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

[amibroker] Re: Optimizing & Robustness of single parameters



PureBytes Links

Trading Reference Links


The next for systems w/more then 2 parameters where exhaustive search 
is not feasible is to incorporate sensitivity testing of parameter 
values in the optimization process so that one winds up with a 
relatively insensitive group of parameter values.  This IS doable 
using an intelligent optimization process.

--- In amibroker@xxxxxxxxxxxxxxx, "Steve Dugas" <sjdugas@xxxx> wrote:
> Thanks Fred - very nice idea!
> 
> FWIW, I just realized that Herman sent the code for his graphs in a 
> follow-up message to the original. Here it is for anyone that might 
be 
> interested.
> 
> Steve
> 
> ----- Original Message ----- 
> From: "Fred" <ftonetti@xxxx>
> To: <amibroker@xxxxxxxxxxxxxxx>
> Sent: Monday, April 04, 2005 9:22 PM
> Subject: [amibroker] Re: Optimizing & Robustness of single 
parameters
> 
> 
> >
> >
> > For systems with a single parameter this is of course fine ...
> >
> > For systems with more then 2 two parameters where the 3d plots 
are no
> > longer of great use it's a simple matter to generate random data
> > points in the +/- n% range of what appears to be the optimum 
value to
> > see how sensitive or not the parameter values actually are i.e.
> >
> > Original System where the default values are assumed to be optimum
> > but sensitivity is unknown ...
> >
> > X = Optimize("X", 10, 1, 20, 1);
> > Y = Optimize("Y", 100, 1, 200, 1);
> > Z = Optimize("Z", 1000, 1, 2000, 1);
> >
> > Sensitivity Tester for +/- 10%
> >
> > Senstivity = Optimize("Sensitivity", 1, 1, 5000, 1);
> >
> > X = Round((Random() - 0.5) * 2.99) + 10;
> > Y = Round((Random() - 0.5) * 20.99) + 100;
> > Z = Round((Random() - 0.5) * 200.99) + 1000;
> >
> > The results based on whatever fitness metric interests you can be
> > observed and sorted in the output of optimization and/or can be
> > exported for further statistical analysis (distribution) in Excel.
> >
> > --- In amibroker@xxxxxxxxxxxxxxx, "Steve Dugas" <sjdugas@xxxx> 
wrote:
> >> Hi Natasha,
> >>
> >> Here is the entire message intact, with thanks to Herman.
> >>
> >> Steve
> >>
> >> -----------------------------------------------------------------
--
> >>
> >> In addition to using 3D surface maps to evaluate the Robustness 
of
> > two or more parameters you may consider assessing the Robustness 
and
> > across-market dependency of single parameters over a Watch list.
> >>
> >> You can do this by superimposing equity lines for different
> > parameter values to see whether they track over time (fan-out) . 
This
> > is a far better method for evaluating systems and finding stocks 
than
> > picking a high equity from the Optimization result table. If the
> > equity ratios remain approximately constant your parameters are 
more
> > Robust than if the Equities do not track.
> >>
> >> In my example I use Steve Karnish's CMO5 system (symmetrical
> > triggers). To test this method on your own trading system you 
should
> > substitute your own code inside the curly brackets of the System()
> > function and substitute the variable named "TestParameter" for the
> > parameter you like to evaluate. Here is a typical chart showing
> > robust behavior:
> >>
> >>
> >>
> >> Here is a chart for a system that may be over optimized and lacks
> > robustness:
> >>
> >>
> >>
> >>
> >> Things to look for:
> >>
> >> 1) Over-optimization: High profits but Equity lines move all over
> > the place.
> >> 2) Robustness: Equity lines "fanout" keeping the system 
profitable
> > under most TestParameter values
> >> 3) Synchronized dips; if all equities dip you can create an 
Equity-
> > composite and use it to qualify signals.
> >> 4) Stepping through the stocks in you workspace you can observe
> > stock-synchronization
> >> 5) Is the optimum value optimum over time?
> >> 6) Straight equity lines
> >>
> >> Happy trading,
> >>
> >> Herman
> >>
> >>
> >>       Yahoo! Groups Sponsor
> >>             ADVERTISEMENT
> >>
> >>
> >>
> >>
> >> Send BUG REPORTS to bugs@xxxx
> >> Send SUGGESTIONS to suggest@xxxx
> >> -----------------------------------------
> >> 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
> >>
> >> Your use of Yahoo! Groups is subject to the Yahoo! Terms of
> > Service.
> >>
> >>   ----- Original Message ----- 
