Hi everybody, I just want to emphasize - and to get your feedback - speed increase is not the only benefit - more important, I think, is the quality of the optimization. Remember, any optimizer will just maximize a function on a finite interval, and given enough parameters, the only way to prevent against curve fitting is more data. My feeling is that these parameter relationships help a bit against curve fitting.
Regards, Alex --- In [email protected], Keith McCombs <kmcco...@...> wrote: > > Tomasz -- > Sorry, I ran only a short test. Therefore, only saw AA's spread sheet > results. And didn't notice the speed increase. > > Would it be possible to exclude the posting of null results from AA's > spread sheet as well? > -- Keith > > Tomasz Janeczko wrote: > > > > It limits the number of iterations to the extension required to > > actually FIND OUT that you want to skip this. I.e. the formula needs > > to be evaluated to find out that "Exclude" variable is true. > > > > The "exclude" tells the backtester NOT to collect signals and not to > > perform actual portfolio backtesting. > > This greatly increases the speed. > > > > Best regards, > > Tomasz Janeczko > > amibroker.com > > > > ----- Original Message ----- > > *From:* Keith McCombs <mailto:kmcco...@...> > > *To:* [email protected] <mailto:[email protected]> > > *Sent:* Tuesday, March 17, 2009 9:37 PM > > *Subject:* Re: [amibroker] Re: Optimization: relationship between > > parameters > > > > Tomasz -- > > I just tried your suggestion. It does not limit the search or the > > number of iterations. It just sets the results of the excluded > > tests to zero. > > -- Keith > > > > redberryys wrote: > >> > >> Oh, great, thank you, Tomasz! > >> --- In [email protected] > >> <mailto:[email protected]>, "Tomasz Janeczko" > >> <groups@> wrote: > >> > > >> > Hello, > >> > > >> > You can do this using this statement: > >> > > >> > Exclude = FastPeriod > SlowPeriod; > >> > > >> > Best regards, > >> > Tomasz Janeczko > >> > amibroker.com > >> > ----- Original Message ----- > >> > From: "redberryys" <redberryys@> > >> > To: <[email protected] > >> <mailto:amibroker%40yahoogroups.com>> > >> > Sent: Tuesday, March 17, 2009 1:55 PM > >> > Subject: [amibroker] Optimization: relationship between parameters > >> > > >> > > >> > > Hi, > >> > > Is there any possibility to specify a relationship between > >> optimization parameters? Many times, such a relationship makes > >> sense - > >> > > for example, the length of a slow avg should be bigger than > >> the length of a fast one. > >> > > Having relationships specified would limit the search space > >> and provide more logical solution, reducing curve fitting as well. > >> > > > >> > > If it is not possible, - for Tomasz - how hard would it be to > >> add this? > >> > > > >> > > Thank you, > >> > > Alex > >> > > > >> > > > >> > > > >> > > ------------------------------------ > >> > > > >> > > **** IMPORTANT PLEASE READ **** > >> > > This group is for the discussion between users only. > >> > > This is *NOT* technical support channel. > >> > > > >> > > TO GET TECHNICAL SUPPORT send an e-mail directly to > >> > > SUPPORT {at} amibroker.com > >> > > > >> > > TO SUBMIT SUGGESTIONS please use FEEDBACK CENTER at > >> > > http://www.amibroker.com/feedback/ > >> <http://www.amibroker.com/feedback/> > >> > > (submissions sent via other channels won't be considered) > >> > > > >> > > For NEW RELEASE ANNOUNCEMENTS and other news always check DEVLOG: > >> > > http://www.amibroker.com/devlog/ > >> <http://www.amibroker.com/devlog/> > >> > > > >> > > Yahoo! Groups Links > >> > > > >> > > > >> > > > >> > > >> > > >
