Fractal Finance: (Multi)Fractality in the FX Markets

This thread treats the topic of fractality in finance.

The focus is not on academics but on the relevance of the topic for trading.

It is recommended that before you participate you read the following books:


Benoit Mandelbrot: Fractal Finance

Benoit Mandelbrot: The (Mis)behavior of Markets

Nassim Taleb: The Black Swan

(will be edited)


No rules on this thread. But please stay focused on the topic.
(will be edited)

Thanks and yes, I understand the difference between pattern and price behavior. In fact, I am very fond on using the term "behavior" as there is incredible uniqueness that currency market presents in regards to price behavior which results in patterns (I have used "patterns" for the chartists here to understand what I mean when I talk about my trend phases). This is especially true because of the "pair" behavior that exists which drive the mathematical correlations that have to exist and conform (hence my phrase... "Do what Markets HAS to do to survive" stems from).

By Affinity, do you mean attraction and repletion as in magnetism in price behavior? if yes, how you would quantify it just by looking at the data?

Awesome! Thanks man. I am only looking for alignment with what I already do (as you know) so this will help support my work/theory :-

EDIT: Moving my posts from the other thread to here for some good discussion

FF,

Quick Question, if I may, to tap into your knowledge on Mandlebrot theories and fractals but a disclaimer first. I have NOT researched Fractals in my journey as I ruled them out based on my elimination on Highs and Lows from my design (plus the repainting element of free fractal indicators... (I had a light bulb moment when I decided to ignore the tip of wicks from my strategy).

The way my own learning and development works (as you can probably tell by now) is that I drive my knowledge and conclusions on my own first and then seek alignment (just in case you are thinking why am I not spending time reading/researching Mandlebrots work on financial markets..... I will but I want to stay true to my own journey first and foremost. I am sure I will get to most of his research in due time)

my question: In your research, do you believe a larger "pattern" (I don't like the word "pattern" as it reflects I am seeing it on chart) or a "data set/series fingerprint" CAN comprise of recursive sequence of smaller fingerprints? By this I mean (and since most here rely on charts), Could a pattern on say Monthly or even yearly or decade timeframe consists of series of exact "same" micro-patterns on WK > D1 > H4 > H1 > M30 > M15> M5 > M1? I am asking this question at the risk of revealing what I do... but no worries, mine cannot be traded manually due to volume of data science involved so I am ok :-)

Appreciate a "long" and detailed response whenever time permits :-

"EDIT: Moving my posts from the other thread to here for some good discussion"


Ok...

can you please make screenshots from your and my posts in the other thread and post them here?

Thx

Moving your post here for reference....

Moving your post here for reference....


Awesome.... we are digging deep bro. You got a truck load of information in this post which I will dissect. Could you please share the link to the source of this image? Interesting to see you say Why and not HOW.... In my research and technique, I know the How and not the Why lol.

"NATURE of the price behaviour" - agree

"it CANNOT predict what the price does next" - agree. In fact, I can confidently argue that risk based models based on price is a complete fallacy. Some here have called me out by saying "i am full of it". I am of the belief that EVERY bit of analysis changes (or becomes obsolete) the second you take an entry. Agree?

