Transient & Recurrent Zones

[Innate]

This reply has been reported for inappropriate content.

Thank you for the last few posts! This is finally where the conversation has been waiting to go. Brownian motion sure does read well… Looks like a new topic to study.

What's it all about? It's all about money.

[simplex]

This reply has been reported for inappropriate content.

FXEZ wrote:

If the odds of flipping a single coin are 50% then before two flips the odds of both coming up heads are indeed 0.5 x 0.5. However, after the first flip, if the first flip comes up heads, the odds for the next single flip coming up heads is still 50%, NOT 25%. The odds change back to the single flip odds for the next single flip. So if you had a coin with 98% tails and 2% heads, for any single flip the odds are 98% tails because the prior flip is independent of the next flip. So it basically comes down to how independent the two events are as to what the true probability is, so the rare-rare idea is interesting but probably not accurate. However if you really want 0.25 odds, flip both coins simultaneously. Can we do this in trading?

That’s exactly the point!

IMO, it’s essential to answer the question whether two individual events are independent of each other or not when trying to combine their individual probabilities to a joint probability.

Two events A and B are independent if and only if their joint probability equals the product of their probabilities. (see: https://en.wikipedia.org/wiki/Independence_%28probability_theory%29)

When looking at ONE single chart and its transient zones for a given h then we will easily be able to spot sequences where a top CTZ is not being followed by a bottom CTZ, but a second top CTZ instead. For typical timeframes and values of h the probability for this to happen can be found as 20 – 25 %.

If all CTZ of one single chart were independent of each other as ‘rare events’, then the probability for two top CTZ being followed by a third top CTZ should be exactly the same, 20 – 25 %. But that is not the case: this probability is significantly lower, typically around 15 – 20 %. (I will be able to provide detailed results after adding a csv output to my recurrence stats indicator).

My conclusion is, that those ‘rare events’ named CTZ are not independent of each other within one chart. There’s a coupling between those events, it may be a loose coupling, but anyway a non-zero one. Visually, price history on a chart is a representation of this coupling.

So, using stats formula that depend on the presence of statistical independence of two events to find a joint probability may lead to questionable results when examining TZ.

This basically is the reason for me to examine very cautiously any proposed strategy that depends on things like the combination of two rare events, those ‘rare-rare’ events.

I’m trying to step back a little and apply good old KISS rule to TZ – or using Saver0’s words: ‘Guys, try not to over complicate things’. I’m trying to focus on probabilities of single and combined events that I can count and calculate at the moment: stats based on single zones, and stats based on sequences of confirmed zones like BB vs BT, BBT vs. BBB, etc. (B=Bottom, T=Top).

 

A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top- and bottom-fishing to people on an ego trip. (Dr. Alexander Elder)

[Saver0]

This reply has been reported for inappropriate content.

smallcat wrote:

At image below, the price p is drawn in Yellow which located at middle of R-TZ, and the +k & -k in dodgerBlue. The lime arrow & line is how i will trade it : 1) buy order in direction of R-TZ and  2) sell order if it passed through the R-TZ.

I see what you are saying. Looks like a nice little technique. Do you have probability stats for this?

Although it might be nice and give you good results, I do not agree to it being what the equation is telling us 100%.

Let me ask you this, what is K to you?

Focus, Patience, Determination & Order in chaos

[smallcat]

This reply has been reported for inappropriate content.

Saver0 wrote:

I see what you are saying. Looks like a nice little technique. Do you have probability stats for this? Although it might be nice and give you good results, I do not agree to it being what the equation is telling us 100%. Let me ask you this, what is K to you?

Hi Saver0,

*) Have no probability yet. Is it some thing like how many % success if i get that signal ?  Ex. For last 1000 bars, i get 200 signal number #1.
Then i must calculate how many of them reach my target (ex. 180). So the probability is 180/200 X 100% ?
The same calculation for signal number #2. It is something like that ?

*) For me K is the amount of price (in pips). Still need properly calculation to get how much is the optimal k value for each TF :
a) static K value :  X pips for each TF (will be different for each TF)
b) dynamic K value:  X % of the HiLo of bar (this also will be in pips).

 

 

 

• This reply was modified 11 years, 4 months ago by smallcat.
• This reply was modified 4 years ago by wise_jedi.

[Saver0]

This reply has been reported for inappropriate content.

