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)
Other Statistical Tendencies
Moderator: moderators
Please add www.kreslik.com to your ad blocker white list.
Thank you for your support.
Thank you for your support.
- michal.kreslik
- rank: 1000+ posts
- Posts: 1048
- Joined: Sat May 13, 2006 2:40 am
- Reputation: 36
- Location: Monte Carlo, Monaco
- Real name: Michal Kreslik
- Gender:
Re: Other Statistical Tendencies
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
- Attachments
-
- Fibonacci_TS_Screenshot.png (173.13 KiB) Viewed 7186 times
- 2.klever
- rank: <50 posts
- Posts: 16
- Joined: Sun Apr 01, 2007 12:56 am
- Reputation: 0
- Location: Edmonton
- Real name: Kirby
- Gender:
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.
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.
Please add www.kreslik.com to your ad blocker white list.
Thank you for your support.
Thank you for your support.
- michal.kreslik
- rank: 1000+ posts
- Posts: 1048
- Joined: Sat May 13, 2006 2:40 am
- Reputation: 36
- Location: Monte Carlo, Monaco
- Real name: Michal Kreslik
- Gender:
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:
Ahh, I'm so happy QueasyLanguage no longer controls my life
Michal
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
Michal
- michal.kreslik
- rank: 1000+ posts
- Posts: 1048
- Joined: Sat May 13, 2006 2:40 am
- Reputation: 36
- Location: Monte Carlo, Monaco
- Real name: Michal Kreslik
- Gender:
Please add www.kreslik.com to your ad blocker white list.
Thank you for your support.
Thank you for your support.