intermilan,

Backtest optimization is like that.
Given any set of stocks and timeframe, and a few indicators, I can  
optimize the performance (avoiding the losses due to temporary sharp  
changes in price due to news events), getting out at the peaks and  
valleys statistically most of the time (or even cheating and removing  
stocks that do not follow my rules).  However, the future conditions  
are not the same and the results are worse.  It is just the nature of  
20/20 hindsight.  And of course, let's not forget the self inflicted  
loses when the system does not meet expectations with a few loses, so  
you lose confidence and start to second guess the system with trades  
that are even worse than the system.

Backtesting should be more for finding systems that are not sensitive  
to fairly large changes in parameters, and various market  
conditions.  I am more interested in that side of things these days.

Dennis

On Aug 30, 2006, at 7:32 PM, intermilan04 wrote:

> I wish I was making 75% up to now :-)
> The 75% is the result of my system which is optimized between
> 2001-2006.  Since I'm always trying to improve my system, I don't
> necessarily have traded with the system verbatim from 2006/1/1.
>
> Now I am having an issue where as soon as I start using a system its
> performance drops :-D  but this is a whole another issue so I didn't
> mention about it here.
>
> intermilan04
>
> --- In [email protected], "sebastiandanconia"
> <[EMAIL PROTECTED]> wrote:
>>
>> Nope, I just meant that he measured all the other years from
>> Jan.1-Jan.1, so he's not comparing apples to apples by looking at YTD
>> performance.  We're coming into a time of year when there are
>> typically major drops followed by major rallies, so if his system
>> captures that behavior it could make up for its miserable 75% profit
>> up until now.:)
>>
>> I hear you about real DD's that exceed that of tested methods.   
>> That's
>> why I think it's so important to understand why a system works,  
>> beyond
>> simply the fact that it's tested-out well, which could just be a
>> mathematical coincidence, a meaningless correlation without any
>> cause-and-effect relationship.
>>
>>
>> Luck,
>>
>> Sebastian
>>
>> --- In [email protected], "Fred" <ftonetti@> wrote:
>>>
>>> "So, two things: First, the obvious one, you can't really know that
>>> your system has "broken down" until you get the final results on
>>> January 1, 2007.:)"
>>>
>>> Really ? ... You mean there is no point at which real DD's exceed
>>> previous experience you wouldn't think that system is broken ?
>>>
>>



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:
    [EMAIL PROTECTED]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 



Reply via email to