|
PureBytes Links
Trading Reference Links
|
DT,
Insightful investigation into DEMA.
The phenomenon however is not related to variable
periods as you get the same results if you replace the variable with a fixed
value.
I would assume the DEMA results for the first 2*len
periods would be erroneous - by definition....
<BLOCKQUOTE
>
----- Original Message -----
<DIV
>From:
Dimitris
Tsokakis
To: <A title=amibroker@xxxxxxxxxxxxxxx
href="">amibroker@xxxxxxxxxxxxxxx
Sent: Saturday, April 26, 2003 5:14
AM
Subject: [amibroker] Built-in DEMA
questions
Some things should be explained about DEMA and its expected
behavior:Let us use a fast DEMA [green line, period=10] and a slow one
[yellow line, period=10]A third, variable period DEMA [the black thicker
line], uses the fast period and, suddenly, 200 bars ago, changes to
slow.The result will immediately leave the green curve, but asymptotically
touch the yellow one. It will not be a real 50-bar DEMA, even after 200
bars.Note that all this 200-bar period the variable DEMA stays higher than
the slow one. [above gif]To answer my [obvious] questions, I search the
available reference [instead of leaving my imagination to fly around].The
explanatory note via <A
href="">http://groups.yahoo.com/group/amibroker/message/38844
sends me to <A
href="">http://www.amibroker.com/guide/afl/afl_view.php?name=DEMAwhere
I read the comment
DEMA internally is implemented via
EMA:Len=10;Graph0= 2 * EMA( C, len ) - EMA( EMA( C, len ), Len
);// for comparison onlyGraph1=DEMA(C,Len);
Some more explanations should be added
because:a. The built-in DEMA
begins from the first bar and the analytic EMA equivalent begins 400 bars
later. How the built-in DEMA is defined for the first 400 bars ? [The
question holds for IB and AA, see the simple exploration
Len=200;G0= 2 * EMA( C, len ) - EMA( EMA( C, len ), Len
);G1=DEMA(C,Len);Filter=1;AddColumn(G0,"ANALYTIC
DEMA");AddColumn(G1,"DEMA");AddColumn(Cum(1),"BARS",1.0);
and the att. gif]
b. Should we
expect a 2*Len delayed response when variable periods is used
?c. My CSCO history is 831
bars [Jan3, 2000 till now]. The Graph0= 2 * EMA( C, len ) - EMA( EMA( C,
len ), Len );begins on Aug2, 2001.c1.Can we
speak for a proper function use back in the critical Oct2, 2001 CSCO
low?The built-in DEMA is 11.53 and the EMA equivalent is 9.08
!!!!!c2.After how many bars the two formulas
are less than 1% diverging ?
Built-in DEMA is, IMO, one of the most powerful AFL
tools, its use in trading systems design gives excellent resultsand this
is accurately translated to interesting profits. To avoid any
misunderstanding [the recent winds are S to SW, not that healthy...]my
questions do not search for built-in functions secrets, it would be useless to
me.
Just a better explanation to understandthe
expected DEMA behavior would encourage users/traders to step into
this great function. 5-10 lines would be more than enough, no need for
1000+ pages.Thank you in advance for any reply.Dimitris
TsokakisSend
BUG REPORTS to bugs@xxxxxxxxxxxxxSend 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 <A
href="">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.
|