mktmole
325 posts
msg #129581
- Ignore mktmole |
7/4/2016 10:11:47 AM
Any efforts to code the below from anyone in the SF Community would be much appreciated.
tks, mm
- - - -
From Stocks and Commodities Magazine July 2016. "The Super Passband Filter" by John Ehlers.
This PB oscillator seeks to PB out both high and low frequencies from market data to eliminate wiggles from the resultant signal thus significantly reducing lag. This PB indicator achieves this by using 2 differenced EMA's of varying periods. (40 and 60). Trigger points for the PB oscillator are added with a RMS cyclic envelope over the Signal line. Output of the PB waveform is calculated by summing its square over the last 50 bars and taking the square root of the averaged sum to create trigger levels.
Buy when PB crosses above its -RMS line. Short when PB crosses below the RMS line.
This EMA calculation is significantly more reactive than the standard EMA or Wilder's EMA.
Standard EMA formula found in most software is:
α = 2/(N+1) standard smoothing factor where "N" = the period.
Using N=40: α = 2/(N+1) = 0.04878, and (1 - α) = 0.95121
Ehler's Super PB Filter formula uses a smoothing factor of 5/N "period".
α = 5/N
Using N =40: α =5/N = 0.12500, and (1 - α) = 0.87500
|
mktmole
325 posts
msg #129667
- Ignore mktmole |
7/8/2016 9:23:29 AM
anyone? any thoughts at all?
|
jimvin
173 posts
msg #129750
- Ignore jimvin |
7/12/2016 12:10:26 AM
Dude...the record shows I have no problems making a public spectacle of myself in the interests of a good cause (which this certainly is) but this is so far beyond me that it would be useless to stick my head above the trench. (Multi-dimensional physics is the outside of my envelope). I wish you and any takers the best with this challenge.
|
graftonian
1,089 posts
msg #129754
- Ignore graftonian |
7/12/2016 8:59:36 AM
I have been attempting to make a little progress with the math on excel, but cannot make the formula work. Too many parentheses, or not enough. If anyone could rewrite this, maybe t'would work.
PB = (a1 - a2) * Close + (a2*(1 - a1) - a1 * (1 - a2))
* Close[1] + ((1 - a1) + (1 - a2))*PB[1] - (1 - a1)
* (1 - a2)*PB[2]
|
Kevin_in_GA
4,599 posts
msg #129763
- Ignore Kevin_in_GA |
7/12/2016 10:38:01 AM
This seems to be another attempt to make a "zero lag" MA. Still not sure this would result in any better result than a standard short term EMA. This would not be anything I would use in my trading strategies anyway, but good luck.
|
mktmole
325 posts
msg #129852
- Ignore mktmole |
7/15/2016 5:15:37 PM
thank you )
|