› Forums › Development › Indicators – Development › Footprints of a TZR indi development
 This topic has 14 replies, 5 voices, and was last updated 9 years, 6 months ago by simplex.

AuthorPosts

December 22, 2014 at 1:35 am #2105
Sometimes when ideas are overwhelming one, it can at worst so to speak suffocate oneself or end up in a big confusion, which will bring dev work to a halt. Best thing would be to… yes, to write things down, and keep the project in check. This topic will hopefully help me to do that and mark my highlights in this TRZ project.
A quarter of a step in trying to compound or accumulate the probabilities is somewhat done. I see something, lets see…
Important notes: Euro session only, picked max probabilities which in current “mathematical form” top out around 70, lower TRZ only.
To do: check upper value calc, show on chart, similar behaviour?; implement middle zones; V2 for the “other” distribution.
Significance of a visual chart analysis?
December 22, 2014 at 10:30 pm #2180Sorry I cant see the images you have posted. Interested in what you are saying..
What's it all about? It's all about money.
December 23, 2014 at 12:48 am #2183Sorry I cant see the images you have posted. Interested in what you are saying..
Hmm, right click on the image, open in new tab/window, then it’s visible.
EDIT: the links are not showing… plain copy/paste then
http://postimg.org/image/mprfyovtt/
http://postimg.org/image/nwxha9mqj/ This reply was modified 9 years, 6 months ago by Rahat Lukum.
December 23, 2014 at 12:59 am #2186Yep that worked, thanks
h
December 25, 2014 at 12:45 pm #2396I took time off for a couple of days as something in the back of my head said that something is not rightly interpreted. And it turned out to be true, I wanted to compound or accumulate probabilities for every single bar but I didn’t! A slight logical error in my thinking. I’ve mended it and now I see probabilities for every bar, screenshot below for illustration.
http://postimg.org/image/4qck76r1n/
I’m still pursuing a way to compound or accumulate probabilities, but clearly the current method to compound them is too fuzzy. Time to delve into theory.
Merry Christmas and a prosperous New Year!
December 28, 2014 at 7:56 pm #2637… accumulate the probabilities …
Hmmm – maybe I can contribute. Which probabilities (precisely) do you plan to accumulate?
A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top and bottomfishing to people on an ego trip. (Dr. Alexander Elder)
December 28, 2014 at 10:25 pm #2643Firstly, I’m not at home with English statistical terms, so pardon my loose use of those and me sounding a bit daft.
Best if I give one example: I have a bar holding multiple probability values, something like the following:
Bar[0,0]=35 Bar[0,1]=5 Bar[0,2]=80
What I want is to have a combined/accumulated/compounded reading for Bar[0]. The answer being 40 – I’m not so sure.
Suggestions very much appreciated!
… accumulate the probabilities …
Hmmm – maybe I can contribute. Which probabilities (precisely) do you plan to accumulate?
December 29, 2014 at 1:34 pm #2672Firstly, I’m not at home with English statistical terms, so pardon my loose use of those
The same is true for me.
Best if I give one example: I have a bar holding multiple probability values, something like the following:
Bar[0,0]=35 Bar[0,1]=5 Bar[0,2]=80
What I want is to have a combined/accumulated/compounded reading for Bar[0]. The answer being 40 – I’m not so sure. Suggestions very much appreciated!
My interpretation of those lines: At bar zero you’re observing three events with probabilities p0 = 0.35, p1 = 0.05, and p2 = 0.8. Correct so far?
Key questions now: What kind of events are the basis for those probabilities? Are those 3 events statistically independent of each other (see: https://en.wikipedia.org/wiki/Independence_%28probability_theory%29) ?
A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top and bottomfishing to people on an ego trip. (Dr. Alexander Elder)
December 31, 2014 at 12:30 pm #2850Thank you for your answer!
I looked into it before too, then other stuff took over, anxiously waiting to return to it.
You’re correctly understanding the probabilities. As far as my research went (I have some things to go through still), bar probabilities can be viewed both ways – as independent and dependent. In some cases there is clear overlap – dependent – and in other cases there’s like a “semioverlap” at most, which I would define as an independent event.
For independent cases I should multiply the probabilities, correct?
0.35*0.05*0.8=0.014
Going Saver’s way (10.014=0.986) would then produce one helluva probability stat …
What are the options for dependent probability accumulation?
