|
THOMAS HELGET
 Elite
  Posts: 610
Joined: 3/22/2006
Location: BALDWINSVILLE, NEW YORK
User Profile |
Fellow OT Professional Users:
In the April, 2019 Technical Analysis of Stocks & Commodities Vitali Apirine offers the Adaptive Exponential Moving Average.
This is not to be confused with Nirvana Systems' own Moving Average (Adaptive).
I have coded it up and attached it here. Please note that the values it generates will not agree with his Excel spreadsheet on page 12 as it uses many more periods in its calculation. Also it does not exactly agree with the screenshot below as I used the IWM instead of the Russell 2000 Small Cap Index he shows in Figure 2 on page 14:

Tom Helget
[Edited by THOMAS HELGET on 3/21/2019 6:30 PM]
Attached file : Figure 2 Using IWM.jpg (114KB - 709 downloads)
Attached file : indAdaptiveEMA.txt (1KB - 245 downloads)
|
|
LSJ
 Veteran
    Posts: 190
Joined: 7/23/2012
Location: Citrus Springs, FL
User Profile |
Tom, thanks for providing another tool. I have plotted the EMA next to the inadaptiveEMA and see they are very close. In what situation would you see one more effective than the other? Does Mr. Apirine make any suggestions?

[Edited by LSJ on 3/22/2019 6:52 AM]
Attached file : MA.jpg (70KB - 642 downloads)
|
|
THOMAS HELGET
 Elite
  Posts: 610
Joined: 3/22/2006
Location: BALDWINSVILLE, NEW YORK
User Profile |
LSJ:
First Mr. Apirine would have compared apples to apples. So, in your screen shot he would have used, say, indAdaptiveEMA(30,30) - the default parameters - and EMA(30) for the comparison.
As regards double crossovers of the moving average he says on p. 12, "The adaptive moving averages generate crossover signals before the EMA does."
On page 15 he says, "You can use a combination of AEMA and EMA with the same lengths for trend identification. AEMAs with different lengths are better suited for double crossovers. Keep in mind that crossovers are prone to whipsaws. Use AEMA together with other aspects of technical analysis to get more confirmation."
And that is all I have for you!
Hope that helps,
Tom Helget
|
|
LSJ
 Veteran
    Posts: 190
Joined: 7/23/2012
Location: Citrus Springs, FL
User Profile |
Thanks for that. I am always on the look out for new ways to do things and no question these are a different period. What I wonder is though in what way would the ema shown not be equivalent to the indEMA.
I am just questioning how could I use to advantage over it's equivalent.
|
|
Jim Dean
 Sage
       Posts: 3433
Joined: 3/13/2006
Location: L'ville, GA
User Profile |
Fwiw, imho ...
If you're looking for a useful alternative to EMA (or WMA) that has definable benefits and is relatively easy to calc, I'd suggest consider using HullEma or HullWma. Easy to google, including formulas.
Benefits: hugs price closely while retaining smoothness ... better ID of slopes
Usage: requires higher input periods for comparable "spacing" ... try ~2x
Consider: crossovers occur earlier, but sometimes create "noise"
FWIW, I'm building in Hull MA options to all my new TradeTight tools
|
|
THOMAS HELGET
 Elite
  Posts: 610
Joined: 3/22/2006
Location: BALDWINSVILLE, NEW YORK
User Profile |
LSJ:
I am not quite sure what you are asking in your query, but as regards Jim Dean's comment the Hull Moving Average is provided by Nirvana Systems in the UIS2 (Ultimate Indicator Suite 2) Module.
In case you don't have it I have also coded it up as well and it is attached here. Because it uses weighted moving averages you should have a healthy warm-up period for it.
Tom
Attached file : HullMovingAverage.txt (0KB - 245 downloads)
|
|
LSJ
 Veteran
    Posts: 190
Joined: 7/23/2012
Location: Citrus Springs, FL
User Profile |
I have been using he Hull also and even though I don't have a set of rules for it, just examining a chart for hull crossovers using and ema looks to have potential.
|
|
Jim Dean
 Sage
       Posts: 3433
Joined: 3/13/2006
Location: L'ville, GA
User Profile |
Hi, Tom
The Hull WMA has the same warmup as a regular WMA ... being a finite series.
Otoh, the Hull EMA (same formula, but using EMA's) requires extra warmup, similar to an EMA's ... I recommend ~3x periods for ~3-signif figure accuracy
(in my TradeTight apps, I use a custom "BEMA"=bounded EMA, that actually creates an EMA with specified accuracy, than uses a finite series like WMA and SMA do ... which assures that BEMA values never change as #historical bars change).
|