Zelo wrote:

Can we find out probability that price go from level A to E and E back to A, or in Forex, can we find out probability price go, e.g from 1.22900 to 1.22100 and from 1.22100 back to 1.22900

This is awesome Zelo! This is the evolution of this concept. Exactly the right direction that we need to go 

Focus, Patience, Determination & Order in chaos

[Saver0]

This reply has been reported for inappropriate content.

smallcat wrote:

Have no probability yet. Is it some thing like how many % success if i get that signal ?

Yes that is correct.

smallcat wrote:

For me K is the amount of price (in pips).

And I agree with this 

Focus, Patience, Determination & Order in chaos

[simplex]

This reply has been reported for inappropriate content.

Zelo wrote:

Can we find out probability that price go from level A to E and E back to A, or in Forex, can we find out probability price go, e.g from 1.22900 to 1.22100 and from 1.22100 back to 1.22900

Nice one! This can become very interesting if you manage to quantitatively define those states A / E as price levels or price ranges, maybe incorporating some k.

A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top- and bottom-fishing to people on an ego trip. (Dr. Alexander Elder)

[simplex]

This reply has been reported for inappropriate content.

smallcat wrote:

Is it some thing like how many % success if i get that signal ? Ex. For last 1000 bars, i get 200 signal number 1 (trading in the direction of R-TZ). Then i must calculate how many of them reach my target (ex. 180).

I think you should modify your success probability a little bit by incorporating SL. IMO, the interesting questions are:

1. How many trades hit your target BEFORE price hits your SL?
2. What is the R:R ratio for this specific trade?

Or do you want to trade without a properly set SL?

simplex

A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top- and bottom-fishing to people on an ego trip. (Dr. Alexander Elder)

[simplex]

This reply has been reported for inappropriate content.

Zelo wrote:

Pip, IMO, Let’s say our S have 100 bars, and we interest only X contain only 11 bars of S, with height of 10, from 1.20 to 1.30, We have 2 state, State 1 is 1.20 to 1.30, State 2 is 1.30 to 1.40 that we want to examine. Can we say that price in X is absorbing in state of 1.20 to 1.30??? I think FX Pricing Model is Wiener process, you was banned when you bring this subject, right? Why we should limit ourself thinking only recurrent and transient? Why not? Because they are stochastic, can we find another way to calculate the probability of price go to the next state, or next level? But we still have good RRR.

Very interesting!

If not recurrent vs. transient price zones, then what? The idea is nice, and the most basic question relating to this is:

How do we define those states Bull, Bear, and Stagnant in a quantitative way, that we can measure, count, and / or calculate?

We all know ‘classic’ TA answers like MA slopes, RSI reversals, MACD readings, etc. But the idea is to leave those behind and find substitutes for them, right?

So how would you propose to ‘measure’ those states? If we can define sound measurements, then it should be feasible to tweak the Recurrence Stats indi in order to provide stats relating to that.

A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top- and bottom-fishing to people on an ego trip. (Dr. Alexander Elder)

[smallcat]

This reply has been reported for inappropriate content.

simplex wrote:

 

I think you should modify your success probability a little bit by incorporating SL. IMO, the interesting questions are:

1. How many trades hit your target BEFORE price hits your SL?
2. What is the R:R ratio for this specific trade?

Or do you want to trade without a properly set SL? simplex

Hm… you are right my friend.

*) I am afraid of trading without SL. Although i had SL on most of my previous orders , i had blown my account several times. I can not imagine if i do not have SL (some times i put a big SL, around 200 pips. And it was touched too    )

1) This remind me on WHEN we must enter a trade, so we must know the TREND at the time we want to enter. Thanks for reminding mate. It think Entry point is the same important as Exit point, and it is more difficult to know when is the best time to entry.
2) About R:R, so before we put order, we must first calculate how many pips our target and our SL for that trade. Is that trade worth or not, some thing like that ?
If we have small R:R (lets say our profit is 1/3 of our SL), but the probability is >= 80%. Is it a good idea to take that trade ?  What are your minimum requirements before you take a trade ?

Thanks in advance bro.

 

 

[Innate]

This reply has been reported for inappropriate content.

Guys,

Has anyone considered using MAE/MFE analysis on the TZ strategy? It is a standard feature in the mql5 strategy tester.

