"Will Trice" <wwtrice[at]paragondynamics.com> wrote
- quote -

I meant over all time.

Timing using, say, certain moving averages certainly "works" during some
parts of the stock market's history. Anticipating when it will cease to work
is the catch.

Elle wrote:
- quote -

I am aware of the dangers of data mining. As I said, the stock screens
in the AAII program were derived from books written by successful
investors, and most of the screens have outperformed the market even
after they were published. The authors of the software say they have
not withdrawn any strategies or modified their rules.

- quote -

I am monitoring the strategies and intend to use them, probably with
some modifications. I want to work on a program to optimize weights,
given a list of stocks produced by several screens.

Elle wrote:

- quote -

Many strategies will work when backtested. Think of simple curve
fitting. Strategies that work on out-of-sample data is an entirely
different matter...

-Will

<beliavsky[at]aol.com> wrote
- quote -

Has it occurred to you that

(1) when back tested, no strategy has consistently yielded successful stock
picks?

and

(2) strategies that back testing shows to be unsuccessful will not be
advertised as such?

- quote -

So are you using some of the strategies? Or are you just experimenting, for
fun? Or are you using this experience to support a hypothesis that, say,
buy-and-holding using, say, indices, as is likely to succeed as any of these
strategies?

- quote -

Anything with fundamentals and forward looking dollar statements.

There are various web sites and computer programs that will let one
screen for stocks meeting specified fundamental (market cap > X, P/E <
Y, etc.) and technical criteria and present the CURRENT results. What
about tracking the HISTORICAL results of a screen?

I'd like to know, for example, for a particular strategy, what stocks
would have been chosen at the beginning of each quarter for the last 10
years. The Compustat database is used by academics and institutional
investors, and I assume it has a price tag to match. I wonder if there
is something costing up to a few hundred dollars geared to individual
investors.

I like Stock Investor from AAII -- it has pre-defined screens for many
strategies, gleaned from books written by successful investors, and it
allows for user-defined screens. I wonder what screening tools people
like (even if they don't produce historical results).

The historical results of style-based indices, such as those by
S&P/Barra or Russell, can be used to track a few simple rules, but I
want to test more complicatd strategies, and I prefer to see historical
returns for equal-weighted as well as cap-weighted portfolios.