Ah, I see your point, that sounds like a good idea. Thanks!

  ----- Original Message ----- 
  From: Howard B 
  To: [email protected] 
  Sent: Saturday, June 06, 2009 1:25 PM
  Subject: Re: [amibroker] Re: Walk Forward Testing - Varies by Computer?




  Hi Steve --

  I have not looked into the details of the differences.  I have QP on one 
computer, Norgate Premium on another.  I like, and use, both.

  When there are differences in the test results run on two data series that 
should be, and appear to be, very close to each other, that is a flag to me 
that the system I am testing is probably unstable.

  Thanks,
  Howard



  On Sat, Jun 6, 2009 at 10:18 AM, Steve Dugas <[email protected]> wrote:





    Hi Howard - between QP and Norgate, just wondering if you ever followed up 
on the differences in data and if you found one to be noticably cleaner than 
the other? Thanks very much!

    Steve

      ----- Original Message ----- 
      From: Howard B 
      To: [email protected] 
      Sent: Saturday, June 06, 2009 12:41 PM
      Subject: Re: [amibroker] Re: Walk Forward Testing - Varies by Computer?


      Greetings --

      Check the settings being used.

      And test runs can be surprisingly sensitive to small differences in the 
data being used.  I regularly find significant differences between runs made 
using Norgate Premium data and Quotes Plus data, although when I plot the two 
price series, I cannot see obvious differences.  

      Thanks,
      Howard


      On Thu, Jun 4, 2009 at 5:03 PM, toddk63 <[email protected]> wrote:




        I just debugged a similar problem. Same system, same database, 
different answer. Turned out to be the "Limit Trade Size" option on the 
"Settings" dialog box. Default is 10, I need to set it to 0. I wish there were 
a function like SetOption("LimitTradeSize"), but there is not...



        --- In [email protected], "jacklweinberg" <jacklweinb...@...> 
wrote:
        >

        > Question: 
        > How is it possible for a walkforward test to produce different 
results on two different computers?
        > Background:
        > 1. AB program, database, system, watchlist, walkforward settings etc. 
are on a usb key that is moved from one computer to another.
        > 2. Before the walk forward test, the same walkforward settings file 
is loaded. The same system is executed. The results are also put into the same 
usb key.
        > 
        > Why do two different computers produce different results? (and 
strangely enough, very different results!)
        >








  

Reply via email to