|
PureBytes Links
Trading Reference Links
|
Hi Ken –
In my opinion, the “symbol space”
is just one more variable to be searched.
If I look at a moving average crossover
system and test all lengths from 1 to 100 in steps of 1 for each of two moving
averages, I have searched 10000 values. Some will be profitable, some
not. Some segments of the space will be stable, some not.
Similarly, if I search 10000 tickers, some will be profitable, some not.
Some groups of tickers (perhaps industry groups) will trade profitably, others
not. Just as I might be comfortable trading the moving average system if
a parameter set is stable, I might be comfortable trading a subset of the
universe of tickers. If the subset that appears to trade well is
logically cohesive, I have more confidence than if it consists of a random
collection of company types.
I sometimes do early phase tests on very
small ticker lists, but I run the risk of not recognizing that a system works
on an industry type. If my test subset of tickers does not include, say, any
money center banks, then I am unlikely to find out that a particular trading system
works well for most money center banks.
I try to keep in mind that selecting only
a small subset of tickers is simply an example of “optimizing over the
symbol space”.
Howard
-----Original Message-----
From: Ken Close
[mailto:closeks@xxxxxxxx]
Sent: Tuesday, November 18, 2003
9:03 AM
To: AmiBroker List
Subject: [amibroker] Optimizing
Selection Criteria vs Timing Signals
Here is a (perhaps) semi philosophical question:<font
size=2 face="Courier New">
We are familiar with the various pitfalls and
prohibitions in optimizing
and "curve-fitting" timing systems.
Do the same pitfalls and prohibitions exist for
keeping a constant
timing signal (whatever it is) yet optimizing
variables that alter
basket selection criteria? Can you
"curve fit" a set of selection
criteria?
It "feels" like the problems would be
different (and perhaps minimized)
because of the dynamic changes that occur,
ESPECIALLY if stop conditions
are used. OTOH, I suppose it would not be
different using the same
criteria on the same basket of stocks over the
same time period as this
would yield the same baskets of stocks each
time. What I mean by that
is that for a given combination of criteria, the
same stocks would be
selected.
Hard to visualize.
Any comments on this aspect?
Ken
<font size=2
face="Courier New">Send BUG REPORTS to
bugs@xxxxxxxxxxxxx<span
>
Send 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 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.
|