I am new to AB and am by no means the most qualified to answer this question
but I have found great use in introducing noise in to the key variables of a
system so as to see how well a system holds up when different
trades/stocks/conditions are dropped out. This is probably most useful when you
are in portfolio mode and considering many stocks at once. I use a dummy
Optimize("steps",1,1,500,1) for instance to set up how many steps I want to go
through, and multiply whatever I want to modulate with one of AB's random
functions. I then take stats from all the different iterations AA spits outs
and usually look at frequency distributions of things like car/mdd in Excel. I
save the excel, then when I make a major change to the system I run a new
pseudo Monte Carlo and compare the new results with the old. I have also found
this technique to be useful when running in tandem with a real Optimization,
you can see what Optimize values seems to appear the most throughout the MC
noise, I have found this sort of approach tends to aid itself to better walk
forwards.
--- In [email protected], "shahariar4" <shahari...@...> wrote:
>
> Dear Frineds:
>
> Pardon me for my ignorance...there has been much talked about the Monte carlo
> analysis here and also converting ami data to tradesim etc...
>
> But is anyone here doing the analysis through Ami? Or is it at all possible
> in Ami (independent of tradesim help or others)?
>
> I am interested to apply this and would be happy if anyone who is working
> with monti carlo comes up with some way for me with Amibroker...
>
>
> I am aware of positionscore of Ami and how close is the function compared to
> Monit Carlo analysis?
>
> Thank you in advance,
>
> Shahariar
>
