The similarity system | Page 651 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Similarity: A Practical Approach 859 replies

Similarity Trading - The ZZBB Method 215 replies

Rainbow similarity 256 replies

Similarity Stuff 17 replies

Directional Similarity 0 replies

Attachments: The similarity system
Exit Attachments

The similarity system

  • #13,001
  • Nov 21, 2014 12:39pm Nov 21, 2014 12:39pm
Quoting maximinus
{quote} Don't be sorry, you are one of the few here who takes the whole thing with rationality and produces useful tools while others produce wonderful trade explorers. Yes easylanguage it is. I've some basic knowledge in EL and I code every indicators that I use, never used indicators made by others unless the code is also provided in .txt format. When you have time, there is really no emergency, throw a simple code with few lines here, so that I can check if that's similar to EL and if I've the level. Yes everyone tells me that once you code in...
Used EL between 1988 and 1995, just took up MT4 and it was an easy transition. Not too many different ways to describe high low or close after all. Any decent developer environment allows you to see examples of what you are trying to code as you code it. Just take the time to figure out the reference "book" that can be open in a side window.
  • #13,002
  • Nov 21, 2014 2:49pm Nov 21, 2014 2:49pm
Quoting priceaction1
point clear I was not go through all the thread . @ Baillie I am sorry,
Hey no problem man!

If more people here would admit their mistakes instead of skirting them, this would be a better place. Subscribing to you.
  • #13,003
  • Nov 21, 2014 3:45pm Nov 21, 2014 3:45pm
Quoting Favorite
{quote} I'm here! Do you want to talk about that, sucker?
You are a blind man
Attached Image
  • #13,004
  • Nov 21, 2014 4:03pm Nov 21, 2014 4:03pm
Wow! What the heck happened here?
I havent' got time to read all posts so I'm usually just
dropping by to see if there's something interesting I can use
for scalping.

From what I understand and pls. Correct me if I'm wrong, Infinitus is
well Infinitus looks like and sounds like he is bashing the guy
and then there's Jay who claimed to have made around 200K from Eurusdd's ideas
and yet also bashing Eurusdd. Am I missing something here?

I can understand Infinitus' reasons but comeon Jay, lets just say Eurusdd just
gave you 200k reasons to at least give it a rest now. Everything is on this thread now
and let everyone see Eurusdd for what they think he is.
It's really good to see how well you've done but it won't hurt to remember who helped
you.

Wishing you the best.
G
I read and watch to learn.
  • #13,005
  • Nov 21, 2014 4:39pm Nov 21, 2014 4:39pm
Quoting LiquidGenius
{quote} Hi infinitus, Thanks for the discussion. I think regardless of the notion of "good or bad", there is a good deal of things to learn from this thread. I'm not quite sure what you're arguing (sorry for not studying your posts much) but I think we can come to some sort of compromise on what's "true" and "false" as far as the propositions go and the numbers to back it. First, regarding the actual probability {image} This is what eurusdd has actually said. Here's a quote from a post I made in the past concerning it: "The auto assumption that...

Hi LiquidGenius,

great work, great response. Thank you very much for that. Will dig through that and answer on that way. So do not be disappointed, when I can not address all in the next hours LOL But the weekend is long, so stay tuned.

Let me say something first:

