|
PureBytes Links
Trading Reference Links
|
> - during the live database period if any charts plunge overnight
> double check to see if they have just gone ex or had a split etc.
Hang on!
That's not right.
For 'live' trading use Yahoo raw data (you just change the format
file AQD in AmiQuote folder .. ref the AQ manual if you haven't done
it for a while).
Raw data will plunge on divs or other adjustments.
brian_z
--- In amibroker@xxxxxxxxxxxxxxx, "brian_z111" <brian_z111@xxx> wrote:
>
> > Can you expand on your comment and share alternatives or options
to
> > "maintain quality" with an AQ updated db.
>
> Opinions on data quality and whether Yahoo is value for money are
> personal.... I am not a White Knight for Yahoo ... it just happens
to
> suit my temperament and objectives i.e. to get a good all round
> knowledge of the markets and to do that cheaply.... also I am more
of
> a chart trader than an algorithmic trader ... if I was going to use
> the data for AT I might take a different view.
>
> I have used Yahoo extensively for backtesting and for research but
> note that my backtesting was educational and not specifically to
> write working systems (I did develop some short term swing systems,
> using Yahoo, that I later honed on intraday data from eS).
>
> Example:
>
> When you 'phoned a friend' I was actually running two instances of
> AB .... one on Yahoo ^GSPC daily and one on ES #F (S&P cash future)
> intraday in daily view and they both look the same to me, a pattern
> trader, because the patterns are identical.... as a pattern trader
I
> am not concerned with +_ 0.01% differences in the datapoint values.
>
> To an algorithmic trader the data might 'have errors' because the
> OHLCV values are slightly different between the eS charts and the
> Yahoo charts.
>
> That is not to say Yahoo is incorrect because the futures data will
> always be different and the indexes are calculated in their own way.
> Also, eS intraday is not adjusted and Yahoo data is so differences
> are expected.
>
> IMO Yahoo EOD data comes from a good source and is good but others
> disagree and will not use it i.e. for the USA.
>
> I am uncertain about data for other exchanges and I won't use it in
> Australia where I found the quality wasn't as good as in the US.
>
> However, Yahoo is evolving all of the time and they have changed
> their international provider so my comment could be unfair relative
> to what is available now.
>
> More importantly, it is div adjusted data and that is what I
wanted,
> at the time, and divAdj data is very hard to find so beggars can't
be
> chosers.
>
>
> Back to your question re AQ databases:
>
> I am just giving you a heads up when using it.
>
> Two things:
>
> - last time I checked AQ references the last bar of the first
symbol
> (usually an index because ^ comes before A in sorting).
> AQ then uses this date as starting date for downloading all symbol.
> If any symbols have no data for that day, at the time of your
> download, but this is later corrected,
> then AQ will not backfill data holes.
>
> - Yahoo adjust data for various reasons and when they do they back
> adjust. AQ only downloads the current bars, as marked by the first
> symbol, and will not backfill adjusted data.
>
> AQ does have some manual allowance for adj factor but I have an old
> dog (DAve and Steve please note!) and a parrot so I don't need any
> more headaches.
>
>
> Some traders use MLDownloader which reports divs, finds new symbols
> and backfills missing bars (not sure if it refills adjusted data).
>
> I didn't bother buying MLD because I tend to be more trusting of
> Tomasz and also I am not that serious about EOD trading as such ( I
> am in intraday to 2-3 day swings).
>
>
> My recommendation:
>
> - download a deep history database for backtesting
> - save it/back it up and don't add any more bars to it.
> - do not use auto download to create the database (set the dates in
> AQ the way I have shown at the UKB and push the data to AB).
>
> for every day use:
>
> - set up a 'live' database with enough history to cover your
> indicators
> - auto download every day
> - dump your live database every week or every month and start again
> - during the live database period if any charts plunge overnight
> double check to see if they have just gone ex or had a split etc.
>
>
> Thats' the best I can do from memory .. I don't use it that much
and
> only tried to learn it well in the beginning.
>
> Others might be able to add something.
>
> brian_z
>
>
>
>
> --- In amibroker@xxxxxxxxxxxxxxx, Ken Close <ken45140@> wrote:
> >
> > Brian:
> >
> > Thanks for the detailed tips...good stuff.
> >
> > You said:
> > >>I meant to warn you that it is not good practise to append new
> data to an
> > existing, AQ created, database for backtesting (susceptible to
> > corruption).... it's OK for temporary working databases (scanning
> for
> > current signals etc) ... for backtesting take a history and
> quarantine the
> > database.
