How AB count depends on backtester logic.

I noticed in my reversal system on futures that in the same bar 
backtester first close open position on stop and then open reversal depends on 
regular signal .

This order is even then if reversal signal is first.
APPlystop from backtester point of view has highest priority for counting than 
regular exit from the system.
Backtester do not check it and ignore that.
If money management is your priority /and should be/ results from backtester 
should be perfect, but they don't.I must count it "by hand", and correct it in 
excel.
This is "dark side " of AB from my point of view.

Maybe those who programing custom backtester can do solve the problem. I can't.
I test AB ,AFL is OK  and I thought that would buy license to use it and 
replace metastock
but play with low level programing for correct results is not for me.

Pit


--- In amibroker@yahoogroups.com, Radek Simcik <radek.sim...@...> wrote:
>
> ok,
> 
> Thank you, I think I understood now but one question is still unanswered:
> 
> What would happen if both happened during the day and my trailing stop is %4.
> -the price went up more than %4 and
> -low more than 4%.
> 
> It seems to me that I am going to experience gain every time although
> we cannot say for 100% what
> happened inside the bar.
> 
> Thank you,
> 
> Radek
> 
> On Fri, Mar 20, 2009 at 12:58 PM, Radek Simcik <radek.sim...@...> wrote:
> > Hi,
> >
> > could anybody explain to me how trailing stop works?  I have below
> > setting in my system ( I short on open and exit the same day on
> > close):
> >
> > ApplyStop(stopTypeTrailing,stopModePercent,4 ,1);
> > SetOption ("ActivateStopsImmediately", True);
> >
> > in that case it seems to me that stop starts to trail to protect my
> > gain only if the gain is more than 4%. I thought that the trade would
> > be exited if price fell more than 4% below low.
> >
> > What would happen if both happened during the day: the price went up
> > more than %4 and low more than 4%. It seems to me that I am going to
> > experience gain every time although we cannot say for 100% what
> > happened inside the bar.
> >
> > Could somebody explain me how ApplyStop(stopTypeTrailing,3,4 ,1);
> > work? It gave me just unbelievable results.
> >
> > Thank you,
> >
> > Radek
> >
>


Reply via email to