Recent Developments

  • October 22, 2017: The long effort to refactor the code that runs, under the hood, the momentum program and Retail Backtest's new program is nearing completion. The revised code permits full use of every possible “out-of-sample” option for fairly determining the parameters that specify the programs. Generally, results already published here have been confirmed and the new code is dramatically faster, permitting several portfolios to be analyzed at the same time on the same computer (in separate processes, in an “embarrassingly parallel” fashion).
  • May 25, 2017: There are now three additional theory-tagged blog entries in Retail Backtest's official blog: "Objectivity and the Trailing Performance Period", "Why Most Published Research Findings Are False" and "Real Vs. Hypothetical". Scintillating titles, no? Well read them anyway. Under the Articles menu select Theory-Tagged Blog Entries.
  • April 10, 2017: There are now three new entries in Retail Backtest's official blog. To view them all just pull down the Articles menu to the Theory-Tagged Blog Entries item. The topics cover a lot of ground: relative strength, regime change and hypothesis testing. All of the treatments are quite substantive.
  • February 25, 2017: Retail Backtest has gone social! A blog has been added, under the About menu, and, a free monthly newsletter has been started. Subscribers to the newsletter will also be notified whenever major changes are occurring in program allocations. Visitors can sign up for the newsletter using their email addresses, or their identities with certain “social providers”. Presently only LinkedIn and GitHub are supported, but others will follow.
  • January 27, 2017: “Does Momentum Work? | Chance or Discovery?” under the Articles menu is now Part B of a two-part article, the first part of which is now “Does Anything Work? | Chance or Discovery? Part A”. Readability is improved. Of greater interest is the online availability of three momentum portfolios, under the Performance menu, with an audit feature for each (download the spreadsheet file and notes at the top of it explain the sense in which an audit can be conducted). Finally, RB's New Program has been introduced. Do read about it on the home page and also on the RB's New Program Overview item of the Performance menu.
  • November 8, 2016: There has been a bit of a cosmetic makeover of the site, with improved navigation. Also, under the “Articles” menu there is now an “Abstracts of the Articles” item. Readers can now focus in on their topic of interest without having to browse the site. And alongside the abstracts is a listing of other pertinent writings pertaining to the prospects for portfolio-management success via dynamic asset allocation.
  • October 23, 2016: The “Articles” menu has a hot new item! Please see “Does Momentum Work? | Chance or Discovery?” In the article a complete scheme for portfolio management is disclosed, consisting of an improved momentum approach. However the article is mainly to demonstrate how to determine if a prospective scheme can be relied upon or not. Heard of “hypothesis testing” and “refuting the null hypothesis?” How about painting contractors going broke because they don't know simple logic? No kidding. It's all there. You are invited to write to Mike O'Connor with your questions and comments. Use the “Contact” menu if you like forms, or to use your regular email client just click on write to Mike.
  • ONGOING: Refactoring is about to begin for the under-the-hood code base of RB's New Program— for greater computational efficiency and for better structure (nothing that affects the results). Note especially that the new program is for portfolio management, not for use with a single ETF. However the plan is still to also provide a program that could be used in small accounts with just a few, or even one ETF. See also this microblog.