|
PureBytes Links
Trading Reference Links
|
Dear Tomasz,
Thank you for your reply.
>After looking closer at optimization results of your
system>on different stocks I have noticed that best combinations of
>X1, Dbuy, Dsell optimization variables vary significantly.
This simply means that [30,70] or [35,65] system is
totally
out of comparison and this was my first
target.
If there was any chance, then we would see the
optimization
results to converge near these levels.
This means that we search for a new way to use RSI, if
any
exists.
>The problem is that once we optimize for single stock we
get "perfect">parameters but only for this one stock and the historywe
just used to>optimize.
As i wrote from the first posting,
<A
href="">http://groups.yahoo.com/group/amibroker/message/5346
this is for a current stock study.
This time, I take the "local" path, not the "global" one. Is
it bad to study
two or three stocks and find a best-fitting system, a
"taylor-made"
parametrization for each one, since we like to trade these
stocks?
Certainly not. Especially if, in the past, we were loosing
money, applying
30/70 systems and, the worst of all, we did not know why we
loose.
May be, through the local path, we find a more general
solution.
May be we write in a better way this page of T/A
booklet.
If we succeed, this will be great. If not, the sure thing is
that the 30/70
general solution it is not a general solution, it is not even
a solution at all and
should be disregarded.
Now, back to the job, there is a
problem indeed, but it is quite different.
I found out that the "optimum" combination for a singlestock
is NOT so
optimum as we may consider.
I had the curiosity to run optimization for a single stock for
parts of the
time period.
|