[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[amibroker] Re: Question about enhanced backtest reports



PureBytes Links

Trading Reference Links

With regards to PositionScore= statement, see the ReadMe from the 
latest beta.

With regards to continuous contracts, maybe not MORE accurately but I 
wouldn't think less accurately either, especially if they are 
properly created.

--- In amibroker@xxxxxxxxxxxxxxx, "Owen Davies" <owen5819@xxxx> wrote:
> Fred wrote:
> 
> 
> > Do you have a PositionScore = statement ?
> 
> No.  What is it?  The term does not show up in a search of the AB 
help file.
> 
> > With regards to the rest, imho futures related systems are better
> > evaluated using continuous contract data which can 
be "manufactured"
> > using a variety of techniques.
> 
> Well, more conveniently, at least.  However, it is hard to imagine 
that test
> results from a continuous contract could reflect real-world trading 
more
> accurately than results from unaltered data would.
> 
> In fact, my experience is that with continuous data profits end up 
looking
> one-fourth to one-third better than on real contracts, and 
drawdowns are
> reduced by that much or more.  The effect seems worst on short-term
> techniques, perhaps in part because you have more trades over which 
to
> cumulate small errors.  And when a contract break is not faired 
adequately
> it can trigger a spurious entry and create the appearance of a 
profit on the
> closed trade.  For long-term trend following, these problsm also 
hide the
> effects of rolling over from one contract to the next.  It could be 
that I
> just don't look at long-term techniques often enough to know that 
the
> distortions with them are just as bad as those in tests of swing 
trading.
> 
> Just as one example, I used to have two continuous contracts for 
the S&P.
> On one short-term system, the tests gave a profit of around +640 
points with
> one contract and a loss of nearly -280 points with the other.  (The 
system
> did not trade very often, or well.)  The contracts covered exactly 
the same
> five-year period, but had been calculated with different methods.  
At this
> point, I do not recall exactly how they were grafted together; it 
was some
> years ago.  The methods were built into the Unfair Advantage 
package.  I had
> already suspected that continuous contracts introduced unacceptable
> problems.  This test convinced me that they could be a lot worse 
than I had
> imagined.
> 
> Nonetheless, I do test with continuous contracts as a first screen, 
on the
> theory that if a system won't appear to make money on them, it has 
no hope
> of making it on real contracts.  If I ever run into a technique 
that would
> look better on real contracts than on continuous ones I will miss an
> opportunity, but this is a compromise I can live with.
> 
> If your results vary significantly from mine, I would love to hear 
more
> about your use of continuous contracts.
> 
> Thanks for your thoughts.
> 
> Owen


------------------------ Yahoo! Groups Sponsor ---------------------~-->
Rent DVDs Online - Over 14,500 titles.
No Late Fees & Free Shipping.
Try Netflix for FREE!
http://us.click.yahoo.com/vhSowB/XP.FAA/3jkFAA/GHeqlB/TM
---------------------------------------------------------------------~->

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 http://docs.yahoo.com/info/terms/