MORE STORIES ON QUANTS

AM | @agumack

"... backward looking data can fizzle out" — Robin Wigglesworth

There are three very interesting stories on quants in today's FT (make it four if you count John Authers' piece on smart beta). Robin Wigglesworth —who writes more persuasively on quants than on markets and interest rates— mentions a Two Sigma competition in which contestants are given three months to code a trading algorithm based on four gigabytes of financial data. The winner will pocket $100,000 (*). His 'Big Read' piece is also pretty interesting, as he mentions the 2007 débâcle of Goldman Sachs' Quantitative Investment Strategies. The sector has since then recovered, and assets under management for quantitatively-oriented hedge funds are approaching the $1 trillion mark. (Don't miss the short section on GS' alternative risk premia unit).

                                                                      * * *

Can we possibly use some of this material in class? Yesterday in the Financial Markets exam I had students estimate two entry points using 20-day and 50-day moving averages. Shortly I'll be testing some (extremely simplified) exercises with 'Natural language processing'. We could take speeches from the Fed chair, or a sample of research reports. Here's Dennis Walsh of Goldman Sachs:

Goldman's algorithms can systematically look for verbal clues from analysts on a call that might indicate whether they were pleasantly or unpleasantly surprised at the results—and therefore upgrade their outlook in response. "There's a tendency towards praise to keep in management's good books, but only marginally. If 20 out of 30 analysts say 'great quarter' then it probably was".



(*) Robin Wigglesworth: "Lessons from the quant quake" and "Funds adopt novel methods to hunt down new talent"; John Gapper: "Technology outsmarts the human investor". See also John Authers: "Popular 'smart beta' strategies reveal their true value in trial". All from the March 9 issue of the Financial Times.
_______________