Page 1 of 1

Other Statistical Tendencies

Posted: Tue Jun 05, 2007 4:51 am
by 2.klever
A while ago I captured a quote from Michal , probably from TRO's Defunct Trend Indicator Thread. After the quote I have a few clarification questions for Michal.

"...let's have two consecutive bars (daily, hourly, 15min, whatever). We know the first bar's open price. We want to know what is the probability of this open price to be at some particular place within future (these very two candles) high - low range.

To make the point more understandable, let's pretend that the statistics shows that the first bar's open is always at the 0% of the following two bars' range. If we knew this, we could always BUY at the open since the price will always go up from the 0% (or lowest low) to 100% (highest high).

I constructed this fuzzy logic indicator based on 60 min candles but it does not really matter what time frame do you choose. If you have a look at the screenshot, it shows you the probability of one bar's open price appearance at the particular two bars' range (two bars = the opening bar and the following one). If you look closely, you could spot several continuous "lines". These are in fact only points which lie on the same position range prediction value level. This means that, for example, it is quite probable that today's open will lie at the exact 50% (the 50% "line") of the today's and tomorrow's total range.

If you look very closely, you will see a startling thing: [highlight=red]the probability of today's open to lie at some particular percentage within the today's and tomorrow's total high/low range increases drastically with the coincidence of this particular percentage with the corresponding fibonacci level.[/highlight]

The data is calculated based on 60 min candles over about 100 consecutive days.

This simple indicator also proves TheRumpledOne's theory on which he bases his "milk the cow" strategy.

Michal,
European Union""

Questions:
1. I understand the statement that todays open will most probably lie at the 50% lin of Today's and tomorrow's total range but I don't understand the highlighted section about the fibonacci levels,,, probably because I don't have the screenshot you referred to or the indicator your were presenting. Could you please explain in a little more detail the fibs part. Don't know if you still have the indicator or not..


Thanks,
2.klever (Kirby)

Re: Other Statistical Tendencies

Posted: Tue Jun 05, 2007 7:24 am
by michal.kreslik
2.klever wrote:I don't understand the highlighted section about the fibonacci levels,,, probably because I don't have the screenshot you referred to


Hi, 2.klever,

I wrote the indicator in 2005 for Tradestation 8.1. A picture is worth thousand words, so here is the screenshot:



The clear visual distinction of the Fibonacci levels is quite startling, isn't it.

Attached to this post you will find another screenshot of the same indicator applied to the chart in the Tradestation environment.

Michal

Posted: Tue Jun 05, 2007 2:02 pm
by pr
I think this is an interesting study, but the formulas are cut off in the picture. Would it be possible to post the full code, or at least the full formulas for PastRange and ExtremeLine?

Posted: Tue Jun 05, 2007 3:20 pm
by 2.klever
PR,

I have tried it and I think the formulas are like this:

vars:
length(3), daylength(96),Boundary(20),PastRange(5),Extremeline(0);

PastRange = highest(high,length) - lowest(low,length);

Extremeline = 100*(open[Length-1]-lowest(low,length-1))/PastRange;



plot1[length-1](Extremeline, "extremeline");
plot2[length-1](Extremeline[daylength]);
plot3[length-1](Extremeline[daylength*2]);
plot4[length-1](Extremeline[daylength*3]);
plot5[length-1](Extremeline[daylength*4]);
plot6[length-1](Extremeline[daylength*5]);
plot7[length-1](Extremeline[daylength*6]);
plot8[length-1](Extremeline[daylength*7]);
plot9[length-1](Extremeline[daylength*8]);
....

...

you only need about 10 plots to see the image well.. set the plot format to left or rigth tick for the plots.

Posted: Tue Jun 05, 2007 3:36 pm
by 2.klever
This concept visualizes and supports some of Gann's theorys, Murray Method Lines as well as the Middle Crossover phenomena.

Posted: Tue Jun 05, 2007 3:56 pm
by michal.kreslik
I'm not using Tradestation any more, so I don't have access to the EL indicator's source code in a text form, but 2.klever's suggestions is correct. Anyway, I'm not sure what "Boundary" meant :)

I can rewrite this simple piece of code for NeoTicker. It's not even going to be so gawky as it is in EasyLanguage. Just look at that idiotic requirement in EasyLanguage to use an individual plot statement for every single plot. In C#, I would just cycle through all of the plot statements in an elegant loop like this:

Code: Select all

for (int plotIdx = 0; plotIdx < plotValues.Length; plotIdx++)
            {
                N.ItSelf().set_Plot(plotIdx, plotValues[plotIdx]);
            }


Ahh, I'm so happy QueasyLanguage no longer controls my life :D

Michal

Posted: Wed Jun 06, 2007 1:57 am
by 2.klever
Can anyone think of a similar way to illustrate and visualize the S/R phenomena of the BuyZone, Middle crossover, or other Support Resistance lines using a graphical method like the one shown above?

Posted: Sat Oct 20, 2007 12:10 am
by balkanbro
hmmm I dont understand the =pictures of the charts ?? the the bottom half of the chart does not show anything.

Posted: Sat Oct 20, 2007 4:20 am
by michal.kreslik
balkanbro wrote:hmmm I dont understand the =pictures of the charts ?? the the bottom half of the chart does not show anything.


Look more closely.

Posted: Sat Oct 20, 2007 9:34 am
by balkanbro
well I am still confused ?? I see the red lines on the chart but still do not understand was is ploted under the chart.

thanks