Vince,
Have you seen the latest version of our article?  In a nutshell, it
establishes both what's right and what's wrong with r-squared/regression.
Essentially, r-squared is fine PROVIDED that the amount of the derivative
optimal, but otherwise it makes little sense.
Please let me know what you think about this.
Regards,
Andy
Andrew Kalotay Associates, Inc.
(212) 482-0900
andy@kalotay.com

Visit our web-site http://www.kalotay.com

 - FAS133 article.pdf