Hurst_Channel_Four

Conventional centered moving average computation of Hurst Channels
 AND
uses Fourier Techniques to extend computation to current (and future if desired) time.

Versions for:  TS2000i
®/TS6®
 

For a time study click on this line

To see how indicator is applied, Click on this line

The routine is currently available and can be ordered either from PayPal or CCNow
See below for "clicks" to order AND further down for more info on routine.

 

 

To order from PayPAL
CLICK  THE    Add to Cart   Button TO ORDER the
Hurst_Channel_Four @ $500.00
   


To order from  CCNow
Product ID: HCF-002    Cost: $500.00

Product Name: Hurst_Channel_Four
Click on: 
Buy It  to Order

 



HERE IS A PICTURE OF TS4.0 VERSION APPLIED TO A THEORETICAL MARKET

CY2254_hurst_4-fourline_4.gif (33706 bytes)

 

The following picture displays the output of the    Hurst_Channel_Four  indicator for TS2k and TS6.  The non-black lines are the computed CENTERED moving averages for lengths equal to 1/2 the period of the indicated cycles -- See the FOURLINE results at the bottom of the chart.  The black lines are future values estimated by a method of Fourier extrapolation and are designed to estimate what the final centered moving average value will be.

The problem with Hurst Channels is that the channels must stop before the current time and consequently we are left in a state of limbo.

It is true that there is no technique that will guarantee that when we try to extend the computation to current (and future) time that the resultant values will be correct.  However, the Fourier analysis of the data not only tells us something about the amplitude of cyclic components of the data but also the phase value at the end of the computation.  These fourier coefficients can be used to extrapolate data into the future and in many cases offer very satisfactory solutions to the lack of data problem.

 

ES_daily_50proj_LBOC.gif (32030 bytes)

 

The following is a listing of the input parameters to this routine:



Input: Price(MedianPrice of Data1);

Input: 
Length1(75), {Longest length for Hurst channel }
Ampl1Mul(1), {Multiplier to adjust projected amplitude}

Length2(21), {Next Long length for hurst channel }
Ampl2Mul(1), {Multiplier to adjust projected amplitude} 

Length3(11), {Middle length for hurst channel }
Ampl3Mul(1), {Multiplier to adjust projected amplitude}

Length4(11), {Longer Short length for hurst channel }
Ampl4Mul(1), {Multiplier to adjust projected amplitude}

Length5(11), {Shortest length for hurst channel }
Ampl5Mul(1), {Multiplier to adjust projected amplitude}

ChanWid1(3), {Multiplier for ATR(Length1) if positive }
             {Percent of Current price if Negative }

ChanWid3(2), {Multiplier for ATR(Length2) if positive }
             {Percent of Current price if Negative }

ChanWid5(1), {Multiplier for ATR(Length2) if positive }
             {Percent of Current price if Negative }
ColrTop1(cyan), {Color of top line of #1 channel. }
             {To allow plotting 3 channels, the first }
             {and Third channels are plotted using the}
             {same plot (1&2) definitions but plotted }
             {on alternate bars. Channel 3 uses the }
             {definitions in the first two plots as }
             {defined in the FORMAT INDICATOR > STYLE }
             {window. So to define the color and the }
             {line width for the longest channel, this}
             {and ColrBot1 (color of bottom line ch1.}
             {mult be assigned a color value which can}
             {be an explicit color or a number between}
             {1 and 16 as desired. }
ColrBot1(yellow),{Color of bottom line of #1 channel }
SizeLin1(3), {Size of lines for #1 channel 0-6 }

BarsProj(15), {Number bars past LBOC to project channel}
ProjColr(red), {Color for projection zone. 0=autoselect}

PlotAll(true); {True= plot values for waves 1, 3, and 5 }
             {of the possible 5 plots to display }
             {1=longest, 3=middle, 5=shortest }
             {False= plot values only for composite of}
             {all five waves. }


Input: StopDate(0), {Set to examine behavior relative to real}
                    {data -- will project as required }
                    {Format: yyymmdd yyy=4 digit yr-1900 }
StopTime(0); {Set 24 hour time (no colon) hhmm }




The following are some more pictures of the routine in action.

$SPX_30min_50barProj_LBOC_fiddle.gif (40058 bytes)

 

CY8822_daily_Theoretical.gif (36953 bytes)

 

QQQ_5min_121-31-Proj51_0529.gif (29894 bytes)

 

The following pictures show a different use of the Hurst/Fourier indicator to determine overbought/oversold zones in an equity.   In this case we examine weekly price/time series of Coke (KO) for this study.   Hurst has a considerable discussion about using 1/2 and full span averages for purpose of determining low risk entry/exit zones.  My experience indicates that 1/2 span of the shortest detected cycle of significant magnitude and 2 times span of a much longer period cycle works better than just the full span of the shorter period.

KO_wk_Hurst-OBOS_Freq.gif (34064 bytes)

KO_wk_Hurst-OBOS_Stoch.gif (30709 bytes)

 



{Hurst_Channel_Four_5 Indicator 

Author: Clyde Lee, www.theswingmachine.com
clydelee@clydelee.com

Purpose: This indicator will plot one, two, or three different channels 
based on the centered moving average of the median price of each 
day plus and minus a selected multiple of the N day average true 
range or a percentage of price. (Note: You may use close/high/low
or whatever, I have found median price great.) Up to 5 lengths
can be included for projection but only 3 can be plotted.

Whenever the ability to compute a centered moving average does
not exist (e.g. current bar closer to last bar on chart than 1/2 
the length to CMA length) then an estimate is made by using a 
Fourier decomposition of the prior moving average data plus the
current price data from that point to the last bar on the chart.

