Crowd Sourcing for Algorithm Trading Experts
Fintech. This seems like such a buzz word in the technology sector, especially in this part of the world. The logic behind this seems so compelling. We used to do financing though brick and mortar banks. Now that we have high bandwidth secured internet, why don't we do the transactions faster, cheaper, better, and more frequently through your computer or your mobile phone?
One area that people have been asking is algo-trading --- trading based on algorithms or you can say "let the computer do it for me". In fact, wouldn't be nice if we have an algorithm that can take in all the data of the market and "invest" for you, wouldn't it be nice? It sounds innocuous enough. After all, don't we now let car drives by themselves? And, we have long embraced smart home and smart city.
Today, we want to take a look at a company that is so somewhat offbeat but very interesting ---Quantopian (https://www.quantopian.com/). The company has raised about US$50M. Based in Boston, this company encourages mostly financial engineers to write algorithm for trading. Instead of coming out with the algorithm itself, Quantopian is a platform on which you can test and deploy your own algorithm. The "big problem" for developing an algorithm in trading is the data source, in terms of data integrity (accuracy), kind of data, and speed of getting those data. And, to test an algorithm, a financial engineer must have access to this timely information. (For those of you who like this kind of story, you may read the book "Flash Boys" by Michael Lewis. http://www.cbsnews.com/news/michael-lewis-explains-his-book-flash-boys/).
The business model is very interesting indeed. Contributor to the trading algorithm, if they get accepted, will get a 10% cut for trading based on their algorithms.
For readers who have read our last blog in which we talked about the Behavioral Economics by Richard Thaler (https://www.jacklau.info/single-post/2017/03/15/A-Bunch-of-Interesting-Stats-and-a-Brain-Teaser-from-Behavioral-Economics), you may remember the conclusion that if everyone is relying on the same data and being equally smart, the net stable point is "net zero" position --- no one makes any profit. (That is obvious, right? If everyone thinks a deal is good, who would sell?)
However, we still think that this model of Quantopian is worth exploring if not duplicating in this part of the world.
The data set provided by Quantopian is strictly United States-centric. If you are trading in mainland China, Singapore, Hong Kong, Japan, you are out of luck. What if someone deploys this in Asia, encouraging financial engineering to flourish in this part of the world more? The trick is timely and comprehensive economic and financial data. Interesting?