Thread: Jurik
View Single Post
  #127 (permalink)  
Old 04-11-2007, 07:31 PM
manz66 manz66 is offline
Member
 
Join Date: Oct 2005
Posts: 32
manz66 is on a distinguished road
Divergence in traditional indicators gives many false signals in day trading

You are absolutely right! Thanks

Quote:
Originally Posted by Pecunia non olet
Some fun with CFB.

....Specifically CFB Adaptive Polarized Fractal Efficiency of Momentum.

https://www.tradestation.com/Discuss...Topic_ID=16096

- - - -

A screenshot from the page above.

https://www.tradestation.com/Discuss...20Momentum.jpg

The screenshot shows stiffer settings where some lag shows up, but this stuff can be tuned a lot better than the image shows.

::--> Note the lack of so-called 'divergence' - that is, unlike MACD and other standard TA oscillators, these have less tendency to move back down toward the center or zeroline while price is still moving up.

IMHO, a big problem with traditional oscillators is that they give false directional readings when a trend occurs, but the momentum of the trend stops increasing or falls off a bit. Price continues to rise, but the traditional oscillators like MACD start to fall, giving a false signal.

This is less of a problem with the indicators in the screenshot. Note the trend on the right side.

"Less false 'divergence' signals, priceless"

A lot of cool stuff can be done with tools like these.

- - - -

This is hard to articulate this late at nite but-

One nonobvious aspect of the problem where price continues to rise, but the traditional oscillators like MACD start to fall and give a false signal is that it sort of forces the use of zeroline crossings of the oscillator as a signal. But maybe, just maybe, the slope-or-state of the indicator is a better signal than a zeroline cross. By slope-or-state I mean, "If the indicator is sloping upwards or is at 100, LongCondition = true". I have seen many instances where slope-or-state of "non-divergence" indicators like PFE of Momentum outperformed zeroline crossings of diverging indicators.... An easier way to say it: In my experience, slope-or-state of "non-divergence" indicators outperforms either slope or zeroline crossings of traditional diverging indicators. Your milage may vary.
Reply With Quote