Read more here: https://www.dukascopy.com/fxcomm/fx-article-contest/?Measurement-Of-Potential-In-Trading=&action=read&id=1861&language=en

 

Here is a script for mt4. Follow this page to the link, the poster explains how to implement it. http://www.forexfactory.com/showthread.php?t=89381

 

Now all you need is a TZ ea to test it with… 

• This reply was modified 11 years, 4 months ago by Innate. Reason: links didnt seem to work

What's it all about? It's all about money.

[aryafx]

This reply has been reported for inappropriate content.

it seems very interesting idea!
thankyou
as we understood : A Sub Sequence theorem description: Let n ϵ N and An^2+1 subsets with n^2 + 1 distinct real numbers.then there is some B sebsets ϵ A with sequence of – increasing/ decreasing -members.
i am thinking if we could use the subsequent theorem for the price (as a variety) then maybe we could use it also for CTZ sequence (as a variety) to find it’s probability sequencing for next PTZ goes to CTZ!(last CTZ in the B-SUBSET of CTZs will be higher or lower than other CTZ?)

[smallcat]

This reply has been reported for inappropriate content.

Zelo wrote:

Can we find out probability that price go from level A to E and E back to A, or in Forex, can we find out probability price go, e.g from 1.22900 to 1.22100 and from 1.22100 back to 1.22900

  This is very interesting bro. So if we get a strong level (may be a SR zone at big TF or a Node or a R-TZ ) we can know the probability where it will go (may be our target is a previous confirm R-TZ or F-TZ like Top TZ or Bottom TZ) , and how big is the probability it will come back to that level again. This is similar as Eurusdd did. Do i get it right ?

 

 

[Zelo]

This reply has been reported for inappropriate content.

Sorry for throwing out ideas and not explain clearly.

Because i don’t like discuss this in public, you know,

FF they banned Saver0, Kiads, Pip because of this.

And I prepared images and hoped some of you guys contact me, 2 times,

but no one did, i thought you guys don’t like it, so i lose interest, deleted images…

This is my last post in this thread.

I will update via my journal.

 

 

 

 

[MTH2014]

This reply has been reported for inappropriate content.

Zelo wrote:

Sorry for throwing out ideas and not explain clearly. Because i don’t like discuss this in public, you know, FF they banned Saver0, Kiads, Pip because of this. And I prepared images and hoped some of you guys contact me, 2 times, but no one did, i thought you guys don’t like it, so i lose interest, deleted images… This is my last post in this thread. I will update via my journal.

My Brother Zelo,

I believe there’s so huge different between our forum here compare to that ‘fucking factory’…

of course I really understand if you don’t want to ‘share’ your finding in public, but if you say there’s no one interested with your ideas,  you count me wrong.., I always interested with any new ideas and i believe there’s nothing wrong to discuss it here in public maybe in general term.., and I’m sure we discuss among brothers here.

Wish You All The Best..

MTH

 

 

Intuition, Experiences and Common sense..
http://www.binaryoptionsedge.com/

[smallcat]

This reply has been reported for inappropriate content.

Zelo wrote:

Sorry for throwing out ideas and not explain clearly. Because i don’t like discuss this in public, you know, FF they banned Saver0, Kiads, Pip because of this. And I prepared images and hoped some of you guys contact me, 2 times, but no one did, i thought you guys don’t like it, so i lose interest, deleted images… This is my last post in this thread. I will update via my journal.

If it is not for public, may be you can “limit” the people who can join the discussion. Of course you have the right to decide who may join and who may not. May be something like “Artificial Intelligence” group, some one must register first to join. May be ?
But we understand your hard work my friend … whatever decision you take, we can accept it …

[aryafx]

This reply has been reported for inappropriate content.

Dear Zolo!
Sorry for inconvenience in FF for you and some of Great men !but be sure that people are appreciative and could understand any sufferer man who doing the best of himself! i am interseted in your subject here as Kiads told kernel of say! please count me too if you like!god bless you

[Revolution]

This reply has been reported for inappropriate content.

Hi Guys,

 

I new here and reading your insights with interest. I have my own opinion on why you all have been “removed” from FF but I will keep it to myself.

I do understand that this forum supposed to be without influence of the “dark side” ;) , so why wont you limit discussion to registered members only as like have been mentioned above “some ideas don’t want to be shared in public” ?

I have followed transient zones on FF tread and I did some testing on TZ only, no other indicators..from 5k$ to 10k in 3 days…however:

– no Sl has been used

– biggest DD was 1,5k after tp was hit

– used heavy lots

– from 30 to 150 points tp used (3-15pip)

 

 

 

[simplex]

This reply has been reported for inappropriate content.

