Problems With the Current Method

You Think You have Problems?
The
lists change regularly; when do you build a new system?
If you are running 2 systems, what about overlaps?
Systems
are developed based on the past, whereas those stocks may not have been on the
list in the past. Hence backtesting only gives you a way of trading those
stocks, not any indication of what would have happened if you had traded the
lists in the past.
End
of day price vs your price.
How do you get started? Buy the current holdings, or start up the system
fresh?
Problem with any mechanical system: If a stock drops out of the list,
or is deleted for some reason, a restructuring of the past can
occur. Errors in the data, such as Fasttrack missing a split (recently
on MATR) may also cause some problems.
I
do not know of any satisfactory purely mechanical solution to these problems.
You have to use some common sense.
Stock Trading is more time consuming than trading funds, unless you just put
in market orders.
These stock systems are not necessarily well correlated to the major market
averages. Hence the timing systems we have used do not necessarily work
well with these systems. It is better to time the systems based on their
own equity curves, and to hedge if it looks profitable.
Question: Is trading frequently more risky than trading seldom? Anybody
read the studies that say frequent traders lose money?