[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [amibroker] testing signal correlation



PureBytes Links

Trading Reference Links

Hi Chuck,

Wednesday, January 21, 2004, 9:31:06 AM, you wrote:

CR> I would run a Scan, making a composite showing the count of the
CR> number ofstocks that meet my criteria.

Right away I run up against my lack of understanding.  I tried this
using a couple of statements near the end of the system:

Filter = Buy OR Short;

AddToComposite(Buy OR Short, "~signal", "V");

But when I scan this, I end up with slightly more than 1800 lines.
However, there are only about 880 trades during the period, so that
cannot possibly be right.  Moreover, when I look at ~signal, there is
a zero in the volume holder.

I gotta be doing something stupid I guess.

Yuki

CR> Then, I would write my system such that I could optimise to
CR> find the"right" number of stocks that have to meet my criteria in
CR> orderfor the overall signal to have merit.

CR> For the Optimization, I would read in the composite file
CR> using a"Foreign" statement.   My "Buy" conditionwould require
CR> that the value read from the composite be greater than somenumber
CR> (the optimised number) and whatever else you want toinclude.  

CR> I'm happy to give you exact logic, but I think you can take
CR> it and runfrom here.   Yell out if I can be of more help.  

CR> Also, let me know whenever you want some more Nikei data.

CR> Ciao

CR> At 09:16 AM 1/21/2004 +0900, you wrote:
CR> Hi guys,

CR> I already have a pretty good idea just from looking over back test
CR> results that signal success seems highly correlated with the number
CR> of signals I get on any particular day.  That is to say, out ofa
CR> basket of, say, 20 stocks, if 8 or 10 of them signal the same way,it
CR> seems highly likely that a very high percentage of those signals are
CR> going to be good, whereas if only a single stock triggers on a
CR> particular day, the odds of success, while still good, are notnearly
CR> as good as when many issues signal at the same time.

CR> I'd like to try and quantify that in a back test, but I am not sure
CR> how to write the code.  I need to include some argument in thebuy
CR> and short statements that requires a certain level of"participation"
CR> before any signals are taken.  (This would allow me to runvarious
CR> back tests for that level, in an attempt to discover if this is even
CR> a condition I should consider requiring.)

CR> Can anyone help with the argument part of this?

CR> Best,

CR> Yuki




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 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/