> >>   From: Natasha !!
> >>   To: amibroker@xxxxxxxxxxxxxxx
> >>   Sent: Monday, April 04, 2005 1:02 PM
> >>   Subject: [amibroker] Optimizing & Robustness of single 
parameters
> >>
> >>
> >>    Hi,
> >>          Just going over the archieves and i dug out a mail i 
didnt
> > quite well understand:
> >>   Since the attachments are not stored in the usergoups any graph
> > to illustrate the same would be much obliged;
> >>
> >>           Thanks.
> >>
> >>   -- N !!
> >>
> >>
> >>   In add From: "Herman van den Bergen" <psytek@>
> >>   Date: Thu Sep 4, 2003  9:44 am
> >>   Subject: Optimizing & Robustness of single parameters ition to
> > using 3D surface maps to evaluate the Robustness of two or more
> > parameters you may consider assessing the Robustness and across-
> > market dependency of single parameters over a Watch list.
> >>
> >>   You can do this by superimposing equity lines for different
> > parameter values to see whether they track over time (fan-out) . 
This
> > is a far better method for evaluating systems and finding stocks 
than
> > picking a high equity from the Optimization result table. If the
> > equity ratios remain approximately constant your parameters are 
more
> > Robust than if the Equities do not track.
> >>
> >>   In my example I use Steve Karnish's CMO5 system (symmetrical
> > triggers). To test this method on your own trading system you 
should
> > substitute your own code inside the curly brackets of the System()
> > function and substitute the variable named "TestParameter" for the
> > parameter you like to evaluate. Here is a typical chart showing
> > robust behavior:
> >>
> >>         Chart for the above and for the below
> > explanation.Thanks ???
> >>
> >>   Here is a chart for a system that may be over optimized and 
lacks
> > robustness:
> >>
> >>
> >>
> >>   Things to look for:
> >>
> >>   1) Over-optimization: High profits but Equity lines move all 
over
> > the place.
> >>   2) Robustness: Equity lines "fanout" keeping the system
> > profitable under most TestParameter values
> >>   3) Synchronized dips; if all equities dip you can create an
> > Equity-composite and use it to qualify signals.
> >>   4) Stepping through the stocks in you workspace you can observe
> > stock-synchronization
> >>   5) Is the optimum value optimum over time?
> >>   6) Straight equity lines
> >>
> >>   Happy trading,
> >>
> >>   Herman
> >>
> >>
> >>
> >>
> >>   Warm regards,
> >>   Natasha !
> >>
> >>
> >>
> >>
> >> -----------------------------------------------------------------
---
> > ----------
> >>   Yahoo! Messenger
> >>   Show us what our next emoticon should look like. Join the fun.
> >>
> >>   Please note that this group is for discussion between users 
only.
> >>
> >>   To get support from AmiBroker please send an e-mail directly to
> >>   SUPPORT {at} amibroker.com
> >>
> >>   For other support material please check also:
> >>   http://www.amibroker.com/support.html
> >>
> >>
> >>
> >>
> >>
> >> -----------------------------------------------------------------
---
> > ----------
> >>   Yahoo! Groups Links
> >>
> >>     a.. To visit your group on the web, go to:
> >>     http://groups.yahoo.com/group/amibroker/
> >>
> >>     b.. To unsubscribe from this group, send an email to:
> >>     amibroker-unsubscribe@xxxxxxxxxxxxxxx
> >>
> >>     c.. Your use of Yahoo! Groups is subject to the Yahoo! Terms 
of
> > Service.
> >
> >
> >
> >
> >
> >
> > Please note that this group is for discussion between users only.
> >
> > To get support from AmiBroker please send an e-mail directly to
> > SUPPORT {at} amibroker.com
> >
> > For other support material please check also:
> > http://www.amibroker.com/support.html
> >
> >
> > Yahoo! Groups Links
> >
> >
> >
> >
> >
> >
> >
> >





------------------------ Yahoo! Groups Sponsor --------------------~--> 
Has someone you know been affected by illness or disease?
Network for Good is THE place to support health awareness efforts!
http://us.click.yahoo.com/Rcy2bD/UOnJAA/cosFAA/GHeqlB/TM
--------------------------------------------------------------------~-> 

Please note that this group is for discussion between users only.

To get support from AmiBroker please send an e-mail directly to 
SUPPORT {at} amibroker.com

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/

<*> 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/