|
PureBytes Links
Trading Reference Links
|
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/
|