RE: [GT] First MAMA-Implementation

[kw <gt AT netsrak.net>]

On Wed, 2008-03-05 at 08:15 -0600, Weigert, Thomas wrote:
> Yes, I guess that is the answer I would have expected. But as you are
> aware of, depending on the nature of the indicator, no matter how much
> history you use, the value will not be the same one. And depending on
> what other indicators you build up, the difference can quickly multiply.
> For example, a very small difference in the ATR leads to a much bigger
> difference in the ADX.
> 
> Th.

I agree to you on the whole line. We should ensure a proper, relaible
and reproducable initialization. For MAs a SMA with the same data window
as the generated MA would be my choice. For a EMA 9 a SMA 9 as first
datum.

I myself don't like the brute force "put enough data in" progamming
technique.


> 
> > -----Original Message-----
> > From: kw [mailto:gt
AT
netsrak.net]
> > Sent: Wednesday, March 05, 2008 2:00 AM
> > To: devel
AT
geniustrader.org
> > Subject: Re: [GT] First MAMA-Implementation
> > 
> > Mr. Ehlers replied to my questions stating he doesn't care about
> > initialization, he simply uses enough historic data to compensate for.
> > As long as an indicator is not crashing duo to the lack of init values
> > everything is okay.
> > 
>