Firstly, I’m not at home with English statistical terms, so pardon my loose use of those
The same is true for me.
Best if I give one example: I have a bar holding multiple probability values, something like the following:
Bar[0,0]=35 Bar[0,1]=5 Bar[0,2]=80
What I want is to have a combined/accumulated/compounded reading for Bar[0]. The answer being 40 – I’m not so sure. Suggestions very much appreciated!
My interpretation of those lines: At bar zero you’re observing three events with probabilities p0 = 0.35, p1 = 0.05, and p2 = 0.8. Correct so far?
Key questions now: What kind of events are the basis for those probabilities? Are those 3 events statistically independent of each other (see: https://en.wikipedia.org/wiki/Independence_%28probability_theory%29) ?
December 31, 2014 at 1:58 pm #2853For independent cases I should multiply the probabilities, correct? 0.35*0.05*0.8=0.014 Going Saver’s way (10.014=0.986) would then produce one helluva probability stat
Yes, that’s correct!
A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top and bottomfishing to people on an ego trip. (Dr. Alexander Elder)
December 31, 2014 at 2:08 pm #2854As far as my research went (I have some things to go through still), bar probabilities can be viewed both ways – as independent and dependent. In some cases there is clear overlap – dependent – and in other cases there’s like a “semioverlap” at most, which I would define as an independent event.
IMO, absolutely NO! You’re having 3 independent events if you’re drawing lottery tickets out of 3 buckets independently. If winning probabilities of those single events are P0, P1, and P2, then the compound probability for all drawings to be successful is P = P0 * P1 * P2.
But I definitely doubt that 3 events taking place on the very same bar of one chart can really be completely independent of each other, if they’re suitable for the purpose of trading signals. Even such different events like rise of volume and a sudden price move will have some nonzero ‘coupling force’, which leads to dependency.
A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top and bottomfishing to people on an ego trip. (Dr. Alexander Elder)
December 31, 2014 at 2:09 pm #2855What are the options for dependent probability accumulation?
Simple testing.
A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top and bottomfishing to people on an ego trip. (Dr. Alexander Elder)
January 10, 2015 at 12:40 am #3645Just pulled everything together for the final push for GEN1 indicator(s). 1 down 7 to go. Stubborn as I am, I’m still going the accumulation route, combined up, down, middle readings are nothing fancy, useless actually.
http://postimg.org/image/ficdnobuf/
 This reply was modified 9 years, 6 months ago by Rahat Lukum. Reason: Why this forum has image feature if it's not working?
 This reply was modified 9 years, 6 months ago by Rahat Lukum. Reason: Is that an image there?
January 10, 2015 at 7:46 am #3655Hello Rahat Lukum, here is a link regarding computing the probability of independent events with examples and pictures, and here is a link for computing the probability of dependent events using playing cards as an example.
If you look at the market prices, you will find high autocorrelation which in simple terms indicates that future prices are highly dependent on previous prices. This makes sense because a price doesn’t go from 1.5 to 0.2 very often! If price moves from 1.3 to 1.5, it will be more likely to trade near 1.5 for some time rather than to trade near 1.3.
If you look at market returns or price differences, you will find that they are not very correlated. This indicates that future returns are fairly independent of current returns. The price can go up or down on the next bar, and it is anyone’s guess as to which will occur. A return or price difference can be simply computed by taking the difference between two close prices, for example, or the close price minus the open price on a given bar.
January 10, 2015 at 1:12 pm #3664here is a link regarding computing the probability of independent events with examples and pictures, and here is a link for computing the probability of dependent events using playing cards as an example.
Great links, FXEZ! These remind me of my former reply in this thread:
Simple testing.
I should have said more precisely that IMO for dependent events in our TZ methodology the functions describing dependencies between two particular events are basically unknown. That’s why I opted for the empirical approach (Simple testing!) to determine some compound probability.
Thanks for clarifying!
s.
A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top and bottomfishing to people on an ego trip. (Dr. Alexander Elder)

AuthorPosts
 You must be logged in to reply to this topic.