"even what pattern should occur next." - this is not important because I am of the understanding the entry setups (patterns or whatever you want to call it) will never be the same for exits i.e. entry setups can never be the same as exit setups (even if you reverse the signals for exits). This lead me to believe that Price Action is a complete Fallacy. Price is NOT everything.... call me nuts !
"multi-fractal models CAN BE EFFICIENTLY USED for predicting the VOLATILITY of the price"... I agree but with a twist.. Here's how. I use multi-fractals (via Inception i.e. time slicing of price structure) to very accurately determine price levels and by levels I do NOT mean the levels visible on a chart ! By levels, I mean THREE in all.... which I refer as overbought, oversold and equilibrium. I know the HOW but I do not understand the WHY aspect of why most if not all my trades fall exactly on such levels (example of what I showed in my trade screenshot from last night). I think it is because, I have found (discovered?) a way to eliminate "manipulation" or "market noise" from my design by IGNORING WICKS! but how can I provide evidence of it? I am trying to figure out the WHY for almost a year now but thus far, my entries are working well because I know with 100% surety, price will either bounce off or breakout from either these THREE levels.
Overbought: Breakout depicts fakeout/stop hunts if you will.
Overbought:Bounce presents filling in the liquidity void if you will.
Same for Oversold.... now, with Equilibrium (which I refer to SR Flips but know that I do not believe in Support or Resistance or Support and Demand zones or pivots or fib levels... none of that crap ), Price could go either way BUT I have not been able to filter out SR FLip Zones because I do not understand the WHY.... you see where I am going with this? Because I do not know the "Why", I had to develop a complete and comprehensive strategy to tackle all three scenarios and trade all three setups which lead me to the discovery of Entries are not as important as exits. I make money with with Exit strategy which in itself is the most complex design. My complex multi-tier comprehensive exits lead me to my "Universal Entries" that will work in ANY market condition leading me to blelieve that Market is Neutral... regardless of whether market is trending or ranging. It will work with any trading style (scalping, day trading, swing trading, position trading or whatever). All this is possible with just one set of entries... similar (I think) to the Mandlebrot signature)... why? because I consider market to be trending at ALL times and use TIME to seek out volatility and then use either breakout or bounce setups for direction and my exits to weaponize my exposure frequency.
Sorry, i got carried away with sharing way too much than I have in the past.

Edit: I just made a few edits

There are not in the screenshots. The best way (i think there is also a scientific/medical reason) to train your mind i.e. focal interpretation is to run the chart setups in slide show a 1000 times to make the setups become second nature. In Trading, I prefer that there is no thinking or analysis involved.. only entry and exit with focus on execution. With your charts, there is ton of information but they all lack the level of reference-bility if you know what I mean. Here, your readers are trying to decipher algorithmic market manipulation. The only way to beat an algo is to become the machine and train yourself to capture all the scenarios. That is the only tool they would have at their disposal imo.

Sorry, one more question...

"multi-fractal models CAN BE EFFICIENTLY USED for predicting the VOLATILITY of the price"... could you share more information or links on how volatility could be predicted? My entire strategy revolves around volatility and I use TIME to seek out market volatility (based on my years of watching price behavior round the 24 hour clock!) Are you referring to predicting "breakout-infussed volatility" as in volume/clusters of orders triggered to inject retail margin liquidation episodes (which I fully understand) or is there more to it ?

There is a reason why I call my strategy "Universal Time Action" as Time is the only element in the universe (or market) which is Neutral and cannot be manipulated.... but the ***deleted***

Moving this post here from the other thread

FF,

Quick Question, if I may, to tap into your knowledge on Mandlebrot theories and fractals but a disclaimer first. I have NOT researched Fractals in my journey as I ruled them out based on my elimination on Highs and Lows from my design (plus the repainting element of free fractal indicators... (I had a light bulb moment when I decided to ignore the tip of wicks from my strategy).

The way my own learning and development works (as you can probably tell by now) is that I drive my knowledge and conclusions on my own first and then seek alignment (just in case you are thinking why am I not spending time reading/researching Mandlebrots work on financial markets..... I will but I want to stay true to my own journey first and foremost. I am sure I will get to most of his research in due time)

my question: In your research, do you believe a larger "pattern" (I don't like the word "pattern" as it reflects I am seeing it on chart) or a "data set/series fingerprint" CAN comprise of recursive sequence of smaller fingerprints? By this I mean (and since most here rely on charts), Could a pattern on say Monthly or even yearly or decade timeframe consists of series of exact "same" micro-patterns on WK > D1 > H4 > H1 > M30 > M15> M5 > M1? I am asking this question at the risk of revealing what I do... but no worries, mine cannot be traded manually due to volume of data science involved so I am ok :-)

Appreciate a "long" and detailed response whenever time permits :-

Links to responses... here and here

one more move from the other thread...

