[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [RT] Re: Failure of studies/patterns



PureBytes Links

Trading Reference Links



Clyde, I think this is some kind of misunderstanding.
 
MESA deals with dominant cycles, this is not a tool that calculates the cycles more accurately. It simply deals with not ideal cycles.
 
For example, it deals with the cycles whose period and/or phase changes. This is the main reason why MESA has been developed. The most tradable cycles mostly exist as dominat cycles (as cycles change constantly), the stable cycles are economical cycles only (like Kitchen or Juglar cycles).
 
You may see about this difference here: http://www.timingsolution.com/TS/Articles/Saga_2/
 
This fact changes the whole situtation drastically:  we constanly should recalculate spectrogrtam (when the new price history is coming). And we cannot apply classical back testing here.
 
Best regards.
Sergey.
 
 
----- Original Message -----
Sent: Sunday, October 18, 2009 12:47 PM
Subject: Re: [RT] Re: Failure of studies/patterns

 



Here is the type of study you mentioned but in this case
only 10 years of data is used and the "normal" Fourier
synthesis is used and results are from four periods.
 
Clyde
 
----- Original Message -----
From: Gene Pope
Sent: Sunday, October 18, 2009 6:29 AM
Subject: RE: [RT] Re: Failure of studies/patterns

Thank you Clyde. I have a question. After having determined the single frequency candidates using MESA and optimizing the shifting as indicated, if we went back and instead performed a brute optimization of all frequencies (including optimizing shifting), wouldn’t we get similar results? In other words, are there more than just these three frequencies that could be “made” to work and if so, what does that tell us?

The reason I ask is that it appears your method “may” be tying into the natural tendency of the markets to rise over longer periods (to varying degrees) and I wonder out loud if the distribution of frequencies couldn’t be more random and still work.

Regards,

Gene

From: realtraders@yahoogroups.com [mailto:realtraders@yahoogroups.com] On Behalf Of Clyde Lee(swb)
Sent: Sunday, October 18, 2009 12:16 AM
To: realtraders@yahoogroups.com
Subject: [RT] Re: Failure of studies/patterns

 



This discussion turned into a discussion of fixed period

"waves" based on a non-analytic method.

I prefer an analytic method for examining periods of

waves and use the Maximum Entropy method to measure

these periods.

However, I take a simple approach for fixed wave projections

and use a trading model based on fixed Cosine waves and

trade at inflection points which are optimized for lead/lag of

the fixed wave over a very long period of time.

The attached pictures indicate that MESA analysis says that

there are 82 month, 108 month, and 262 month LONG

period waves in the DJIA from 1921 to now.

The results of each model and the composite model are

shown in the attached pictures.

Have fun ! ! !

Clyde



__._,_.___


Your email settings: Individual Email|Traditional
Change settings via the Web (Yahoo! ID required)
Change settings via email: Switch delivery to Daily Digest | Switch to Fully Featured
Visit Your Group | Yahoo! Groups Terms of Use | Unsubscribe

__,_._,___