Thanks, Steve. That is a useful idea.

But let's say I had thousands of such numbers. Rather than setting
each number manually to an array item, is there a way to perhaps dump
the numbers into a batch file or something? And then somehow pull from
that file?

As a last resort, I could hand code it. But if there is a way to do it
in a more automated fashion, that would be best, as I anticipate
having to do this many times during different optimizations.


--- In amibroker@yahoogroups.com, "Steve Davis" <_sdavis@...> wrote:
>
> If the number of conditions is less than BarCount, you could store
> your non-sequential numbers in an array and then fetch them from the
> array later. Like this:
>
> // Store my numbers
> array[0] = 65;
> array[1] = 126;
> array[2] = 867;
> array[3] = 1023;
>
> // Iterate over my numbers
> for (i = 0; i <= 3; i++)
> number = array[i];
>
> Does that help?
> -Steve
>
>
> --- In amibroker@yahoogroups.com, "ozzyapeman" <zoopfree@> wrote:
> >
> > Looping and optimizing a set of sequential numbers is easy. But
how does
> > one loop or optimize through a large set of non sequential
numbers, that
> > have no defined step value or relation?
> >
> > Without getting too bogged down in tangential details, I have a
DLL that
> > created a new function in AmiBroker. That function spits out thousands
> > of possible buy/sell conditions. Those conditions were then
optimized in
> > a separate AFL, and I ended up with hundreds of "good" conditions
that I
> > now want to step through, for backtesting purposes.
> >
> > For e.g.:
> >
> > for (a = step through all the numbers below)
> >
> > condition65 = ...stuff
> > condition126 = ...stuff
> > condition867 = ...stuff
> > condition 1023 = ...stuff
> >
> > [... a few hundred more]
> >
> > Buy = VarGet("condition"+NumToStr(a,1.0,0));
> >
>