|
PureBytes Links
Trading Reference Links
|
b, I'm just trying it out the latest beta.
I'll report of my findings.
UM
----- Original Message -----
From: "b" <b519b@xxxxxxxxx>
To: <amibroker@xxxxxxxxxxxxxxx>
Sent: Tuesday, January 20, 2004 2:19 AM
Subject: Re: [amibroker] Getting CSI data into AB (a question for UM)
> UM,
>
> Thank you for the thorough reply.
>
> It has shown me some new ways to approach the storing of
> more data than can normally be fit into AB's 6 data fields.
>
>
> By the way, have you had a look at UA 2.7.7 (just
> released)? CSI support indicates that it has a new call
> that will retrive raw and adjusted data in a single step.
>
> Snip from CSI support email:
> "Unfair Advantage version 2.7.7 will have a function
> RetrieveAdjAndUnadjStock that will obtain the data you are
> requesting."
>
> b
>
> --- uenal.mutlu@xxxxxxxxxxx wrote:
> > From: "b" <b519b@xxxxxxxxx>
> > >
> > > That sounds like some nice programming. Your daily
> > update
> > > speed is short enough so that it is a minor issue.
> >
> > Yes, I'm doing 3 API calls to get all possible cases:
> > 1) Raw (unadjusted original) quotes
> > 2) Split adjusted
> > 3) Dividend adjusted
> > 4) 2+3
> >
> > Ie. all combinations are possible. I'm using a secondary
> > DB which
> > stores the raw data + split dates&ratios + dividend
> > dates&amounts.
> > From that I can quickly generate desired data for AB. I
> > personally
> > use raw (unadjusted) data in AB for my backtests. In ABs
> > OI I store a flag:
> > 0=none, 1=Dividend, 2=Split, 3=Both
> >
> > That is: one knows quickly if that bar is affected by any
> > Splits or Dividends.
> > In case one is interessted in the Split ratio (or
> > Dividend value), or in the
> > adjusted data series, than one can quickly get them via
> > usual AB arrays
> > from the secondary DB using a "foreign" like AFL function
> > call. This secondary
> > DB is accessible via the ABtool plugin, and offers a very
> > flexible and
> > easily extendable fast stocks database. One can
> > practically store any
> > data one likes, even variable length data. It is also
> > possible to store intraday
> > data either in the same or in a new database. Multiple
> > databases can be
> > used and queried at the same time.
> >
> > For about a month now I'm collecting intraday quotes for
> > all US tickers
> > from free internet sources using this DB. I can even plot
> > these intraday
> > quotes in AB eventhough an AB EOD DB is in use :-)
> >
> > BTW, the Sectors and Industries a stock belongs to get
> > automatically
> > assigned in AB since CSI has this data available.
> >
> > > Does your daily update process have a way to detect if
> > a
> > > stock has split? I assume your program puts the CSI
> > data
> > > into the following AB fields:
> > >
> > > O = Open backadjusted for splits etc
> > > H = High backadjusted for splits etc
> > > L = Low backadjusted for splits etc
> > > C = Close backajusted for splits etc
> > > V = Volume
> > > OI = Close NOT adjusted (ie, the "raw" close)
> > >
> > > If your program just appends the CSI data for the new
> > day,
> > > the OHLC fields will be out of ratio with previous days
> > if
> > > there was a split today.
> >
> > As said, I use raw data and than this does not apply.
> > But it is possible to refresh the whole data of a ticker
> > after
> > a split. This can be detected and performed
> > automatically.
> >
> > > Let me explain the problem as I see it.
> > >
> > > Let's say stock ABC is already in the AB database with
> > > perfectly adjusted data current as of yesterday.
> > >
> > > But a split takes effect before the open today. CSI
> > will
> > > process that split on their computers and calculate all
> > the
> > > prior prices for ABC's entire history. Your download
> > from
> > > CSI will communicate those changes to the CSI database
> > on
> > > your hard drive. But how will those new adjusted prices
> > for
> > > the entire history of that stock get into your AB
> > database?