Innate wrote:

Guys, Has anyone considered using MAE/MFE analysis on the TZ strategy? It is a standard feature in the mql5 strategy tester. Read more here: https://www.dukascopy.com/fxcomm/fx-article-contest/?Measurement-Of-Potential-In-Trading=&action=read&id=1861&language=en Here is a script for mt4. Follow this page to the link, the poster explains how to implement it. http://www.ff.com/showthread.php?t=89381 Now all you need is a TZ ea to test it with…

Interesting links! Might help a lot – thanks!

• This reply was modified 11 years, 4 months ago by simplex. Reason: stripped FF link down
• This reply was modified 4 years ago by wise_jedi.

A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top- and bottom-fishing to people on an ego trip. (Dr. Alexander Elder)

[GU_Night]

This reply has been reported for inappropriate content.

Zelo wrote:

Sorry for throwing out ideas and not explain clearly. Because i don’t like discuss this in public, you know, FF they banned Saver0, Kiads, Pip because of this. And I prepared images and hoped some of you guys contact me, 2 times, but no one did, i thought you guys don’t like it, so i lose interest, deleted images… This is my last post in this thread. I will update via my journal.

I’m too very interested in your ideas Zelo! What about starting a group for only your ideas, Zelo? I want to join such a group for sure, and I’m sure I’m not the only one. 

There are children playing in the streets who could solve some of my top problems in physics, because they have modes of sensory perception that I lost long ago. ~Oppenheimer

[hannele]

This reply has been reported for inappropriate content.

GU_Night wrote:

Zelo wrote:

Sorry for throwing out ideas and not explain clearly. Because i don’t like discuss this in public, you know, FF they banned Saver0, Kiads, Pip because of this. And I prepared images and hoped some of you guys contact me, 2 times, but no one did, i thought you guys don’t like it, so i lose interest, deleted images… This is my last post in this thread. I will update via my journal.

I’m too very interested in your ideas Zelo! What about starting a group for only your ideas, Zelo? I want to join such a group for sure, and I’m sure I’m not the only one.

Count me in

cheers

hannele

[simplex]

This reply has been reported for inappropriate content.

smallcat wrote:

I am afraid of trading without SL. Although i had SL on most of my previous orders , i had blown my account several times.

So am I – and I believe it’s ok to follow this instinct in order to survive as a small fish in the FX sharks’ basin. But ANY SL level can be hit in rough markets.

smallcat wrote:

1) This remind me on WHEN we must enter a trade, so we must know the TREND at the time we want to enter. It think Entry point is the same important as Exit point, and it is more difficult to know when is the best time to entry.

The point is that trend detection will always (hmm – really?) incorporate some kind of averaging algorithm, and this will introduce lag.

smallcat wrote:

2) About R:R, so before we put order, we must first calculate how many pips our target and our SL for that trade. Is that trade worth or not, some thing like that ? If we have small R:R (lets say our profit is 1/3 of our SL), but the probability is >= 80%. Is it a good idea to take that trade?

I’m afraid I don’t have elaborate rules for that yet. I started to think about calculating hit probabilities a little more than two weeks ago, when somebody pointed me to the similarity thread over at FF.

What I’m thinking about at the moment is not to care about short term momentum, but just after a new candle has been opened to look at the potenial fractal and mid-bar zone that have been defined with that particular open, while h being extremely short: 1 or 2 bars only, TF H4 or Daily. If both zones meet general entry conditions (to be defined) we have to find our entry.

My first idea was to go down to M5 / M15 at that point and then detect a reversal by some classic TA indicator, or by nonparametric stats like Spearman algorithm – both introducing lag.

My latest idea is just to wait which zone (potential mid-bar or fractal) is being eaten up by price action first, and then set a trade in the opposite direction. Since probability for fully resolving with such a small amount of h is small (little more than 50 %), k comes into play: it should be dynamically defined as a share of the initial width of those PTZ and enhance probabilities.

I think it could be ok to use some multiple of ATR to calculate SL. ATR could also play a role as a weighing factor when calculating k for that particular trade.

This might be supported by a currency strength indicator to sort out which pair and direction to trade and which not to trade (replacement for single pair trend detection). I am fully aware of the fact that this would introduce lag into our process.