it was never about good or bad. It is about the right application of math and statistics.
  • #13,006
  • Nov 21, 2014 6:03pm Nov 21, 2014 6:03pm
Quoting LiquidGenius
{quote} Hi infinitus, Thanks for the discussion. I think regardless of the notion of "good or bad", there is a good deal of things to learn from this thread. I'm not quite sure what you're arguing (sorry for not studying your posts much) but I think we can come to some sort of compromise on what's "true" and "false" as far as the propositions go and the numbers to back it. First, regarding the actual probability {image} This is what eurusdd has actually said. Here's a quote from a post I made in the past concerning it: "The auto assumption that...
Nice post LiquidGenuius! Thanks very much. It was your blog and stats that inspired me to make my own indicators and play around with the probababilities. Can't thank you enough for that.
You reap what you sow.
  • #13,007
  • Nov 21, 2014 6:07pm Nov 21, 2014 6:07pm
Quoting stt
{quote} thats good to know. anyone serious in trading should equip himself with stats knowledge and some programming (lot of libraries in R/python now so you dont need C++/java knowledge). My background was in computer science and math before moving into finance. so quant trading was an easy fit for me. My team (just like any other team) research strategies and their implementation - but note our objective is to come up with strategies that have capacity (and our expectations for returns are very modest - 4% unlevered is good, 6% is great and 9%...
Yeah I'm sure little returns when working with big capital is huge. And risk must be very small as well. I actually am reading Chan's book and it is really a good read, however looks like I need to learn some coding basics. I have heard that Python is a fairly easy language to get started with so I am definitely looking into that. In the future I will most likely be learning C and C++ as well. The future will definitely be technology based so learning programming could definitely help a lot. It's just stuff I do in my free time when I'm not scalping the crap out of these markets haha. Is your fund diversified in different markets or just trading equities? And what got you into trading FX?
  • #13,008
  • Nov 21, 2014 6:15pm Nov 21, 2014 6:15pm
Quoting LiquidGenius
{quote} First, regarding the actual probability {image} This is what eurusdd has actually said. Here's a quote from a post I made in the past concerning it: "The auto assumption that came from this is that if the h value is correct, 3% of bars should be transient or contain a transient zone. Is this correct? I assumed so for a VERY long time, and many others did as well. However, after thinking about it for a long time, what I'm beginning to think is that the indication is here is that not that the percentage of bars that are transient is 3%, but...
a) 97% recurrent - 3% transient (or 80% - 20% like NorthTrader found)

what does this mean in the bigger context? It simply means, that there are only very few price levels in the market, that did not became recurrent, yet.

So when one has enough time, low leverage or deep pockets, we have a very high probability, that we will see the specified price level again in the future.

What we do not know is, when the specified levels will be revisited again. Money Management has to account for that.


But keep in mind: some price levels have not been revisited in years


The 3% number out of my view is the number for the transient bars. Transient bars are rare. And when they happened, there is a great probability that one of the next bars will be recurrent


One has to simply look at how EURUSDD advised to trade:

http://www.forexfactory.com/showthread.php?p=7679105#post7679105

He states:

You do not know if that region is transient or recurrent yet because it is still forming. What you do though is wait. The moment price goes above that region, you can assume it is transient and then prices very close above it recurrent. Those prices are the ones you will expect a re-visit within at least one of the remaining h bars


One posting later:

if price goes above the region before it is done forming, you can assume it is transient and the prices above it recurrent. Those prices are the ones you will expect to be re-visited, not the prices in the region.



I do not know if that is a wise approach to trade, because it is very fragile, and optimally we should look for solutions that are not fragile.

This method is fragile, because when price gets transient, we can assume that there was a reason or shift in social mood which made the price moving/transient. When we are not able to determine the force behind the move, we may end up betting on recurrence but waiting for a very long time to occur.

Neio once nicely laid the problem down, that a series of transient bars can occur in trending markets.

http://www.forexfactory.com/showthread.php?p=7689031#post7689031

Transiency is like "a trend starts". Non-fragile methods go with the trend. Betting on recurrence hence is fragile. You may end up betting on a e.g. 10 pip move in front of a 100 pip move against you

The question is: when bar got transient, where is the best place to bet on recurrence, so that the fragile component may be very low. Or is it much wiser to bet in direction of new trend when transiency occurs?

As I see it, no agreement on a best way to trade a transient bar is reached in the threads. That is a bit strange, when most think that betting on recurrence of transient bars has such a high probability.



to be continued......
  • #13,009
  • Nov 21, 2014 6:22pm Nov 21, 2014 6:22pm
Quoting LiquidGenius
{quote} So, when you face a scenario like this: {image} The probability that the current low (which is entering a price range where no bar in the past 99 bars has entered) the probability that the low is taken out is somewhere between 80 and 86% (80 being a low estimate according to NorthTrader, 86% being the rounded number of "probability of resolving" according to the indicator). This is NOT 97% and I don't think anyone can argue with good statistics that it is. There is no error here..

