Quantitative traders are hoping that computer-driven strategies will be able to “figure out this crazy stock market,” according to a recent article in Bloomberg.
“Machine-learning takes quant investing to the next level because the robots are programmed to adapt and improve their performance based on the data they sample over time, without needing explicit human instructions,” the article explains. Although the concept is not new, it adds, the difference now is that the tools have been getting increasingly “cheaper and easier to use” and the backdrop has changed: “Whipsawing markets and the over-crowding of many quant strategies have battered their performance and started to undermine investor enthusiasm for this once red-hot corner of the investing world.”
Advocates of machine-learning argue that it can offer a competitive edge that quants are now missing because it can allow them to glean new signals from massive amounts of data and therefore improve returns. Aberdeen Standard Investments’ Boyan Filev echoes the sentiment: “In terms of the scope of data we can analyze, it’s much greater than before. We’re analyzing billions of data points every month.”
But some believe that machine learning needs more fine-tuning and argue that the new techniques come with the risk that computers will identify patterns that “don’t actually work in the real world.” Martin Kallstrom, a partner at Lynx Asset Management, asserts, “It’s very important that the algorithms use information-rich data and not all possible data out there because in financial markets, there’s so much noise.”