The similarity system | Page 587 | 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

  • #11,721
  • Nov 6, 2014 10:26am Nov 6, 2014 10:26am
Quoting Kiads
{quote} Hi My Friend, here is the 'rule' 1. after all sub-sequences formula calculation give 'her' result, 2. look at the previous candle form, if bullish, then the result is for bearish direction, if bearish then the result is for bullish direction. 3. then the 'bomb' mark will appear if in bearish direction, current candle price should bellow previous candle high, and if in bullish direction, current price should above previous candle low. Meaning 'she' not always show up at candle open, but only if conditions met, because this calculation...
Thank you again, Kiads !

I think my indicator is close to yours
Need add some filter

http://i.imgur.com/vYkWYkX.png

I think i will learn how to draw a line so it will look like your indicator.

Good night everyone !

zelo
When you lose, I'm the loudest who laugh at your face )
  • #11,722
  • Nov 6, 2014 10:40am Nov 6, 2014 10:40am
The concept can be applied in almost in any sequence , i did applied with fractals and zz legs , after some days of testing good numbers came

There are plenty of trade can be done in any time frame,

i'll post some algo soon to predict next fractals and next zz leg


Regards
James
  • #11,723
  • Nov 6, 2014 11:37am Nov 6, 2014 11:37am
Quoting Zelo
{quote} Thank you again, Kiads ! I think my indicator is close to yours Need add some filter {image} I think i will learn how to draw a line so it will look like your indicator. Good night everyone ! zelo

Once again Congratulation My Friend, I know you can do it since the beginning you share it here, because you have that strong spirit to make it done...

You can learn how to create that continue updating horizontal lines (that actually using trend line) from supdem (supply demand) indicator that use continue updating rectangles...

hope it help and best regards
MTH
  • #11,724
  • Nov 6, 2014 11:39am Nov 6, 2014 11:39am
Quoting iexam
The concept can be applied in almost in any sequence , i did applied with fractals and zz legs , after some days of testing good numbers came There are plenty of trade can be done in any time frame, i'll post some algo soon to predict next fractals and next zz leg Regards James
Yes My Friend this SubSeq can be applied to anything related to time series data feed.
MTH
  • #11,725
  • Nov 6, 2014 11:56am Nov 6, 2014 11:56am
Quoting Kiads
{image} ok time to rest.... {image} MTH ps. there's sub-sequence of sub-sequence...
Hi Kiads,

Impressive results. For EURUSD D1, I'm getting probability of the sequence of 98.16%

Are you using something thats entire different for your sequence calculations now?
Focus, Patience, Determination & Order in chaos
  • #11,726
  • Nov 6, 2014 12:14pm Nov 6, 2014 12:14pm
Quoting Kiads
{quote} Yes, You know about that My Friend, that still 'rarely' discuss in forum. I'm member of this 3d chartist..lol. and I learn this similarity concept including sub-sequences, zones etc in relation with my 'space' learning.. that 'angle' in my previous post is just basic, but when you select correct 'h' you can find the 'center value' of every wave movements in 3 dimensional point of view just like this example bellow, the 'center value' is that 'Yellow Trend Lines'. {image} best regards MTH PS. even i'm that kind of noob, my posts...
If you are good at something, never do it for free--Joker
  • #11,727
  • Edited 12:51pm Nov 6, 2014 12:16pm | Edited 12:51pm
Quoting Zelo
{quote} Thank you again, Kiads ! I think my indicator is close to yours Need add some filter {image} I think i will learn how to draw a line so it will look like your indicator. Good night everyone ! zelo
Zelo my guy, killing the indi production! Keep up the good work brother! If possible please upload to thread once done, I and many others would be forever grateful!

Getting ready to try and short UJ when price is right. targeting blueline/lower TZ zone...

Attached Image (click to enlarge)
Click to Enlarge Name: Screen Shot 2014-11-06 at 11.16.42 AM.png Size: 76 KB


USDCHF also looks good for a drop.. 3 zones need to be cleared some day just dont know when 0.0

Attached Image (click to enlarge)
Click to Enlarge Name: Screen Shot 2014-11-06 at 11.21.32 AM.png Size: 84 KB


I like to think as the price moves up further away from these zones retail thinks price is rising and breaking out so they are buying but really market is positioning themselves short so the run up is the MM going short but price is rising because some one has to buy their sells(sonic PVA Analysis concept/exploit the market using volume PDF on sonic thread and this one).

Update: Nice Engulf on UJ near potential TZ. entered short with stops above PTZ
Attached Image (click to enlarge)
Click to Enlarge Name: Screen Shot 2014-11-06 at 11.48.51 AM.png Size: 72 KB



Quick question about an idea that I believe eurusdd mentioned but I cant remember. Wasnt it said that the closer h gets to being filled creating the zone the less likely the potential tz zone will be fulfilled correct? or is this wrong?

