The similarity system | Page 538 | 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

  • #10,741
  • Oct 22, 2014 7:13am Oct 22, 2014 7:13am
Quoting hilavoku
{quote} Sure? When the probability would be even lower like 1% then the probability of having a second TZ would be 26% ?
You would have MULTIPLY those numbers. So 0.02*0.02 = 0.0004 (=0.04%)

What I dont quite understand yet:
In order to have two consecutive TZ they MUST be mid-bar TZs, otherwise the 2nd bar would invalidate the first TZ and therefore it wouldnt be 2 consecutive TZs anymore, right?
  • #10,742
  • Oct 22, 2014 7:14am Oct 22, 2014 7:14am
Quoting fsm
{quote} You would have MULTIPLY those numbers. So 0.02*0.02 = 0.0004 (=0.04%) ?
This is ~what I wanted to say.
0.0026*0.02=0.0052% ?
  • #10,743
  • Oct 22, 2014 7:20am Oct 22, 2014 7:20am
Quoting NorthTrader
{quote} Another good point! I was wondering why our numbers were so different. Yes, my stats are for the next bar after a mid-bar TZ being another mid-bar TZ. A quick visual scan of the chart seems to confirm it, but I didn't have time to check the whole data set visually.
Ah. Then yes your stats are roughly in the same ball park I think +\- 5% given the 2 h values I checked.

edit:
clarification: As I understand it, North Traders number would be correct, or roughly correct depending on TF and h value if he's calculating the probability that the next bar is also transient GIVEN(or assuming) that the current bar in question is also mid bar transient. This is of course because the set of mid bar TZs is already small.
However, if the question is simply what are the chances that any 2 bars are consecutively mid bar transient, then the odds are small and roughly close to what I presented earlier

Quoting fsm
{quote} What I dont quite understand yet: In order to have two consecutive TZ they MUST be mid-bar TZs, otherwise the 2nd bar would invalidate the first TZ and therefore it wouldnt be 2 consecutive TZs anymore, right?
Not completely but close. You could have the additional possibility of a mid bar TZ followed by a regular (top/bottom) TZ
"Mrrraggglhlhghghlgh"
  • #10,744
  • Oct 22, 2014 7:23am Oct 22, 2014 7:23am
Quoting hilavoku
{quote} Sure? When the probability would be even lower like 1% then the probability of having a second TZ would be 26% ?
I think so, given LG's number of 0.0026 holds.

We're talking here about the conditional probability (http://en.wikipedia.org/wiki/Conditional_probability#Kolmogorov_definition).

P(TZ on the candle i+1 occurs if the TZ occurs at the candle i)=P(TZ on candle i and TZ on candle i+1)/P(TZ on the candle i)

To my understanding:

LG measures P(TZ on candle i and TZ on candle i+1), whereas
NT measures P(TZ on the candle i+1 occurs if the TZ occurs at the candle i).

Cheers,
k
  • #10,745
  • Oct 22, 2014 7:26am Oct 22, 2014 7:26am
Head Hurts!

Attached Image


LOL
A pip is only worth it if you know how much you risked to earn it
  • #10,746
  • Oct 22, 2014 7:38am Oct 22, 2014 7:38am
Quoting vlady1974
Head Hurts! {image} LOL

Attached Image
  • #10,747
  • Edited 8:35am Oct 22, 2014 7:44am | Edited 8:35am
Sorry kprsa,

you are right! At first sight I thought they were common unconditional probabilities. Next time I will take a closer look before posting.

PS: So the first TZ is like B and the second one is A which is the restricted to the sample space B?!
  • #10,748
  • Oct 22, 2014 7:50am Oct 22, 2014 7:50am
Quoting hilavoku
At first sight I thought they were common unconditional probabilities.
No problem, it is best to ask me to clarify as you did, I try to avoid too many explanations when posting.

