Teaching Machines to Make Your Bets
You teach a kid to walk, talk. You send them to school for a number of years, and then you send them on their way--out into the world, hoping they turn out okay: smart, well mannered and independent.
But what about teaching a machine to think for itself? Yes, that is happening. Machine learning is a relatively new industry, but it’s exploding. Companies like Google and Netflix are using it without consumers even knowing. Machines that have the ability to learn are also being introduced in the financial world, and that’s where Spencer Greenberg, Jeremy Newton, Jonathan Sturges and Alex Fleiss come in. They are the four guys behind Rebellion Research, a New York hedge fund using a machine to make its investment decisions. But it’s not just any machine. Like so many proud parents, the Rebellion Research guys like to boast about their machine (they call it Starr) and think it’s smart enough to make some very intelligent investment plays.
“There are some interesting advantages that computers have compared to humans. For one thing, they are able to find patterns, and I think there are a lot of subtle patterns in the financial market,” said Rebellion Research CEO Spencer Greenberg, a 28-year-old Columbia grad. “It also has the advantage of speed, so a machine algorithm may be able to process 7,000 stocks a day, while no team of analysts could do that.”
Greenberg, along with his three co-founders (two of whom he grew up with), said that using machines to predict the markets is going to become more widespread as time goes on. This mathematician, who is comfortable with the term “nerd,” is confident that the algorithms Rebellion has devised are working.
Greenberg is so enamored with mathematics he believes a mathematical formula can be applied to all aspects of human interaction. In his spare time, he helps run a website called askamathematician.com that answers an assortment of questions, including ones regarding your love life.
“Well, one thing people don’t realize is if on average it takes you a year to find someone you would really like to date and you’re able to double the amount of people you meet each month with taking a sacrifice in the quality of those people….then on average you find someone in six months rather than a year,” he explained. “So simple ways of thinking mathematically will surprisingly have a bearing on our love lives.”
Well, that’s certainly an interesting--and rational--way of looking at the dating game.
As for the fund, which has around $14 million in assets, it’s currently making money. According to a report from the Wall Street Journal, from its launch in 2007 to 2010, it has beaten the S&P 500 stock index by an average of 10%. Apparently the machine is doing just fine out on its own.
Six Shooter with Rebellion Research's Spencer Greenberg and Jeremy Newton:
When times get tough, what inspires you?Greenberg: It can be helpful to remember that making mistakes is an important part of learning. The essential thing is to learn from those mistakes so that they don’t happen again. I also try to remember that sometimes things go poorly even when you did nothing wrong. Many aspects of business and life in general are inherently unpredictable. If you reason very carefully, you will often be able to make good decisions. But inevitably you will still be wrong sometimes, and occasionally things will turn out badly.
Newton: Remembering that in this uncertain world, just because the outcome you obtained was not what you wanted, that doesn't mean that your decision was wrong.
What rule do you live by?Greenberg: I make an effort to keep my beliefs at arm’s length, knowing - both intellectually and from experience - that sometimes even those things I believe strongly in will turn out to be false. When you have less attachment to your beliefs you are able to be less biased and change your mind faster when the evidence proves you wrong.
Newton: Always challenge assumptions and don't be afraid to alter your beliefs in the presence of new information.
What is your favorite quote and why?Greenberg: "Fletchers bend the arrow; carpenters bend a log of wood; wise people fashion themselves." This quote helps remind me that none of us are the best version of ourselves that we can be. With conscious effort we can improve. Learning to spot logical fallacies and becoming aware of human cognitive biases can make us better reasoners. Studying positive psychology can help teach us how to be happier. Learning about cognitive behavioral therapy can give us greater emotional control. Studying entrepreneurship can make us better entrepreneurs. By putting time into self-improvement, we can begin to fashion ourselves into who we would like to be.
Newton: "The important thing is not to stop questioning." – Einstein
Why is your fund different from others?Greenberg: There are a few things that make our firm particularly unique: 1. We use machine learning, which is a type of artificial intelligence. This means that our system learns automatically from market activity, and changes itself over time. 2. Our investment decisions are very long-term compared to most quantitative funds. There has been a trend towards having computers make faster and faster trades, but we take the opposite approach. Despite being driven by a computer, our system does something that is much closer to what professional investors do. While many funds in the quantitative space hold stocks for seconds, minutes or days, we have held stocks for as long as two years before.
What is your biggest tip to other entrepreneurs?Greenberg: When people tell me about their business ideas, I notice that they sometimes haven't thought them through carefully. Before you start a business, I'd suggest that you be able to thoroughly answer the following questions: 1. Who is your competition going to be? What are they doing well? What are they doing poorly? 2. Who are the users of your product going to be? How do those potential users respond when you tell them about your product idea? How do you know that they want your product enough to buy it or start using it? 3. How are you going to make a profit? How many users of your product will you need to become profitable? 4. What are the most likely reasons that your business would fail? What can you do in advance to help prevent those issues from occurring?
Newton: Don't get complacent. Always be executing a plan to improve the business.
Why are machines better equipped to handle your money?Greenberg: When it comes to financial decision making, software-based systems have several major advantages over humans: 1. They can process massive quantities of data very rapidly--far more than the human brain can handle at any one time. 2. They can detect subtle patterns that are very difficult for humans to find. 3. They are free of biases, fear and excitement, all of which can distort the judgment of even experienced investors.
Newton: Many shorter-term price movements in the stock market are fueled by human emotions of fear and greed and lead to a herd mentality--where investors move in and out of an investment idea en masse. Most traders and investors are influenced at least in part by these two emotions (as evidenced by the existence of a herd mentality), but the only way to reliably make money in the stock market is to find investment ideas before they catch the herd's attention. Machines are nice in this task, because they can identify the herd mentality in the market without having the innate bias toward following the herd.
Recommend an entrepreneur you think we should cover by e-mailing us at email@example.com