Hi All,

Having looked at the thread on Hull MA, I thought this might be of interest to others.

When I got interested in technical analysis, I was very keen on using indicators, etc, like most newbies, and came across the following formula for significantly reducing the lag inherent in moving averages.

That was a while ago, and am now more interested in reading price and volume, but thought others might be able to make better use of it than I have.

The basic formula is:
&EMAOne := MovingAvgMethod (Close , Periods , 2)
&EMATwo := MovingAvgMethod (&EMAOne , Periods , 2)
&Diff := &EMAOne - &EMATwo
&EMAOne + &Diff

Replace "MovingAvgMethod" with whatever MA you prefer.

And here's a pic comparing several different MAs with/out Zero Lag, including Hull:


Regards,

[Minetoo]

phase21, please post the eld's (or code) for your zero lag Hull and EMAs.

[phase21]

Hello Minetoo,

The code I posted above is a direct copy from a function in Trade Navigator. I use TN for trading and quick prototyping of ideas. You can define functions in TN, but it doesn't have a programming language to speak of.

However, you should be able to plonk this code/algorithm into a template you use for defining functions in your favourite software, hence no "*.eld".

But, below is some code which does the same thing in AmiBroker. AB has a "C"-like programming language.

Define the function:
/*======================== ========================= ========================= ===
Zero-Lag Moving Averages

This AFL contains functions to calc moving averages with minimal lag.
========================= ========================= ========================= ==*/

//-----------------------------------------------------------------------------
function ZeroLagMovingAvg(ValToSmo oth, BarsToUse)
/*
This AFL smooths the input data, and then removes as much lag as possible.

Note: Original source of the formula not known.
-----------------------------------------------------------------------------*/
{
// Declare local variables
local result ;
local MAOne ;
local MATwo ;
local MADiff ;

// Initialise the local variables.
result = Null ;

// Your code goes in here
MAOne = MA(ValToSmooth , BarsToUse) ;
MATwo = MA(MAOne , BarsToUse) ;
MADiff = MAOne - MATwo ;
result = MAOne + MADiff ;

// Declare which variable/value to be returned to the user/calling function
return result ;
}


Call the function:
/*======================== ========================= ========================= ===
Zero-Lag Moving Averages

This AFL plotts the moving averages with minimal lag.
========================= ========================= ========================= ==*/

#pragma nocache ;
#include_once <ZeroLagMovingAvg.afl> ;

ValToSmooth = ParamField("Price field", 3) ;
BarsToUse = Param("Bars to use", 13, 2, 1000, 1, 1) ;
SmoothedVal = ZeroLagMovingAvg(ValToSmo oth, BarsToUse) ;

Plot(SmoothedVal, "ZeroLagMovingAvg", colorBlue, styleLine | styleDots | styleThick, 0, 0, 0) ;


Note: I tend to define things as componentised as possible, a-la building blocks, and draw them all together as late as possible/appropriate. That way I can separate the visuals from the calcs.


Regards,