Vlady and Kiads: Sorry for the nerdspeak. :nerd:

k
  • #10,749
  • Oct 22, 2014 8:14am Oct 22, 2014 8:14am
Quoting kprsa
{quote} No problem, it is best to ask me to clarify as you did, I try to avoid too many explanations when posting. Vlady and Kiads: Sorry for the nerdspeak. :nerd: k
Hi My Friend kprsa, sorry for my bad English, i wanna ask you, is your indicator base on standard mt4 zigzag ?
if it yes, then the latest MT4 build 711 zigzag is not same with the old standard zigzag from previous build, and hv some unrecognized error (100 % error free when recompile). This is found out by Mladen from TSD.

Thanks in advance
Best regard
MTH
  • #10,750
  • Oct 22, 2014 8:16am Oct 22, 2014 8:16am
Quoting kprsa
{quote} No problem, it is best to ask me to clarify as you did, I try to avoid too many explanations when posting. Vlady and Kiads: Sorry for the nerdspeak. :nerd: k
Nah man, just joking around.
You know what I did after visiting your link.... visited Amazon.co.uk for these babies (http://goo.gl/uFrOCD)

A pip is only worth it if you know how much you risked to earn it
  • #10,751
  • Oct 22, 2014 8:19am Oct 22, 2014 8:19am
I took a long trade on EURUSD at 1.2737, expecting it to reach 1.2750. Turns out it took a sharp drop. Is the trade still valid?
less than
  • #10,752
  • Edited 8:42am Oct 22, 2014 8:22am | Edited 8:42am
"Price is 97% recurrent, 3% transient"


I have a few suggestions, especially for anyone new. Remove ALL indicators. While I am amazed at all the coding, other ideas being sparked etc. the most important thing period in this thread is the quote I opened this post with.


Open a 15 min chart. Choose an H value of 6. Look at every bar for a few days back. See if you can draw a line or box (it becomes a box it price moves more than a pip in 'free' untouched space) from the current bar you are looking at back six bars without touching another bar. When you have done that, look to the right of the current bar / mid-bar. If looking at historical prices see how many times that line or box you drew was penetrated and fully cancelled out by price. Now watch as this happens in real time. This is ALL you need in my opinion to create a profitable edge. Again, it's so powerful and so simple:


"Price is 97% recurrent, 3% transient"
Attached Image (click to enlarge)
Click to Enlarge Name: TZ.jpg Size: 66 KB
  • #10,753
  • Oct 22, 2014 8:26am Oct 22, 2014 8:26am
Quoting Kiads
{quote} Hi My Friend kprsa, sorry for my bad English, i wanna ask you, is your indicator base on standard mt4 zigzag ? if it yes, then the latest MT4 build 711 zigzag is not same with the old standard zigzag from previous build, and hv some unrecognized error (100 % error free when recompile). This is found out by Mladen from TSD. Thanks in advance Best regard MTH
No worries, I understand you loud and clear.
No, my indicator does not use Zigzag directly (but is in principle connected with it). Thanks for letting me know.
Mladen is a superhero.

Cheers,
k
  • #10,754
  • Oct 22, 2014 8:28am Oct 22, 2014 8:28am
Quoting vlady1974
{quote} Nah man, just joking around. You know what I did after visiting your link.... visited Amazon.co.uk for these babies (http://goo.gl/uFrOCD)