Still a bit rough, but what do you think?

A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top- and bottom-fishing to people on an ego trip. (Dr. Alexander Elder)

[smallcat]

This reply has been reported for inappropriate content.

Innate wrote:

Guys, Has anyone considered using MAE/MFE analysis on the TZ strategy? It is a standard feature in the mql5 strategy tester. Read more here: https://www.dukascopy.com/fxcomm/fx-article-contest/?Measurement-Of-Potential-In-Trading=&action=read&id=1861&language=en Here is a script for mt4. Follow this page to the link, the poster explains how to implement it. http://www.forexfactory.com/showthread.php?t=89381 Now all you need is a TZ ea to test it with…

Thanks mate, this is interesting. i have downloaded it, will try later in future if i have some free time …

[smallcat]

This reply has been reported for inappropriate content.

simplex wrote:

The point is that trend detection will always (hmm – really?) incorporate some kind of averaging algorithm, and this will introduce lag.

Hi bro,
I am agree about lagging of “averaging”. I tried 3 EMA crossing to find the entry point in trending last years, but because of lagging, some times the signal given is near the end of the trend

simplex wrote:

… while h being extremely short: 1 or 2 bars only, TF H4 or Daily. If both zones meet general entry conditions (to be defined) we have to find our entry. My first idea was to go down to M5 / M15 at that point and then detect a reversal by some classic TA indicator, or by nonparametric stats like Spearman algorithm

I am not sure about this, but i think using h=1 or h=2 , will give us recurrent zone every where. Each bar will be recurrent.  May be some one can get benefit from this, but it is still difficult for me.
TA indicator like a momentum/trend/volume/oscillator  indicator ?   I remember GG’s post here :
http://penguintraders.com/forums/topic/currency-strengthpowervalue-indicator/#post-1692

simplex wrote:

… k comes into play: it should be dynamically defined as a share of the initial width of those PTZ and enhance probabilities. I think it could be ok to use some multiple of ATR to calculate SL. ATR could also play a role as a weighing factor when calculating k for that particular trade. This might be supported by a currency strength indicator to sort out which pair and direction to trade and which not to trade (replacement for single pair trend detection). I am fully aware of the fact that this would introduce lag into our process. Still a bit rough, but what do you think?

If we understand “K” properly + currency Strength, may be we can get some pips in big TF just looking at PTZ.
I like ATR, but do not know how to use it to define our SL. Any clue ?

 

[simplex]

This reply has been reported for inappropriate content.

smallcat wrote:

I am not sure about this, but i think using h=1 or h=2 , will give us recurrent zone every where. Each bar will be recurrent. 

I don’t think so. A large number of bars will be equipped with rather large POTENTIAL fractal and mid-bar zones. As time passes, those large zones are partially eaten up by price action. Many bars will keep some part of those initial zones to become confirmed. The ratio of final width of a confirmed zone to the initial width of a potential zone will basically define k. This is where ATR may join in as a weighing factor to achieve some ‘universal’ algorithm defining k. I ran stats about the first part, without incorporating ATR. It looks promising, but still has to be evaluated more properly. In general: the smaller the value of h, the more potential zones for entry you will find, but the smaller k will be in order to achieve statistically safe entries.

smallcat wrote:

May be some one can get benefit from this, but it is still difficult for me. TA indicator like a momentum/trend/volume/oscillator indicator ? I remember GG’s post here : http://penguintraders.com/forums/topic/currency-strengthpowervalue-indicator/#post-1692

Sure, I’m aware of that post and the ambiguity of RSI, Stoch, etc. readings. Just read my reply to his post. That’s why I’m here in this forum, trying to find alternatives.

smallcat wrote:

If we understand “K” properly + currency Strength, may be we can get some pips in big TF just looking at PTZ. I like ATR, but do not know how to use it to define our SL. Any clue ?

I would use currency strength as a basic filter: just to sort out which pairs to trade at the moment, and in which direction. No entry signals from this side, IMO. For the role of ATR, there’s my rough idea of defining k (see above), and there are ‘classic’ ways to set a SL or TP. I think testing a combination of both might be worth the effort. This article gives an idea of a classic approach, though there are many more of them: http://www.investopedia.com/articles/trading/06/stopplacement.asp

 

A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top- and bottom-fishing to people on an ego trip. (Dr. Alexander Elder)

[Author]

Posts