|
PureBytes Links
Trading Reference Links
|
<SPAN
class=050450521-12052003>Siva,
I hope I
understand this correctly.
<SPAN
class=050450521-12052003>
If you time
compress right to left:
<SPAN
class=050450521-12052003>
On Monday, you do
a backtest, you get certain results based on the data as of
Monday.
On Tuesday, the
bars are now different, based on different periods. If you backtest again
based on Tuesday, would not the results be entirely different because all the
past bars have now changed?
<FONT face="Vladimir Script" color=#000080
size=5>Rick
<FONT face=Tahoma
size=2>-----Original Message-----From: siva_26sg
[mailto:siva_26sg@xxxxxxxxx]Sent: Sunday, May 11, 2003 12:04
PMTo: amibroker@xxxxxxxxxxxxxxxSubject: [amibroker] Help
needed - Back test shows unbelivable
resultsHelloI need help and any advice I can
get on the following. I'm in a state of shock!!!Use of Synthetic
bars (timeframe compression) in back testing and its validityJust
to make sure what I mean by synthetic bars. I have written a routine that
will compress the daily bars to any number of bars – example for weekly; I
can compress it by 5 bars and for monthly 22 bars. The compression is done
from right to left which is different from normal weekly and monthly
compression, which is done on calendar day end of a week or month.
Synthetic bar compression is dynamic and each day (new bar) on the right
edge is used for re-computing. Example if today is Monday than the weekly
(5 bar) compression will start from previous Tuesday and if today is
Tuesday than the weekly start will be previous Wednesday…. and so on. You
know what I mean.Questions I have is that, Is this a valid method
of compression for dynamic time frame?Creating EMA of this compressed
bar will provides dynamic weekly EMA, which can be used in a daily chart.
Though there is a slight difference between true weekly EMA and this, is
it a valid method for analysis? I have also noticed that this method has
lesser lag and it is more responsive.Does this method of
compression violate the rule that we should not look ahead for back
testing?Will the dynamic nature of the compression invalidate back
test result for historical data?The reason I'm asking these
questions is that when I back tested this with a simple trading rule for
entry and exit using a 22-day compression with 3 period EMA (this
simulates 66 day EMA), I was getting results that I couldn't believe my
eyes!!! I have tested with 400 counters over an average of 10 years of
data and 395 came out with 0 losing trades and the remaining 5 with 2 to 3
losing trades. Though the drawdown was sometimes high, the overall profits
and winning trades are just out of this world.Since this to me
seems too good to be true, I would appreciate your expert opinion and
guidance.Thanks in advanceRegardsSivaSend
BUG REPORTS to bugs@xxxxxxxxxxxxxSend SUGGESTIONS to
suggest@xxxxxxxxxxxxx-----------------------------------------Post
AmiQuote-related messages ONLY to: amiquote@xxxxxxxxxxxxxxx (Web page: <A
href="">http://groups.yahoo.com/group/amiquote/messages/)--------------------------------------------Check
group FAQ at: <A
href="">http://groups.yahoo.com/group/amibroker/files/groupfaq.html
Your use of Yahoo! Groups is subject to the <A
href="">Yahoo! Terms of Service.
Yahoo! Groups Sponsor
Send BUG REPORTS to bugs@xxxxxxxxxxxxx
Send SUGGESTIONS to suggest@xxxxxxxxxxxxx
-----------------------------------------
Post AmiQuote-related messages ONLY to: amiquote@xxxxxxxxxxxxxxx
(Web page: http://groups.yahoo.com/group/amiquote/messages/)
--------------------------------------------
Check group FAQ at: http://groups.yahoo.com/group/amibroker/files/groupfaq.html
Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.
|