You have twitched the OCD in me with this post :-(

"BUT, of course it is NOT the case that when you have a grail pattern on the day chart, that now there must be smaller grail patterns on the M15...." Agree. this is not the right thinking or approach that would work otherwise every EA would be profitable. The mtf "pattern" can't be a one here and one there. There needs to be some sequencing to trigger/simulate a divergence that will result in a fat tail in pricing and account equity. Crucial Point does this on tick +M1 charts. I expanded this to include over 32 timeframes (which I refer to as spatial context) to simulate divergence in price (as I consider market to be trending 100% of the time i.e. Diverging with varying SCALE within a given time series. With this universal rule i.e. Mandlebrot signature or sequencing of signatures or behaviour, one can identify/trap movements regardless of the driving principles behind those moves. There needs to be an alignment (or magnatism or confluence or Affinity (?) between these various signature sequencing to capture a bounce or a breakout....but all this is just for entries or standardization of entries. What happens at the next tick is anybody's guess same with any other entry signal. You are able to predict because you have the necessary data sequencing tools to predict direction that is just not available to us retail traders. We can only ANTICIPATE direction and take a firm stance in one direction or the other and play with our exit/trade/money mgmt techniques....agree?

Edit: I will share some more on trend phases. With the above sequencing principle, I slice and dice data series into 4 different trend phases for each UP and Down Direction i.e Trend Reversal, Trend Reversal Confirmation/Continuation, Trend Continuation (depicting Pullback), Trend Continuation Confirmation. I determine these by only looking at the previous closing price and I m a non-believer in price or HH/LH type definition of trend (remember I ignore wicks which means I ignore Highs and lows).
With these 4 distinct trending market Signatures, A comprehensive strategy can be devised that is independant of market conditions....thoughts?

Edit - made some edits

Can someone make screenshots from my comments in the other thread and post the screenshots here?

I cannot always do everything on my own.

Thx, FF

Here the MFMAR paper.

FF

Here some other papers:

https://papers.ssrn.com/sol3/papers....act_id=2338439

https://papers.ssrn.com/sol3/papers....act_id=2448648

https://papers.ssrn.com/sol3/papers....ract_id=182778

Stay safe,
FF

ufff.....ok here then:

So, understand:


I have shown you WHY a price goes up and down and that it is predictable (see my thread on FX Exposed: Market Manipulation, the videos etc.).

NOW,

this thread shows METHODs of

- simulating
- modeling
- describing
- analyzing

the price BEHAVIOUR.

So in laymen terms, "how it moves", NOTm "why" and NEITHER in any prediction context whatsoever.

SO this here is closer to academics.

FF

Any active FX trader will prefer the other thread which exposes the live market manipulation (DM vs SM).


Here is how Mandelbrot / Taleb approached markets WITHOUT HAVING that specific manipulation knowledge:


They did PURE, STATISTICAL analysis, looking at HOW prices behave / move.


So in other words, they took massive data sets of prices (by the way, mostly on HIGHER TFs....they consider lower TFs "noise").


Then they used mathematics to model / describe / characterize the GENERAL behaviour of the prices.

In that context,

Mandelbrot applied NON-mainstream model techniques (from outside of finance, mainly physics) to financial prices.


"Established" academics were angry about it since basically, this exposed how INACCURATE / PLATONIAN / WRONG their approaches and models were.

See the "CAPM" as an example (Capital Asset Pricing Model).


Still, it became clear that Mandelbrot's methods (which started by describing turbulence in physics) were VERY useful for financial researchers and MUCH more efficient with regard to price analysis.

Again, remember,

academics are NOT traders,

they are NOT interested in predicting anything but rather in DESCRIBING / ANALYZING "how things are".


Also note that the current state of financial academia is the equivalent to medicine a few hundred years ago.

FF

SO understand;


a multi-fractal model will NOT predict where prices go next.

It is just a BETTER "ruler" for bringing some "order" into the price chaos.

Stay safe,
FF