Exactly, I argued this earlier, asking:

would you bet a big part of the farm on a 97% probability?

would you bet the same on an 80% probability?
  • #13,010
  • Nov 21, 2014 6:27pm Nov 21, 2014 6:27pm
Quoting LiquidGenius
{quote} There is no error here. However, if we look to prove eurusdd's claim and aim to find a price zone that IS 97% recurrent (the proposition I quoted), there IS a way to do this, which is different from the method that NorthTrader, myself, and others have tried. The approach does meet eurusdd's propositions: that if we are currently at price P, and none of the previous h bars has been through P, the probability that one of the next h bars will make P recurrent is 97%. Your first question I would expect is, "is it tradeable?" (and if yes, then...


Great work, man. I expected such a result. Hence my constantly questioning of the indicators.

Extremely thankful for that........
  • #13,011
  • Nov 21, 2014 6:39pm Nov 21, 2014 6:39pm
Quoting LiquidGenius
{quote} So, there is indeed a way to use the idea of transient/recurrent bars to create a scenario in which 97% of the time, price will return to it (for a sake of all the 'haters' saying you should read the thread, I will say that the method is mentioned/explained/strongly hinted here and not 'give it away'). But again, I do not think that it's tradeable. Is there a possibility that it is though? This point I think is unknown, as (AFAIK) not a lot of people have put in resources to discover it. combined with some sort of MTF or other information...


When we look at the charts with the indicator on it, it looks very much like supply and demand zones. I think one may trade it accordingly.

This was my first impression, when I saw the first pictures of those charts. Someone even argued somewhere in the thread, that the indicator is derivate of the SupDem indicator.


If the method of similarity via the way of the indicator would generate really strong tradeable signals, trader would easily find and agree on trading method JMHO

I can not state enough: when someone produces gains with this method/indicator, chance is high, that it is not mainly because of the method/indicator but their individual trading skills.
  • #13,012
  • Nov 21, 2014 6:46pm Nov 21, 2014 6:46pm
Quoting LiquidGenius
{quote} Second, I can't find the posts that you mentioned it, but I believe you said you were an elliot/wave trader. Transient bars are no more than n bar fractals. Fractals are simply swing points. Connection of swing points=waves? {image}



we Elliotticans try to judge the next move by applying probability. The wave count with highest probability is the main count.

Often enough it gets invalidated


probability is not certainty
  • #13,013
  • Nov 21, 2014 6:49pm Nov 21, 2014 6:49pm
Quoting LiquidGenius
{quote} I've editted out a lot of stuff here (on screen statistics that I use. Note: currently not profitable) but when viewed this way, are the ideas that far off base to be worth something? I know there are various problems with using this fractal/wave idea, but I think that is best left up to the trader to find out for him/herself. Personally, I thank eurusdd for presenting the idea in the manner that he did. I have been trying to trade waves for a long time now (unsuccessfully) and this thread was my "light bulb moment" and basis point that...


fair enough


My impression is, that many here in the thread, who think to have profited, have learned something about supply and demand without noticing it.
  • #13,014
  • Nov 21, 2014 6:54pm Nov 21, 2014 6:54pm
Thanks again LiquidGenious for this fine posting
  • #13,015
  • Nov 21, 2014 11:31pm Nov 21, 2014 11:31pm
Quoting infinitus
Thanks again LiquidGenious for this fine posting
Definetly I agree on what above!
  • #13,016
  • Edited 2:02am Nov 22, 2014 12:54am | Edited 2:02am
Quoting infinitus
{quote} fair enough My impression is, that many here in the thread, who think to have profited, have learned something about supply and demand without noticing it.
I don’t know if you are saying this exactly, but I agree that the way it can be traded as suggested by Eurusdd is difficult to trade profitable. Because that way only works imo in ranging markets (without too much drawdown). Or/and as Neio said it is not a tradable 97% most of the time.

