Originally Posted by haha
Many people say this, but I believe it comes from a-perfectionist definition of the problem. If on the other hand someone defines the task as only investing at a PE10 level that has in the past data been associated with attractive long term returns, there is no problem. Yet that rule would have likely kept the investor out of many recent markets, so if he is bench-marking he is not going to like it..
I'm concerned that the model is "over fit", but the PE10 level / momentum scheme in "Rock Breaks Scissors" showed a nice benefit over buy and hold (Using Shiller PE to Time the Market
). Average real return of the S&P since 1881, according to this analysis, was 6.23%. By using the scheme in the book, the real return was 7.93%...nothing to sneeze at. It requires that the investor be out of the market for very, very long swaths of time, though. So "nobody" is going to do it...at least if they need to prove their worth every quarter.