|
PureBytes Links
Trading Reference Links
|
Hi Ara,
Wednesday, February 1, 2006, 7:09:35 AM, you wrote:
AK> but you say the numbers look OK .... ???
There is no question that it works, Ara. Thanks to everyone for
contributing to this.
I do not know why, the first time I ran the backtest (after creating
the composite), I got no trades. I do know that this happened once
and only once, however. (Maybe I had a line wrap problem or
something, as I did "beautify" the code between 1st and subsequent
attempts.) But I have not had a miss since, and have been playing
around with the setups required part of the buy statement, and
getting the correct results based on that in subsequent backtests.
So I believe this code is okay. It seems, rather disappointingly
however, to indicate that number of setups has little or no bearing
on system metrics, until the numbers fall so far off either side that
the resulting number of trades becomes statistically unreliable.
Rather surprisingly, those falloffs, meaningless as they are, would
tend to point toward a smaller percentage of basket setups as
resulting in better outcomes than an overwhelming percentage of
basket setups. But, we are getting to the very edges of low and high
percentages of basket setups, and that dramatically limits the number
of trades -- too dramatically for me to say anything or assume
anything definitive about the numbers.
I'd am going to try and take this one step further. I have opening
criteria that can either allow an EOD setup from the previous day to
live, or that can extinguish that setup immediately. Thus, the setup
count I am putting in the composite at the moment is not always the
same number that the backtester will "see" after the next day's open.
The number will never be larger, but it is rather often smaller. I
believe I can modify the code to populate the composite with "net
setup" numbers without much problem. Excuse me if I holler if I get
into trouble doing that. ^_^
The final step I'd like to take is to measure the percentage of "net
setups" that become signals, and use that percentage as a backtest
variable to see if I can draw any conclusions. Here, I have strong
casual observational evidence that a high percentage of setups
becoming signals is a good barometer. But, sometimes "casual"
evidence can be misleading, so I'd like to test that, too. But I
have little idea how to do that.
Thanks again for the help.
Yuki
------------------------ Yahoo! Groups Sponsor --------------------~-->
Try Online Currency Trading with GFT. Free 50K Demo. Trade
24 Hours. Commission-Free.
http://us.click.yahoo.com/RvFikB/9M2KAA/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/
|