Pip
Life's Good & Why Wouldn't It Be? Im a Pirate After All!
  • #11,728
  • Edited Nov 7, 2014 3:11am Nov 6, 2014 1:17pm | Edited Nov 7, 2014 3:11am
Trying to understand how the subsequence theorem can be applied. Congrats to those who got it!

If I understand it correctly the sub-sequence theorem cannot give any clues about the next expected value (low or high or something else...) will be.
It only shows that in a sequence with n=2 (n^2 + 1 = 5 values) there will be a sequence n+1=3 of either increasing or decreasing values with 93.75% probability of fullfilling at least one of the conditions exposed by EURUSDD.
The 4 conditions can be found here (http://www.forexfactory.com/showthread.php?p=7831903#post7831903).

Then there exist only 2 situations where the condition is not met.
This is shown in this example (http://www.forexfactory.com/showthread.php?p=7832070#post7832070) provided by EURUSDD.
Therefore the 2 situations, in the example of using highs, are:
1. Higher high (HH), lower high(LH), HH, LH
2. LH, HH, LH, HH

These 2 situations can also be found by using the tree example by EURUSDD posted here (http://www.forexfactory.com/showthread.php?p=7833206#post7833206).

Since we cannot predict values we use the probability, as said by EURUSDD
Quoting Eurusdd
So for 10 consecutive highs if the theorem is not true for the first five, then the probability it is false for the second five also is 0.3%
The question is what does this mean, why could this give me an edge?
This means that after 5 bars that failed to meet the conditions, the next 5 bars are likely to meet at least 1 of the 4 conditions. But since there are 4 conditions, how can this give an edge about the next value?

I tested the idea that if for example we have a potential rare event (potential to become one of those 6.25%) HH, LH, HH, ?
we can expect that rare event not to happen, a bit like the PTZ and TZ game. But here I defined as potential where the last condition is still unclear,
meaning if the example sequence in formation HH, LH, HH, ? turns out as HH, LH, HH, LH we have the rare event, but HH, LH, HH, HH would meet one of the conditions and so becomes invalid.

Results have shown that on a sample of 5000 bars on 1H TF EURUSD with value being the highs:
- 9.74% of all bars don't meet the conditions of the theorem (rare event) (HH, LH, HH, LH or LH, HH, LH, HH)
- 21.3% of all bars are potential candidates for the rare event (HH, LH, HH, ? or LH, HH, LH, ?)
- 54% of potential candidates get invalidated, ie. meet one of the conditions (HH, LH, HH, HH or LH, HH, LH, LH)

Attached Image (click to enlarge)
Click to Enlarge Name: rare_potential.png Size: 44 KB

A triangle above the bar means that the last 5 bars (including the candle with the triangle painted above) didn't meet the conditions of the theorem.
An X below the bar means a potential rare event.
If X and a triangle are painted on the same bar, means potential rare event confirmed.
Only X means potential rare event invalidated by next bar's high (the bar with the X).

Of course, this is not that good edge. Just one of my ideas how one could use this theorem.

Now to the real question.
The indicators some of you created.
I see that the indicators draw arrows or bombs when specific conditions are met.
Are those conditions these rare events (6.25%)? Or something else?
By looking at the charts you posted, I can't recognize those.
I thought there were only 2 conditions that invalidate the theorem.

Then, why do you expect the price coming back to those prices that did not meet the conditions of the theorem?
Just from observation or is there some logical connection?

Sorry, I still don't understand. Hope someone can help
the 3rd wave is the crowd's wave
  • #11,729
  • Nov 6, 2014 1:19pm Nov 6, 2014 1:19pm
Quoting PiratePip
Quick question about an idea that I believe eurusdd mentioned but I cant remember. Wasnt it said that the closer h gets to being filled creating the zone the less likely the potential tz zone will be fulfilled correct? or is this wrong? Pip
EURUSDD has said that the further price moves away from a DS zone, the more violent the return to it will be.

I'm not sure this applies to TZ directly, as the data I have collected seems to indicate that the attractive/repellent forces of TZ's/PTZ's degrade logarithmically with time as well as with price movement away from the TZ/PTZ. You can see this by comparing probabilities of different h-values in K's tester. I have also confirmed this with live testing.
  • #11,730
  • Nov 6, 2014 1:51pm Nov 6, 2014 1:51pm
Quoting saver0
{quote} Hi Kiads, Impressive results. For EURUSD D1, I'm getting probability of the sequence of 98.16% Are you using something thats entire different for your sequence calculations now?
My Friend, I learn this sub-sequence theorem from our chief EURUSDD in this thread in relation with my price base chart own thread (shifted renko) so, try to create it to match my price base environment, and of course there is some 'adjustment' needed from the sub-sequence equation. And once again EURUSDD was 100% right this theorem work on anything and i believe it's work at best with price base chart..lol.