> >
> > As I have learnt, CSI internally stores only the raw data
> > +
> > split dates&ratios + dividend dates+amounts. You can
> > query
> > the DB for the type of data you want: ie. the 4 possible
> > cases
> > mentioned above.
> >
> > > When your program gets today's prices to add to the AB
> > > database, it will get today's OHLC prices which reflect
> > the
> > > split. But the AB data base for ABC stock will not have
> > > been updated for yesterday and all the preceeding
> > history.
> > > Thus the OHLC arrays will show a huge gab from
> > yesterday to
> > > today because all but today's data has not be adjusted
> > for
> > > the most recent split.
> >
> > Since I prefer raw data it was not necessary to do so
> > yet.
> > But true, if one wants adjusted data to be stored in AB
> > than it
> > indeed would be necessary to perform like you describe.
> > But since the number of tickers having a split on a day
> > is
> > limited it would not take much time to completely
> > retransfer
> > all quotes for such split affected tickers.
> >
> > > Does your program have a way to detect which stocks had
> > > splits since the last update?
> >
> > Yes, of course. Even (or especially) with raw data I need
> > to know
> > whether there was a split. During daily update this of
> > course is
> > detected.
> >
> > >And a way to readjust the prior history of stocks with a
> > split today?
> >
> > As said above: it is possible and easily doable, but in
> > my case
> > it was not necessary because raw data doesn't change with
> > splits.
> >
> > BTW, my tool is still in development, and I'm not sure if
> > I ever should
> > release it to public. If there is interest by people who
> > would like to
> > realize some advanced things, fe. data mining, or
> > creating individually
> > structured databases than contact me privately.
> >
> > UM
> >
> > > --- uenal.mutlu@xxxxxxxxxxx wrote:
> > > > Hi b,
> > > > the fundamentals I posted in my prev posting today
> > > > can be retrieved via their API; but as you might know
> > > > the API is not one of the fastest. It took me more
> > > > than 6 hours on a 2600MHz AMD machine to initially
> > > > fill the AB quotes database from the CSI DB.
> > > > Daily updates of the AB DB take up about 15 minutes;
> > > > this of course has to be added to the CSI DB daily
> > update
> > > > time which usually takes about 5 min. BTW, I'm
> > talking
> > > > of timings of my own plugin I wrote; maybe there is a
> > > > much faster way with other plugins or methods.
> > > > UM
> > > >
> > > > ----- Original Message -----
> > > > From: "b" <b519b@xxxxxxxxx>
> > > > To: <amibroker@xxxxxxxxxxxxxxx>
> > > > Sent: Monday, January 19, 2004 6:24 PM
> > > > Subject: [amibroker] Fundamental Data in QP and CSI
> > (was
> > > > Institutional
> > > > Sponsorship)
Send BUG REPORTS to bugs@xxxxxxxxxxxxx
Send SUGGESTIONS to suggest@xxxxxxxxxxxxx
-----------------------------------------
Post AmiQuote-related messages ONLY to: amiquote@xxxxxxxxxxxxxxx
(Web page: http://groups.yahoo.com/group/amiquote/messages/)
--------------------------------------------
Check group FAQ at: http://groups.yahoo.com/group/amibroker/files/groupfaq.html
------------------------ Yahoo! Groups Sponsor ---------------------~-->
Buy Ink Cartridges or Refill Kits for your HP, Epson, Canon or Lexmark
Printer at MyInks.com. Free s/h on orders $50 or more to the US & Canada.
http://www.c1tracking.com/l.asp?cid=5511
http://us.click.yahoo.com/mOAaAA/3exGAA/qnsNAA/GHeqlB/TM
---------------------------------------------------------------------~->
Yahoo! Groups Links
To visit your group on the web, go to:
http://groups.yahoo.com/group/amibroker/
To unsubscribe from this group, send an email to:
amibroker-unsubscribe@xxxxxxxxxxxxxxx
Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
|