We assume that each of the periods selected represents the FULL 
PERIOD that we wish to extract (see Hurst on this matter) so
we take 1/2 times each specified period to compute a Centered
Moving Average and use the full period to do a computation of the
Fourier Amplitude and Phase of the CMA data for two full cycles 
before the end (last bar on chart or last time period) minus the
stepback of 1/2 cycle. We then use this pair of coefficients to 
calculate the probable value of the CMA values after the limit of 
centered moving average.

Using the calculated coefficients, data is extrapolated past the 
end of where valid CMA values are calculated and can be extended
past the data plotted.

The extrapolated values are plotted in black or white depending
on the background of the chart. The REAL centered moving 
average values will be colored as you select them in the style
of the indicator. An option exists to stop the plot of the
longer period data at the point of the last CMA computation
and just show the composite of the long, medium, and short
cycle data in the projection.

A routine called FOURLINE is provided which will help to 
establish what are the real cyclic elements of a symbol at the
present time. Remember, use 1/2 the indicated periods for best
results in evaluating where major turning points occurred and
what can best be used for future extrapolation.

NOTE: Because TradeStation goes goofy if you try to plot past
the last point defined by "max bars to right" this becomes a
limiting factor EVEN WHEN YOU STOP THE COMPUTATION MUCH EARLIER.

Also, since TS only allows 4 plot statements, the two Major 
channels have been programmed to use the PlotX statements but
on an alternate bar basis so it is necessary to make sure that
these first two plot definitions call for POINT rather than
line defintion.

It appears to me that two sets of proper length are all that is
required but since Hurst mentioned 3 sets (even though he erased
the first after it was established) I have provided that facility.

It is often true that the Fourier coefficient estimates taken over
the last 2 cycles of data in the symbol being evaluated are not
necessarily the true amplitude values and consequently I have
provided a "fiddle" factor (amplitude multiplier) for each of the
3 channels which may help in establishing more realistic values
for a projection. For some reason unknown to me, it appears
that sometimes the current phase of the cycle has flipped a
full 180 degrees from the calculated value. You can see this
when at the beginning of a projection the projection suddenly
goes in the opposite direction of the average. When this
happens, just assign a value of -1 to the multiplier of the
offending cycle and see what happens.


Reference: the Profit Magic of STOCK TRANSACTION TIMING
J.M. Hurst, Prentice-Hall, Inc. 1970 ....

Copyrighted 4/2001-2002 by Clyde Lee. This is a copyrighted application
and is intended for the use of the person paying a fee for the right
to use ONLY. Normal copyright privileges and penalties if violated are
in effect.


}
Input: Price(MedianPrice of Data1);

Input: 
Length1(75), {Longest length for Hurst channel }
Ampl1Mul(1), {Multiplier to adjust projected amplitude}

Length2(21), {Next Long length for hurst channel }
Ampl2Mul(1), {Multiplier to adjust projected amplitude} 

Length3(11), {Middle length for hurst channel }
Ampl3Mul(1), {Multiplier to adjust projected amplitude}

Length4(11), {Longer Short length for hurst channel }
Ampl4Mul(1), {Multiplier to adjust projected amplitude}

Length5(11), {Shortest length for hurst channel }
Ampl5Mul(1), {Multiplier to adjust projected amplitude}

ChanWid1(3), {Multiplier for ATR(Length1) if positive }
             {Percent of Current price if Negative }

ChanWid3(2), {Multiplier for ATR(Length2) if positive }
             {Percent of Current price if Negative }

ChanWid5(1), {Multiplier for ATR(Length2) if positive }
             {Percent of Current price if Negative }
ColrTop1(cyan), {Color of top line of #1 channel. }
             {To allow plotting 3 channels, the first }
             {and Third channels are plotted using the}
             {same plot (1&2) definitions but plotted }
             {on alternate bars. Channel 3 uses the }
             {definitions in the first two plots as }
             {defined in the FORMAT INDICATOR > STYLE }
             {window. So to define the color and the }
             {line width for the longest channel, this}
             {and ColrBot1 (color of bottom line ch1.}
             {mult be assigned a color value which can}
             {be an explicit color or a number between}
             {1 and 16 as desired. }
ColrBot1(yellow),{Color of bottom line of #1 channel }
SizeLin1(3), {Size of lines for #1 channel 0-6 }

BarsProj(15), {Number bars past LBOC to project channel}
ProjColr(red), {Color for projection zone. 0=autoselect}

PlotAll(true); {True= plot values for waves 1, 3, and 5 }
             {of the possible 5 plots to display }
             {1=longest, 3=middle, 5=shortest }
             {False= plot values only for composite of}
             {all five waves. }


Input: StopDate(0), {Set to examine behavior relative to real}
                    {data -- will project as required }
                    {Format: yyymmdd yyy=4 digit yr-1900 }
StopTime(0); {Set 24 hour time (no colon) hhmm }




Reference: the Profit Magic of STOCK TRANSACTION TIMING
J.M. Hurst, Prentice-Hall, Inc. 1970 ....

Copyrighted 4/2001-2002 by Clyde Lee. This is a copyrighted application
and is intended for the use of the person paying a fee for the right
to use ONLY. Normal copyright privileges and penalties if violated are
in effect.

 

To order from PayPAL
CLICK  THE    Add to Cart   Button TO ORDER the
Hurst_Channel_Four @ $500.00
   


To order from  CCNow
Product ID: HCF-002    Cost: $500.00

Product Name: Hurst_Channel_Four
Click on: 
Buy It  to Order

 


 

Return to Home Page

For additional information either
email:   clydelee@clydelee.com
Phone: clyde lee, SYTECH Corporation, (713) 783-9540