All the concepts, similarity, DNA, TZ and SubSeq, deeply related with price base chart.

This is example how to visualize probability in price base chart like my shifted renko. If you remember that EURUSDD mention about 16 branch with 2 failed that meant 14/16 = 87.5% and that 2 failed = 12.5% and 50/50 = 6.25% so 87.5 % + 6.25% = 93.75% and thanks God that my shifted renko default shift is 12.5% or half of it is 6.25% of 1/16 step.

So, with 16 point renko brick size, with default 6.25 shift every step will be 1 pip/point.

With constant OHLC input data flow per 1 point, the sub-sequence calculation will become ideal because 'she' receiving pre-filtered data feed. so, gaining 100% probability is not impossible.

Also every shift step in my shifted renko with same size brick can represent series of binary like if bullish form is 1 then bearish from is 0, and because there is no 'doji' in price base chart, then there is no noise for that dna series and the result will be more reliable.

For TZ, in price base chart will only consist of Top and Bottom TZ with very 'equal' form. You can see it when i apply Xixi TZ with statistic data, there will be almost always 50:50 ratios between Top and Bottom TZ, and that mean represent 'coin toss' in probability theorem. So, when there is 2 or more Top or Bottom TZ in a row, then the probability of next TZ become the opposite of previous TZ will be nearly 100%.

(Btw, Xixi my friend, i think your stat is in 'infinity loop' mode so it keep counting bigger and bigger percentage for the same TZ numbers)

My example here is shifted renko chart use ADX as momentum acceleration tools..

Attached Image (click to enlarge)
Click to Enlarge Name: GBPJPY M11 shifted renko 16point 6.25% shift with ADX def accel as example of SubSeq and TZ impl.png Size: 318 KB


hope it help and best regards
MTH
  • #11,731
  • Nov 6, 2014 1:52pm Nov 6, 2014 1:52pm
Xixi...

Is there any way to show the currently forming TZ on your indi (left h).

Thanks in advance.
  • #11,732
  • Nov 6, 2014 2:26pm Nov 6, 2014 2:26pm
Quoting juriFX
Trying to understand how the subsequence theorem can be applied. Congrats to those who got it! If I understand it correctly the sub-sequence theorem cannot give any clues about the next expected value (low or high or something else...) will be. It only shows that in a sequence with n=2 (n^2 + 1 = 5 values) there will be a sequence n+1=3 of either increasing or decreasing values with 93.75% probability of fullfilling at least one of the conditions exposed by EURUSDD. The 4 conditions can be found here (http://www.forexfactory.com/showthread.php?p=7831903#post7831903)....
wow. this post is deep ooh. nice!!! i have never posted something this cute::::
  • #11,733
  • Nov 6, 2014 2:38pm Nov 6, 2014 2:38pm
USDJPY got stopped on UJ. will be looking to short again as price pops over top TZ back down to blue line.
Attached Image (click to enlarge)
Click to Enlarge Name: Screen Shot 2014-11-06 at 1.36.48 PM.png Size: 39 KB
Life's Good & Why Wouldn't It Be? Im a Pirate After All!
  • #11,734
  • Nov 6, 2014 5:23pm Nov 6, 2014 5:23pm
Quoting Kiads
{quote}You can see it when i apply Xixi TZ with statistic data, there will be almost always 50:50 ratios between Top and Bottom TZ, and that mean represent 'coin toss' in probability theorem. So, when there is 2 or more Top or Bottom TZ in a row, then the probability of next TZ become the opposite of previous TZ will be nearly 100%.
Brilliant! always opening my mind kiads. again thank you!
Life's Good & Why Wouldn't It Be? Im a Pirate After All!
  • #11,735
  • Nov 6, 2014 5:38pm Nov 6, 2014 5:38pm
Quoting Barbaresco
{quote} I am trying M1 in reverse, and moving up to M15 for the regular version. Another thought regarding the reverse version is removing the confirmed TZ as a condition and simply working from all PTZ's that have broken out by Y. Will report stats here as I have them.

Updating M1 data:

M1 in reverse (data from M1 running regular was given on pg 583):
Betting on 2 confirmed TZ's in a row based on a move Y pts away from the PTZ - also does not work - worse than trading it forward - didn't even compile the stats.

Conclusion is that in M1, after a confirmed TZ and a new PTZ at the same level, a 5 pt move away from that 2nd order PTZ can't be traded back to the PTZ profitably, nor can it be traded away from the PTZ profitably. Win ratio running forward (trade back to the PTZ so as not to have 2 TZ in a row) was approx. 42%.