But there are indicators made by kprsa and Northtrader, who give statistics that I can’t deny, they say that fractals under certain conditions (H) were taken out > 80% of the time within H. And that is in my opinion important ‘within H’, as you highly probable know; ‘within a certain time’.

If one starts thinking with this statistics in mind, they can come up with strategies that work in all market conditions, and it is different from demand and supply at least because the trades have a expiration time, if the fractal is not cleared within H one can close the trade. Interesting thing here is in my opinion that if the zone is not cleared within H, often the market has tried to do so against the expiration time.

I don’t have statistical proof to back the next sentence up; but I believe that the possible TZ’s with the current trend have a higher probability to be taken out within H than the ones against the current trend, but that is just common sense.

As you said in a earlier post, most traders here have shown a casual way to deal with probabilities. I agree with that, I’m one of them too. As far as I understand you have to trade probabilities all of the time and I don’t do that, because these percentages/probabilities are only a part of my setup. I implemented it in the way I traded before I read about TZ zones, and it has nothing to do with the way it can be traded as shown by Eurusdd.
  • #13,017
  • Nov 22, 2014 2:41am Nov 22, 2014 2:41am
Quoting GU-Night
{quote} I don’t know if you are saying this exactly, but I agree that the way it can be traded as suggested by Eurusdd is difficult to trade profitable. Because that way only works imo in ranging markets (without too much drawdown). Or/and as Neio said it is not a tradable 97% most of the time. But there are indicators made by kprsa and Northtrader, who give statistics that I can’t deny, they say that fractals under certain conditions (H) were taken out > 80% of the time within H. And that is in my opinion important ‘within...


Hi GU-Night,

thanks for response.

Yes, that was my first impression ever, that this concept can work out in ranging markets, but may kill one in suddenly trending or strong trending markets.

If such a trend or longer impulse occurs, one has to have good money management, and trading method has to be changed accordingly.


but I believe that the possible TZ’s with the current trend have a higher probability to be taken out within H than the ones against the current trend

I would second that sentence. Have to think about it for a while to come up with explanation. It may have something to do with the fact,that in order for price to make progress, you simply have to leave zones behind. Maybe it is then, that when countertrend TZ's would all be worked off too, there would be no progress possible, because price would maybe form a stable state. But I am not sure about that at the moment.


I have tried to improve my trading with calculating probabilities since the last 4 years in one way or another. E.g., I have a way of approximating the chance for reversal of price in a system of Murrey Math within a MTF setup. It works great in sideways correction, but in especially strong trending moves it can kill you with ease.

Another example is my way on looking at trading volume. In consolidation phases price builds Gaussian- or Gaussian like distributions of price/volume. So you can bet on mean reverting, when price is at the outer boarders of the Gaussian curve. Works pretty well in sideways phases, but can be a pain in trending moves.

When EURUSD started moving down back in May, both my statistical systems gave me very strong buying signals in the 1.34 region. After a bit of consolidation, price just decided to go more south. On most of the distribution-curves I had price simply left the proverbial building, building long routes of so to say transient zones, which were often enough not even slightly tested. At some point price had run out of nearly all the distributions on the different TF's. In my "world" of thinking "mean reverting" one could have made a big bet on price to mean revert to the upper side, but price visited the 1.25 area before the first significant mean reverting process happened. Price did this, although the calculated chances were stuck against it over weeks.

It was all of my experience over the course of the last 4 years that made me super suspicious about the claimed very high probability rates here in the EURUSDD threads.


One last word to the indicators:

guys, you have really done a lot of very good work here. That is for sure. And when this indicators only help a few to improve decision making the goal is reached.

I think with the indicator is somewhat similar to the use of Elliott Wave. In the back-mirror view the turning points, transient zones, wave countings etc. are crystal clear and one asks oneself, why it was not so clear when life trading.

Simply said, it is the ongoing, building zones or waves. When they develop, I think, we have to reduce our expectancy rate very much, because only in hindsight we can see that it developed one way or another



One word regarding "to deal with probabilities in a casual way":

as long as we are aware, that we are dealing with it in a sloppy way, it may be OK.

