Bests & Worsts Reviews from Amazon

I started to read this book with high expectations. I believe that a strong dose of commonsense sometimes can do more than a lot ot analytical expertise (I have a Ph.D. in Economics from MIT, and through the years I have come to recognize the limitations of hard analytical methods). The first few chapters of the book were OK. However, I was expecting the book to end with a bang, but it did end with a whimper. The author does not provide a model that can be used to forecast anything. He basically tells you to build your own. There are two possibilities: 1) You have enough technical expertise to build an econometric model of the US economy that would potentially be useful to forecast the stock market. 2) You have no analytical skills. In case 1) you will find out that the book is basically useless, as it contains nothing new or original. The author makes a lot of noise about tracking rates of change instead of levels, but anyone who does econometrics knows he just transformed his variables so that they are stationary, a requirement for doing meaningful regressions. In case 2) after reading the book there is nothing you can do to forecast the stock market except possibly to read the comments in Mr. Ellis web site. The leading relationships that Mr. Ellis claims exist between variables like personal consumption expenditures and the stock market are not easy to find (assuming they exist) and the graphical tools used by Mr. Ellis are totally insufficient for the task. Let me end quoting from the book (Appendix D, page 263): "Those choosing to construct their own charts ... may notice some discrepancies between charts constructed from data on these Web sites and the charts in this book. This results from the "rebasing" of the statistics by the government and other bureaus that use them. As far back as the early 1980s, the elapsed time between peaks and throughs in the leading and lagging indicators was extremely clear in the years the data was reported and for five to ten years afterward. However, in following years these lead/lag times often shrunk or disappeared altogether as the data was restated and rebased by the issuing bureaus."