|
PureBytes Links
Trading Reference Links
|
<SPAN
class=780275118-08022004>No confusion, that's exactly what I thought you meant
(:-). I'm just ruminating about what that says about the underlying market
behavior.
<SPAN
class=780275118-08022004>
<SPAN
class=780275118-08022004>It's an interesting notion, not obviously true or
untrue, to me anyway, that the market as a whole has changing periodicities that
apply across many stocks. Whether it's tradeable or not (probably not), it's
more clear that individual stocks have changing time constants. It doesn't
strike me as inherent that different stocks would be synchronized with each
other at all, or why that would be true if it was.
<SPAN
class=780275118-08022004>
<SPAN
class=780275118-08022004>Just to be clear, I'm
not disagreeing in any way with the idea that tweaking parameters per stock
could be destructive overfitting, or that re-optimizing parameters on a
whole-market or sub-market level might generate profitable results. Kind of
rolling these ideas around in my head and
wondering...
<SPAN
class=780275118-08022004>
<SPAN
class=780275118-08022004>Dave
<BLOCKQUOTE dir=ltr
>
I'm
think my comments are, at best, ambiguous and I apologise for the
confusion. The confusion came about due to talking about "dynamic
parameters" vs. "stock-specific parameters".
<FONT face=Arial color=#0000ff
size=2>
While
I would really discourage anyone from using a different parameter set for each
stock, I would recommend using identically calculated dynamic
parameters for each bar for each stock.
<FONT face=Arial color=#0000ff
size=2>
Now,
that statement may lead to even more confusion unless I give some examples and
these are only my opinions:
<FONT face=Arial color=#0000ff
size=2>
Good
example:
<FONT face=Arial color=#0000ff
size=2>
<FONT face=Arial
color=#0000ff size=2>5/15 cross for all stocks
<FONT face=Arial color=#0000ff
size=2>
Bad
examples:
<FONT face=Arial color=#0000ff
size=2>
<FONT face=Arial
color=#0000ff size=2>5/15 cross for IBM
<FONT face=Arial
color=#0000ff size=2>5/17 cross for MSFT
<FONT face=Arial
color=#0000ff size=2>4/13 cross for INTC
<FONT face=Arial
color=#0000ff size=2>etc.
<FONT face=Arial color=#0000ff
size=2>
Good
example using a percentage of the Hilbert calculated cycle length on any
day:
<FONT face=Arial color=#0000ff
size=2>
<FONT
color=#0000ff size=2>0.25% / 0.75% cross for all
stocks
<FONT face=Arial color=#0000ff
size=2>
Bad
examples:
<FONT face=Arial color=#0000ff
size=2>
<FONT face=Arial
color=#0000ff size=2>0.25% / 0.75% cross for IBM
<FONT face=Arial
color=#0000ff size=2>0.33% / 0.80% cross for MSFT
<FONT face=Arial
color=#0000ff size=2>etc.
<BLOCKQUOTE
>
<FONT face="Times New Roman"
size=2>-----Original Message-----From: Dave Merrill
[mailto:dmerrill@xxxxxxx]Sent: Sunday, February 08, 2004 8:38
AMTo: amibroker@xxxxxxxxxxxxxxxSubject: RE:
[amibroker] Determining cycles
<SPAN
class=322173113-08022004>As I suspected.
<SPAN
class=322173113-08022004>
<SPAN
class=322173113-08022004>Just for my understanding then, what you mean by
"adjusting the lookback period" is changing the period for EMAs or other
functions over time, but using the same value for all issues? Thinking
through the underlying trading idea here, that'd say that the "frequency" of
the overall market changes, but that there's some degree of consistency
across various issues at any given time?
<SPAN
class=322173113-08022004>
<SPAN
class=322173113-08022004>Dave
<BLOCKQUOTE dir=ltr
>
<FONT face=Tahoma
size=2>-----Original Message-----From: Chuck
Rademacher
<FONT face=Arial color=#0000ff
size=2>Dave,
<FONT face=Arial color=#0000ff
size=2>
<FONT face=Arial color=#0000ff
size=2>IMO, that would be extreme overfitting. It takes about
fifteen minutes doing out-of-sample research to convince me that
individual parameters for each (or even a few) stocks would make a serious
dent in my assets. I feel this way regardless of whether you are
talking about RSI, MA, EMA, whatever. To me, the beauty of
trading stocks instead of 40 futures markets is that you can find a
single parameter set that works across thousands of stocks generating
thousands of trades. I like statistics on my
side!
<BLOCKQUOTE
>
<FONT face="Times New Roman"
size=2>-----Original Message-----From: Dave Merrill
[mailto:dmerrill@xxxxxxx]Sent: Saturday, February 07, 2004
4:25 PMTo: amibroker@xxxxxxxxxxxxxxxSubject: RE:
[amibroker] Determining cycles
<SPAN
class=031482321-07022004>Chuck, just out of curiosity, why do you say
you'd NEVER use different periods for each stock? Is your concern there
with overfitting? Something else?
<SPAN
class=031482321-07022004>
<SPAN
class=031482321-07022004>Dave
<BLOCKQUOTE dir=ltr
>
<FONT face=Arial color=#0000ff
size=2>Ara,
<FONT face=Arial color=#0000ff
size=2>
<FONT face=Arial color=#0000ff
size=2>I should tell you that I have never found moving averages to
work well, over time, except for market timing using something like a
5/15 EMA crossover. Every attempt that I've made to find one
moving average across a basket of stocks has not been
profitable. Dynamically adjusting the lookback periods,
however, does improve things considerably. For the record, I
would NEVER use different lookback periods for each
stock.
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
Yahoo! Groups Sponsor
ADVERTISEMENT
Yahoo! Groups Links
To visit your group on the web, go to:http://groups.yahoo.com/group/amibroker/
To unsubscribe from this group, send an email to:amibroker-unsubscribe@xxxxxxxxxxxxxxx
Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.
|