Observations: Without a stop loss (time or point based), most trades (in either direction) do end up winning, but the risk so far outweighs the return that it would not be profitable for very long. It seems that once price moves far enough away from an M1 PTZ to be traded back to the PTZ, the probabilities are no longer in our favor such that it is similar to trading a random price the same way, instead of starting with a PTZ. It also seems that a 2nd order TZ is no less likely than a 1st order TZ, or, at least not enough to give a profitable setup.

Hypothesis: After any PTZ, there is a relatively small window in time and price where probabilities will support a trade back to that PTZ. Ex. if price moves less than X, within Y time, a trade may be taken back to the PTZ. The SL would be PTZ+/-X depending on direction, and the timed SL would be the full rest of h. This will not be possible to test in M1.

Next Results:
M1, h=32, enter in direction of price on any PTZ, SL if PTZ in opposite direction, TP of 5. This gave approx. 60% W:L, but the SL's were 2-4 times the TP's. Next would be to try an SL based on a confirmed zone, to try to trade the approx. 15% PTZ's against the approx. 3% TZ's for 80% W:L.

Discouraged.
  • #11,736
  • Nov 6, 2014 7:43pm Nov 6, 2014 7:43pm
I can't believe how the level of discussion in this thread advanced in the last couple of weeks. Amazing!

Keep up the great work guys. I see some really interesting ideas that are worth developing out there. I hope I'll be able to go through all of it during the weekend.

For those looking for some trade recommendations, keep your eyes on EURNZD & NZDCHF, those are going to be my next 2 big trades.
I think by now I don't really need to give much more details right? You should be able to spot the direction I'm going to trade towards.

I'll be open for discussion.

Best,
Jay
  • #11,737
  • Nov 6, 2014 7:52pm Nov 6, 2014 7:52pm
Quoting FX-Jay
I can't believe how the level of discussion in this thread advanced in the last couple of weeks. Amazing! Keep up the great work guys. I see some really interesting ideas that are worth developing out there. I hope I'll be able to go through all of it during the weekend. For those looking for some trade recommendations, keep your eyes on EURNZD & NZDCHF, those are going to be my next 2 big trades. I think by now I don't really need to give much more details right? You should be able to spot the direction I'm going to trade towards. I'll be open...

hy, in the post that i have read of eurusd and yours eurusd said that you win no matter where price goes i really don't understand how you will win if the price doesn't return a little bit to your target http://www.forexfactory.com/showthread.php?p=7680220#post7680220 post 5884
  • #11,738
  • Nov 6, 2014 7:57pm Nov 6, 2014 7:57pm
Quoting Barbaresco
{quote} Updating M1 data: M1 in reverse (data from M1 running regular was given on pg 583): Betting on 2 confirmed TZ's in a row based on a move Y pts away from the PTZ - also does not work - worse than trading it forward - didn't even compile the stats. Conclusion is that in M1, after a confirmed TZ and a new PTZ at the same level, a 5 pt move away from that 2nd order PTZ can't be traded back to the PTZ profitably, nor can it be traded away from the PTZ profitably. Win ratio running forward (trade back to the PTZ so as not to have 2 TZ in a row)...
I've been trying to trade back as you described with various stoploss and "K" estimations for about 3 days now on demo... with no luck, glad to know atleast I'm not alone on this boat
A Positive Mind, and Positive Spirit lead to True Happiness
  • #11,739
  • Nov 6, 2014 8:41pm Nov 6, 2014 8:41pm
Quoting FX-Jay
I can't believe how the level of discussion in this thread advanced in the last couple of weeks. Amazing! Keep up the great work guys. I see some really interesting ideas that are worth developing out there. I hope I'll be able to go through all of it during the weekend. For those looking for some trade recommendations, keep your eyes on EURNZD & NZDCHF, those are going to be my next 2 big trades. I think by now I don't really need to give much more details right? You should be able to spot the direction I'm going to trade towards. I'll be open...
jay, your post (http://www.forexfactory.com/showthread.php?p=7842544#post7842544) gives some hints?
Humble & Kalcker CLO2 = Covid killer
  • #11,740
  • Nov 6, 2014 9:22pm Nov 6, 2014 9:22pm
Quoting Rparm
{quote} I've been trying to trade back as you described with various stoploss and "K" estimations for about 3 days now on demo... with no luck, glad to know atleast I'm not alone on this boat
You have to have a system that can trade maybe break even and apply TZ as a filter or direction indicator. I have a momentum trading EA that I was never able to get working well but is now at 50% a year on a daily chart. Hope to adapt it to hourly / 15 minute entries and improve the entries (maybe with COBB) to improve on these modest but still possibly life changing results. Remember, Warren Buffet'a career is based on 20% annual returns.
For example if you try to take advantage of the example that FX-Jay posted with 97% chance to return to TZ, you still an EA that can pick a good entry and not get stopped out if you want to backtest that automatically. That is my next project.
0 traders viewing now