|
PureBytes Links
Trading Reference Links
|
The problem lies in the first straight line, that line should be zero, but
it is not. The other line follows the PlusDM variable. I realise now I
should not have included that part in the mail, it only confuses. I can make
a code that signals for the occasions when the error causes problem,
probably that is what I should have included the first time.
I have read the threads on precision problem in Tradestation. The decimal
problem is only a printout problem, the decimals are there, but due to some
strange bug, they can not be printed further than to the third decimal.
Making the internal numbers bigger is not a solution since the data involved
in calculation is high/low/close. You might be right that the errors occur
only if there are decimals involved (it is likely, there is some strange
calculation engine in Tradestation). However, decimals will always occur, at
least if the source is stock symbols with decimal price systems, so the bug
must be corrected by Omega if the indicators shall be freed from the effects
of it.
>From Gary I got this much more readable version producing the error, it
focuses on the error instead of the dmi calculation which makes it easier to
understand where the problem is:
******************************************
{Test Indicator, must plot Zero line}
Value1 = 10.387 - 10.280;
Value2 = 10.280 - 10.173;
If Value1 = Value2 Then Value1 = 0;
plot1 (Value1);
******************************************
> -----Original Message-----
> From: Jack Griffin [mailto:jack_2231@xxxxxxxxx]
> Sent: den 17 juli 2001 00:10
> To: Bengtsson, Mats; Eskimo Omega List
> Subject: RE: I think I have to change my opinion from bad
> precision to bug gy c alculations
>
>
> --- "Bengtsson, Mats" <mats.bengtsson@xxxxxxxx> wrote:
> > I did not understand plot2 being lower than plot1 in
> > TS PRO. Plot1 is fixed,
> > so it is as I expected the same bug in TS PRO if
> > plot1 plots the same on
> > both of them.
> >
> > But plot2, how can it differ? The part included does
> > not contain any
> > accumulator, that comes later in the original code.
> > Could it be that you are
> > looking at different dates or got different scales?
>
> Well the data in Pro is from a different source. And
> I didn't check each bar. I tried to be sure the
> dates, # bars, scale were 100% the same. Plot1 was
> the same as I recall while plot2 seemed lower in Pro
> by that factor. I didn't try to find out why, just
> posting what I found out. You have yet to pinpoint
> exactly which line/s actions cause the error with
> further experiments. Did you follow the recent thread
> on the code list in which many of these issues were discussed?
>
> I would guess (without looking at your code in detail)
> that anything done in TS beyond decimal #2 is wrong or random. And
> that the easiest way to avoid the problem is to make the internal
> numbers bigger, which is not hard to do.
>
> Jack
>
>
>
This message contains information that may be privileged or confidential and is the property of the Cap Gemini Ernst & Young Group. It is intended only for the person to whom it is addressed. If you are not the intended recipient, you are not authorized to read, print, retain, copy, disseminate, distribute, or use this message or any part thereof. If you receive this message in error, please notify the sender immediately and delete all copies of this message.
|