Re: How to do ... appendum

["Glen Wallace" <gcwallace@xxxxxxxx>]

You could probably do it with bunch of Fml( ) functions, but a better
solution likely is a .dll file for Version 7.0's MSX


----- Original Message -----
From: jr <jrbox@xxxxxxx>
To: <metastock@xxxxxxxxxxxxx>
Sent: November 9, 1999 11:48
Subject: Re: How to do ... appendum

> Glen,
> of course, in the final example of my initial mail
> I did like this when I used H0 (is B there) again
> and again. However, that still will not allow for
> more than 20 variables and I want to take, say,
> the first 50 autoregression coeffizients. There
> has to be some other solution or MS is useless
> for a purpose like that (Also, the 'code' grows
> fast, so the space limit is reached very soon.)
>
> Regards - Jasper
>
>
> ----- Original Message -----
> From: Glen Wallace <gcwallace@xxxxxxxx>
>
> > Jasper:
> >
> > I'm not really clear on what your problem is, but if you are running
> > into a maximum number of allowable variables, simply eliminate the
> > H1, H2, H3 ... definitions. For example, rather than:
> >
> > > H1:=(C-A)*(Ref(C,-1)-A)
> > > +(Ref(C,-1)-A)*(Ref(C,-2)-A)
> > > +(Ref(C,-2)-A)*(Ref(C,-3)-A)
> > > ;
> > > r1:=H1/H0;
> >
> > the following will reduce the number of variables used by nearly half:
> >
> > > r1:= ((C-A)*(Ref(C,-1)-A)
> > > +(Ref(C,-1)-A)*(Ref(C,-2)-A)
> > > +(Ref(C,-2)-A)*(Ref(C,-3)-A)) / H0;