Price Action v Indicators

In my backtests CCI and Stoch produce best results. I tested EURUSD for the period June 6 to November 16 with 60 min charts using Tradestation:

indicator #trades $ profit curve

CCI 223 22'340 steady up
Stoch 245 21'856 steady up
MACDhisto 203 10'766 steady up
MACD top/bot 200 15'445 unsteady
EMA 211 13'729 unsteady
MACDcross 97 6'723 steady
2 EMA cross 47 5'527 unsteady

These figures are of course not representative but a good indication for the value of the CCI and the Stoch.

Happy trading, Felix
 
Quote from ozzietrader:

Hi TL Trader,
What do you suggest for forex traders where volume is basically rendered useless?
More...

Seems to me you got two choices. FX futures , which for me anyway the volume doesn't seem to be as helpful/useful as with other markets. Or don't use volume at all. Since no one else really has it either theoretically it may not be as big a disadvantage as with other instruments.
 
Quote from mgabriel01:


Im trying to get my arms around the reasoning behind using price action v technical indicators to find entry/exit points

[/B]
More...


>mgabriel01


How's this?

Price movement = the heart.
Indicators = the pulse.
A doctor can tell more by examining the heart than by feeling the pulse.

Having learned to analyze price movement from the start, I
could now never imagine using indicators, as I would then indirectly be analizing what I am going to trade.
Sort of like going through an interpreter rather than speaking
the language yourself. Too much can be lost in the translation.
 
While browsing around, studying CCI, I ran across this statement

"A close approximation of the CCI can be replicated easily by the use of Bollinger Bands, which will eliminate the need for any programming as these bands are included with most software packages and can be plotted right over the price bars."

at http://tuckerreport.com/indicators/cci-basic/ .

The accompaning article makes it plain that the mysticism around Woodies' CCI can be analyzed as simple moving average arithmetic.

In line with other thoughts on Bollinger...
 
As someone who did a lot of work on trying to improve CCI I can tell you that its an approximation and not the real thing especially when it comes to divergence.

A 14 sma and price is a good approximation to the basic cci (just line on close) too.

But ... the CCI has a normalizing factor that takes into account the range of the last XX bars. And its not the same as a standard deviation on which bollinger bands are based. So your divergences will be different - which is right I do not know :(

You could make it exact by using the CCI calc'd normalizing factor to build your 100 and 200 Bollinger bands.

Interesting idea. I'm going to do that for Sierra Chart (CCI bands).
 
After I learned to use MACD in different time spans to confirm direction, I noticed stochastics seemed faster and more leading.

Then I started studying CCI theory. The follow up of the article I noted above is http://tuckerreport.com/indicators/cci-improvement/

My trading improved with CCI, and two std deviation (1.5 and 2.5) Bollingers have not yet proven themselves. Its probably me, not the concept.

But at the same time I am ripping Time&Sales data apart in Excel, and about to plot CCI against the raw data.

The reason is that constant volume bars ignore the fact that most significant price moves appear to be initiated when the supply or demand dries up, and volume falls off. Constant volume hides this artifact, pure price action, the "tape," does not.
 
Quote from kiwi_trader:


But ... the CCI has a normalizing factor that takes into account the range of the last XX bars. And its not the same as a standard deviation on which bollinger bands are based. So your divergences will be different - which is right I do not know :((
More...

Tucker goes on to suggest an adaptive CCI, which, if I understand him, is an attempt to vary the length of the sample period to match variations in the fundamental cycle length. He noted that Lambert, the developer of CCI, suggested a sample period equal to 1/3 of a cycle.

Even without FFT, it is easy to see that the cycle length varies with the mood and angst of the market. He discusses something similar.

The divergence looks to vary with the sample period, which if "incorrect" suggests an erroneous dP/dV, or a misleading >rate< of divergence.
 
I read through Tucker's work and also looked into adaptive CCI. The thing is that the smoothing is only a good thing if trader has wide stop losses. However, this brings a usual problem to forefront, if several of those losses happen in the row, how many traders will be able to pull trigger again.

Tucker did say that everything needs to be confirmed and validated by each person, and I totally agree with it. Some might prefer adaptive smothed CCI, but I am totally Ok with using price action along with regular CCI 14 on constant volume bars.

Regards,
redduke
 
Quote from kiwi_trader:

As someone who did a lot of work on trying to improve CCI I can tell you that its an approximation and not the real thing especially when it comes to divergence.

A 14 sma and price is a good approximation to the basic cci (just line on close) too.

But ... the CCI has a normalizing factor that takes into account the range of the last XX bars. And its not the same as a standard deviation on which bollinger bands are based. So your divergences will be different - which is right I do not know :(

You could make it exact by using the CCI calc'd normalizing factor to build your 100 and 200 Bollinger bands.

Interesting idea. I'm going to do that for Sierra Chart (CCI bands).
More...

Why restrict it to the CCI?


Goinglite
 
You must log in or register to reply here.
Back
Top