|
PureBytes Links
Trading Reference Links
|
Hi Fred,
Thank you for your suggestion & comment.
- Drawdowns
28% is certainly not low, but I actually have a 3% stop order built
into my system. Since it is a daytrade system, each day the maximum I
lose is 3%. (the 28% system drawdown is the result of series of losses
in several days)
So I am personally not too worried about it; it is a soft-landing
rather than a hard-landing system.
- Objective Testing
I used to optimize my system with past 3 years of data(2003-2005). My
CAR was 400% yet I was going broke. Why? Because my system
encountered the infamous "curve-fitting" problem, where it generates
impressive CAR in a given time frame, but completely fails to work
outside the time frame.
So I re-optimized my system with 10-year frame and, my return is now
at 110%.
I have also optimized with half data (2001 to end of 2005) and did
some forward-testing between 2006/1/1 - 2006/4/13, but the result
wasn't very good. In my opinion, 5 years of data isn't enough to
escape from the curve-fitting problem.
As far as I'm forward-testing the 10-year-optimized system, I'm
getting 98% return between 2006/1/1 and 2006/4/13. So far, it is
working just the way it is supposed to.
Regards,
intermilan04
--- In amibroker@xxxxxxxxxxxxxxx, "Fred" <ftonetti@xxx> wrote:
>
> A Comment and a suggestion ...
>
> - DrawDowns ... I could be wrong but I suspect most people can't
> tolerate 28% DD's ... To bring that number down to the point where
> at least some people would be comfortable with it using real money
> one would I think have to cut it half. Doing that with an existing
> system by restricting how invested one is will result in the CAR
> being reduced to the square root of its original number.
>
> - Objective Testing ... Take your data, cut in half ... Optimize
> your system over half of the data and then test the parameter values
> on the other half. This rudimentary view of out of sample testing
> will give you some idea of what you are likely to experience in real
> trading as opposed to totally in sample results.
>
> --- In amibroker@xxxxxxxxxxxxxxx, "intermilan04" <intermilan04@>
> wrote:
> >
> > Since I have optimized my system between 1996-2006, I guess the
> > answer would be the same time period.
> >
> > --- In amibroker@xxxxxxxxxxxxxxx, "Fred" <ftonetti@> wrote:
> > >
> > > That doesn't answer my question ...
> > >
> > > In the development of the system what range of data ( time
> period )
> > > did you use ? The same time period ? An earlier one ?
> > >
> > > --- In amibroker@xxxxxxxxxxxxxxx, "intermilan04" <intermilan04@>
> > > wrote:
> > > >
> > > > The numbers are the result of backtesting my system with
> > > > NASDAQ and NYSE tickers (around 7000 tickers) between
> > > > 1996/1/1~2006/1/1.
> > > >
> > > > --- In amibroker@xxxxxxxxxxxxxxx, "Fred" <ftonetti@> wrote:
> > > > >
> > > > > Are the numbers you posted in sample or out of sample ?
> > > > >
> > > > > --- In amibroker@xxxxxxxxxxxxxxx, "intermilan04"
> <intermilan04@>
> > > > > wrote:
> > > > > >
> > > > > > I know it depends on what you want personally for
> risk/reward,
> > > but
> > > > > I'm
> > > > > > curious as to what other people's systems (developed in
> > > Amibroker)
> > > > > are
> > > > > > performing like. You don't have to share your code or the
> idea
> > > behind
> > > > > > your system (unless you want to), but I'm curious.
> > > > > >
> > > > > > Over the last 10 years, say, what is your annual profit %,
> max
> > > > > > drawdown, % winning trades, etc.?
> > > > > >
> > > > > > I have a long system that has returned around 110% since
> > > 1996. Its
> > > > > > winning % is 47%, and the system drawdown is 28%. It is a
> > > > > > reversal-based, swing-daytrade system.
> > > > > >
> > > > >
> > > >
> > >
> >
>
------------------------ Yahoo! Groups Sponsor --------------------~-->
GFT Forex Trading Accounts As low as $250 with up to 400:1 Leverage. Free Demo.
http://us.click.yahoo.com/lpv1TA/jlQNAA/U1CZAA/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/
|