|
PureBytes Links
Trading Reference Links
|
Tomasz, i just updated ami, Same problem.
I had used a file called Sample ASX.exe that i downloaded from
the amibroker group file section.( cant see it in there anymore ).
When scanning for a composite does AMI scan the stocks in the
watchlist if i had selected it in the use filter of the AA.
It still seems like it is testing hidden files that where
originally in the sample ASX data set.
peter.
--- In amibroker@xxxx, "Tomasz Janeczko" <amibroker@xxxx> wrote:
> Peter,
>
> Which version are you using?
> I suggest downloading 4.20.5
(http://www.amibroker.com/bin/ab4205.exe) and trying again.
>
> Best regards,
> Tomasz Janeczko
> amibroker.com
> ----- Original Message -----
> From: peter j
> To: amibroker@xxxx
> Sent: Friday, November 01, 2002 1:41 AM
> Subject: [amibroker] IS THERE A BUG IN THE DATABASE.
>
>
> The database in AMI is incorrect when using the sector and
industry assignments.
> (this is not an actual not an AMI problem).
> when using e.g. the sample ASX that sets up these assignments
and then loading the data history there is a major problem.
> I came across this recently when using the Add to composite
formula (below) for counting stocks traded each day.
>
> AddToComposite(1,"~MyCounter","X");
> Plot(Foreign("~MyCounter","C"),"MyCounter",4,2);
> Buy=Sell=Short=Cover=0;
> Filter=1;
>
>
> There was a huge jump in the number of shares trading on the
day for a short period of time which matches the start date and last
date of the data which is supplied in the SAMPLE ASX data base.
>
> To further test this i ran a test to find the last day traded
and only selected the stocks that traded up to todays date and added
this to a watchlist. but there was still a large jump.
>
> I then created a new data base without using the sample ASX
sector and industry assignments.
> This worked, and there was no large jump in the number of stocks
trading on the day. (both the databases had the same number of stocks
which is why i cant figure out why there are different number of
stocks traded on the day from 1 data base to the other )
>
> When running backtests there were different results as well.
> For the past 9 months or more i have had false backtest results
without even knowing it.
>
> Does anyone know why this happens as it seems to do it when
there was a small sample of data
> and then loading the whole data history into it.
>
> Its almost as if there were hidden data files as there were the
same number of stocks in the watchlists, that i ran the test on.
>
>
> I have attached 2 files which shows the results .
> (CountA) is the result of the normal database.
> (countB) is the result using the sampleASX database.
>
> the problem occurs on the 3/12/2001 till 1/3/2002 in the
countB image.
>
>
> If this is causing a problem what happens when updating data
after the data supplier has adjusted the data for splits E.T.C.
>
> thanks to Herman van den Bergen for the add to composite
help file as i would probably not have noticed this for a while to
come.
>
>
> Peter.
>
>
>
>
>
>
>
>
> --------------------------------------------------------------------
----------
> Do you Yahoo!?
> New DSL Internet Access from SBC & Yahoo!
> Yahoo! Groups Sponsor
> ADVERTISEMENT
>
>
>
>
> Post AmiQuote-related messages ONLY to: amiquote@xxxx
> (Web page: http://groups.yahoo.com/group/amiquote/messages/)
>
> Check group FAQ at:
http://groups.yahoo.com/group/amibroker/files/groupfaq.html
>
> Your use of Yahoo! Groups is subject to the Yahoo! Terms of
Service.
>
>
>
> --------------------------------------------------------------------
----------
>
>
>
>
>
>
> --------------------------------------------------------------------
----------
|