> >
> > This has me worried, as I am working myself not only towards
> backtesting on
> > which to depend, but for an AQ maintained db to be my working db
for
> > O/H/L/C/V data.
> >
> > Can you expand on your comment and share alternatives or options
to
> > "maintain quality" with an AQ updated db.
> >
> > Ken
> >
> > -----Original Message-----
> > From: amibroker@xxxxxxxxxxxxxxx
[mailto:amibroker@xxxxxxxxxxxxxxx]
> On Behalf
> > Of brian_z111
> > Sent: Friday, November 14, 2008 8:39 PM
> > To: amibroker@xxxxxxxxxxxxxxx
> > Subject: [amibroker] Re: Another AQ Question
> >
> > a) For an existing database that includes symbosl with no data:
> >
> > - Tools >> Database Purify
> > - Select reference >> All Symbols >> select All Symbols/N ==
> > 1, Report Missing Quotes
> > - Deselect other report options
> > - Default threshold is OK
> > - Analyse
> > - Click on column header to sort (separate Missing from No
> > Quotes)
> > - Shift/click first and last NoQuotes symbols
> > - Right Click and send to watchlist
> >
> > Normal procedure to delete WL symbols from DB
> >
> >
> > b) For a new EOD database:
> >
> > - create a new blank database
> > - leave AB open
> > - open AQ and import ticker list into AQ
> > - uncheck AutoImport
> > - set AQ dates to download one bar (quicker)
> > - download and symbols will record errors
> > - AQ >> Edit >> Toggle (switches checked symbols to non error
> > symbol)
> > - Check AutoImport
> > - Set download dates to required history
> > - Download (only symbols with data will be `pushed' to AB)
> >
> > I didn't have time to check method B completely... I am pretty
> certain it
> > works ... if not tweaking it a little will get results (save
> filtered list
> > and restart database).
> >
> > AFAIK you can't sort the AQ columns or delete all error symbols
> from the AQ
> > list.
> >
> > I couldn't find the toggle to turn on the 'hidden' AB help manual
> either.
> >
> > I meant to warn you that it is not good practise to append new
data
> to an
> > existing, AQ created, database for backtesting (susceptible to
> > corruption).... it's OK for temporary working databases (scanning
> for
> > current signals etc) ... for backtesting take a history and
> quarantine the
> > database.
> >
> > There are some tricks to using AQ ... a few are documented at the
> UKB
> > although the DBM section is incomplete and becoming dated now :-(
> >
> > brian_z
> >
> >
> >
> >
> >
> >
> > --- In amibroker@xxxxxxxxxxxxxxx, "Ken Close" <ken45140@> wrote:
> > >
> > > As I load symbols into AQ and fetch data, a bunch of symbols are
> > reported as
> > > errors, obviously symbols which have changed names or are no
> longer
> > listed.
> > >
> > > My question is whether there is a convenient way to purge the
> > database of
> > > the out of date symbols, by somehow reading the window of
results
> > for those
> > > having errors, and then turning that into some mass deletion
list.
> > >
> > > Does such a method exist?
> > >
> > > Thanks.
> > > Ken
> > >
> >
> >
> >
> > ------------------------------------
> >
> > **** IMPORTANT ****
> > This group is for the discussion between users only.
> > This is *NOT* technical support channel.
> >
> > *********************
> > TO GET TECHNICAL SUPPORT from AmiBroker please send an e-mail
> directly to
> > SUPPORT {at} amibroker.com
> > *********************
> >
> > For NEW RELEASE ANNOUNCEMENTS and other news always check DEVLOG:
> > http://www.amibroker.com/devlog/
> >
> > For other support material please check also:
> > http://www.amibroker.com/support.html
> >
> > *********************************
> > Yahoo! Groups Links
> >
>
------------------------------------
**** IMPORTANT ****
This group is for the discussion between users only.
This is *NOT* technical support channel.
*********************
TO GET TECHNICAL SUPPORT from AmiBroker please send an e-mail directly to
SUPPORT {at} amibroker.com
*********************
For NEW RELEASE ANNOUNCEMENTS and other news always check DEVLOG:
http://www.amibroker.com/devlog/
For other support material please check also:
http://www.amibroker.com/support.html
*********************************
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/amibroker/
<*> Your email settings:
Individual Email | Traditional
<*> To change settings online go to:
http://groups.yahoo.com/group/amibroker/join
(Yahoo! ID required)
<*> To change settings via email:
mailto:amibroker-digest@xxxxxxxxxxxxxxx
mailto:amibroker-fullfeatured@xxxxxxxxxxxxxxx
<*> 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/
|