E.g.: I think it is pretty meaningless at the end for trading, if you think you have a probability for XY of 80% or of 97%. I would take both trades anyway. What one should wish more when one gets handed a probability of over 80%

Another critical question is, how we handle probability or how we look at it:

when the indicator postulates a probability of over 97% in the past, it does not necessarily mean that we can have the same expectancy in an ongoing process.
  • #13,018
  • Edited 5:09am Nov 22, 2014 4:38am | Edited 5:09am
Quoting infinitus
{quote}I think it is pretty meaningless at the end for trading, if you think you have a probability for XY of 80% or of 97%. I would take both trades anyway. What one should wish more when one gets handed a probability of over 80% Another critical question is, how we handle probability or how we look at it: when the indicator postulates a probability of over 97% in the past, it does not necessarily mean that we can have the same expectancy in an ongoing process.
May I give you my 2 cents?

The probalility of what the ZT zone is closed in H bars is even lower !!!

May I explain?

1/ The indicator give us the ratio of 97/3 (or 98/2, or 99/1, if you like). What does it mean? IT meas that the probability of what the TZ is closed is 97%. WHEN? it should be in H bars, should not it? NOT! (I will explain a little bit later) So, WHEN? - in 5 mins, 5 hours, 5 months and so on. Moreover, you should NOTE, you can see that the indicator gives you "a window of probability taking" that is from 1000 to 2500 bars. So if we spread our vision to the rest of the market, let us say, we take 250000 bars, for example, we will be limiting by another ratio. I will not exactly say what the ratio is, but I personally think it will be estimated as 80%-to-20% (this is the Pareto rule http://en.wikipedia.org/wiki/Pareto_principle).

2/ So, is this the ratio of 80-to-20? NO, for sure, it is not. WHY? Look at this table (http://www.forexfactory.com/showthread.php?t=434603&page=651), for example. It is just one of many similar tables that could be found out, if you use the statistics calculated with indicators published in here. I have found that such tables (I took another one) bring us to the Negative Exponential Distribution of probabilities (http://en.wikipedia.org/wiki/Exponential_distribution) that means that the real functional bar range is up to 30-50% of H bars only. In other words, no more than 50% of the H range choosen will be effective for price coming back to the PTZ levels - the probability of this event is exponentially decaying very fast then, and the price will not come to test the PTZ zone, it will go in the opposite direction more precisely.

I hope it is helpful for trading

Edited. So, the real probability of what the price comes back to the PTZ and closes it after the key bar has been formed is no more than 0,5*0,97=0,485 that is 48,5% as a limit.

Best,
D.P.
Hunting High and Low
  • #13,019
  • Nov 22, 2014 5:58am Nov 22, 2014 5:58am
Quoting Dr.Phoenix
{quote} I hope it is helpful for trading Edited. So, the real probability of what the price comes back to the PTZ and closes it after the key bar has been formed is no more than 0,5*0,97=0,485 that is 48,5% as a limit. Best, D.P.



Great work, D.P.


That was my impression from the beginning, that the real probability is very much lower than most people thought. But that it would be so low even I was not aware.

Have to think a bit longer about what you wrote.


That would not be tradeable at all - coin toss........

So what is it, that made some traders here believe that one can trade the similarity thing and the indicator successfully?
  • #13,020
  • Edited 6:12am Nov 22, 2014 6:01am | Edited 6:12am
Quoting infinitus
{quote} Great work, D.P. That was my impression from the beginning, that the real probability is very much lower than most people thought. But that it would be so low even I was not aware. Have to think a bit longer about what you wrote. That would not be tradeable at all - coin toss........ So what is it, that made some traders here believe that one can trade the similarity thing and the indicator successfully?
YES, basing only on the suggestions proposed in the thread, it comlpetely is.

The indicator and the similarity thing themselves do not give us superior advantages in trading (how many think), if they are used "as is", without additional knowledge about markets and proper money management. However they could be used as an additional element in somebody's trading system.

Best,
D.P.
Hunting High and Low
2 traders viewing now