At university I liked very much the Schaum's series, they were very cheap and I liked the approach. Check (http://www.amazon.co.uk/Schaums-Outline-Probability-Statistics-4th/dp/007179557X/ref=sr_1_1?s=books&ie=UTF8&qid=1413984116&sr=1-1&keywords=schaums+outlines+probability) those (http://www.amazon.co.uk/Schaums-Outline-Probability-Variables-Processes/dp/0071822984/ref=sr_1_2?s=books&ie=UTF8&qid=1413984116&sr=1-2&keywords=schaums+outlines+probability) as well. :nerd:
k
  • #10,755
  • Oct 22, 2014 9:04am Oct 22, 2014 9:04am
Quoting kprsa
{quote} At university I liked very much the Schaum's series, they were very cheap and I liked the approach. Check (http://www.amazon.co.uk/Schaums-Outline-Probability-Statistics-4th/dp/007179557X/ref=sr_1_1?s=books&ie=UTF8&qid=1413984116&sr=1-1&keywords=schaums+outlines+probability) those (http://www.amazon.co.uk/Schaums-Outline-Probability-Variables-Processes/dp/0071822984/ref=sr_1_2?s=books&ie=UTF8&qid=1413984116&sr=1-2&keywords=schaums+outlines+probability) as well. :nerd: k
Copy that. Reminds me of many hours with headache and throwing books into corners and screaming at them. :nerd:
  • #10,756
  • Oct 22, 2014 9:09am Oct 22, 2014 9:09am
Quoting kprsa
{quote} At university I liked very much the Schaum's series, they were very cheap and I liked the approach. Check (http://www.amazon.co.uk/Schaums-Outline-Probability-Statistics-4th/dp/007179557X/ref=sr_1_1?s=books&ie=UTF8&qid=1413984116&sr=1-1&keywords=schaums+outlines+probability) those (http://www.amazon.co.uk/Schaums-Outline-Probability-Variables-Processes/dp/0071822984/ref=sr_1_2?s=books&ie=UTF8&qid=1413984116&sr=1-2&keywords=schaums+outlines+probability) as well. :nerd: k
Which one to start with first mate? Probability & Statistics or Probability, Random Variables, Random Processes?
Seems to me that since our topic is not "random" it would be the 1st one. Yes?
A pip is only worth it if you know how much you risked to earn it
  • #10,757
  • Oct 22, 2014 9:18am Oct 22, 2014 9:18am
Quoting vlady1974
{quote} Which one to start with first mate? Probability & Statistics or Probability, Random Variables, Random Processes? Seems to me that since our topic is not "random" it would be the 1st one. Yes?
Yes, the first one is more than sufficient. The second one repeats a lot but has Markov processes etc inside.

k
  • #10,758
  • Oct 22, 2014 9:22am Oct 22, 2014 9:22am
Quoting vlady1974
Head Hurts! {image} LOL
Head hurts? This is the fun stuff!
Although I only have a highschool level understanding of statistics heh.
"Mrrraggglhlhghghlgh"
  • #10,759
  • Edited 9:38am Oct 22, 2014 9:24am | Edited 9:38am
Hi,guys,I have some problem..
So we must to find a confirmed top,because we don't know if price will keep going up?
thx
Quoting Blackeagle
{quote} I know probability is on your side based on Eurusdd's explanations but what if the price DOES NOT go back to transient zone? Therefore I understand that if my zone turns out to be h-recurrent, my sell position will win, but I don't understand how my sell position wins if my zone turns out to be h-transient and price DOES NOT want to make it recurrent this time and price DOES NOT want to make it recurrent this time and keep going up? I believe this is where SL plays its part but then how about Eurusdd's win-win condition? Anyone? Kins Regards, Blackeagle
This is Eurusdd say
Quoting Eurusdd
Let us supposed we know the yellow area but the market has not develop the blue area yet. We can then assume that when the region B is matured, it will be transient. So, we declare the area A as recurrent. We then move ahead to find sell position above B. Now, if the market pulls back into B, we win. If B turns out to be transient, our assumption of A as recurrent will give our sell a profit. Both ways, we will win!!!
http://i.imgur.com/mDehfGk.png​
  • #10,760
  • Oct 22, 2014 9:27am Oct 22, 2014 9:27am
Quoting LiquidGenius
{quote} Head hurts? This is the fun stuff! Although I only have a highschool level understanding of statistics heh.
Mate... the last time I saw this stuff was 1994!
But I've bought the books recommended by K and I'm gonna sink into it. I'm not one to back away from a challenge.

This gives me inspiration:

Quoting iDoubleStoch
{quote} Through raw data Research, Development, Engineering, Testing, Optimization, careful Implementation and the reliance on empirical findings for all trading decisions both strategic and tactical. And, by trading within a Business and Revenue Model that sets historically appropriate expectations for revenue growth and draw-down such that fear, greed and emotion never impacts business operations, functions and protocols. You fear the unknown. Research tells you everything you need to know about what is probable. Everything in trading begins...
A pip is only worth it if you know how much you risked to earn it
15 traders viewing now