large cap stocks with a p/e less than 20 and a growth rate over 7%). Unfortunately, for most investors, these screens fail — searching for specific stocks tells us nothing about the success of such a strategy.

Screening 2.0, lead by analysis and money management firm, Validea, allows us to recreate history’s best investment strategies, computerize them, and then look for stocks that guru investors like Ken Fisher and Marty Zweig would have purchased themselves.  Screening 2.0 is the marriage of search technologies and artificial intelligence with quantitative investing.

More Resources

Make sure you check out the Tradestreaming for the Internet’s best stock screening resources.

<– Previous: Crowdsource your portfolio I