PureBytes Links
Trading Reference Links
|
Chuck,
I don't disagree with your evaluation. My comment was a very simple
one relating to the effect of combining FIXED POSITION SIZE and FIXED
NUMBER OF POSITIONS not the practicality of increasing bet size
instead of the number of positions. This probably all started way
back when all of us were aware that AB had no capability to restrict
numbers of positions so that one did not unintentionally exceed being
more than 100% invested when trading portfolios ergo the initial
necessity of PT or something like it. There have been some who
advocated the idea of testing based on the combo of the above. My
statement was and is that this method ?! at best hides both the good
and bad aspects of individual systems even if ones intention is to
invest in order to gain assets that one withdraws from the account
for living expenses. IMHO a 10% DD on account that is only 25%
invested is highly misleading unless because of market conditions or
available candidates one is only invested to that extent.
--- In amibroker@xxxxxxxxxxxxxxx, "Chuck Rademacher"
<chuck_rademacher@x> wrote:
> Sometimes, I find it difficult to write everything that I am
thinking on a
> complex subject. I would like to take another shot at why I think
it is
> better, for backtesting & evaluation purposes, to use profits to
increase
> the number of positions rather than the size of each trade.
>
> I feel that all we can do when developing a system is to try to
pick stocks
> (or other instruments) that are going to move in a particular
direction.
> IMO, it is only luck that a particular stock might go up
geometrically
> instead of just a few percentage points. Generally, I'm trying
to develop
> systems that pick a stock to go up. I've never been able to pick
stocks
> that are going to go up the most. This is largely due to the fact
that I
> don't know that the whole market is going to get overheated over
the next
> few days or weeks. I didn't know that 9/11 was going to happen,
but my
> systems sure made a lot of money on the short side. Due to the
disastrous
> effect of that event on so many people, I would hate to call
this "luck",
> but it was as far as trading systems are concerned.
>
> Based on hundreds of systems and thousands of trades, I know that
on average
> I'm going to have 48% to 52% winners with an average net profit of
4% to 6%.
> The last thing I want to see in my research is a system that
doesn't trade
> very often and accidentally picks a winner that makes 2000%. I
definitely
> don't want this outlier to be exaggerated by coming along right
after a big
> profit in another stock and using that profit to buy more of the
second
> stock. To me, this is "luck". It certainly has nothing to do
with the
> merits of the system.
>
> Therefore,
>
> When evaluating the merits of a system, I don't want to see any
compounding
> of profits.
>
> If I want to see how much money the system might have made for me,
I would
> use the profits to increase the number of positions.
>
> I cannot think of a single reason why I would want to see
compounding of
> position size in my backtesting results.
>
> But, that's just me.
------------------------ Yahoo! Groups Sponsor ---------------------~-->
Get A Free Psychic Reading! Your Online Answer To Life's Important Questions.
http://us.click.yahoo.com/Lj3uPC/Me7FAA/ySSFAA/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/
|