The similarity system | Page 537 | 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

  • #10,721
  • Edited 4:19am Oct 22, 2014 12:11am | Edited 4:19am
  •  Neio
  • Joined Oct 2012 | Status: Can't think outside?Enlarge the box | 688 Posts
Quoting MetaCoder
{quote} If I understood eurusdd's explanation on the stoch mismatch, it's only to find price levels the market is likely to return to. Direction/strength would be gleaned from other indicators.

Stoch at 0/100 levels is really the same as a potential TZ. It is going into areas where price has not been in h number of bars.

A stoch configured with 300,1,1 is essentially the same as configuring the TZ indicator with a h of 300.

Stoch obviously does not show mid-bar transients

Stoch is showing the highest/lowest close within x number of bars. The TZ indicator shows places were price has not been in x number of bars.


So when price hits 0/100 levels stoch cannot follow so you can expect at some time price will need to come back down/up to mirror price again. As if you have the right stoch settings the stoch and price will like to mirror each other most of the time.
An expert is a man who has made all the mistakes which can be made
  • #10,722
  • Oct 22, 2014 12:53am Oct 22, 2014 12:53am
The way you choose to display the realization of a stochastic process is important. For most of our chart, the time component is what is used to display changes in price. That s why we see so much randomness. Now, if you replace the time component with a space component, you will get a renko chart, which in most cases show less randomness. There are other ways to display your data in order to see exactly what price is doing!
As far as the time component is concerned, recurrent prices rule and I will prove that rigorously soon. Little technicality....
  • #10,723
  • Oct 22, 2014 1:40am Oct 22, 2014 1:40am
Quoting Eurusdd
That s why we see so much randomness..
thanks to you I no longer see any randomness what so ever. now i only see the algo used by the powers that be
  • #10,724
  • Oct 22, 2014 1:48am Oct 22, 2014 1:48am
Quoting Eurusdd
The way you choose to display the realization of a stochastic process is important. For most of our chart, the time component is what is used to display changes in price. That s why we see so much randomness. Now, if you replace the time component with a space component, you will get a renko chart, which in most cases show less randomness. There are other ways to display your data in order to see exactly what price is doing! As far as the time component is concerned, recurrent prices rule and I will prove that rigorously soon. Little technicality....
Thank You Sir , I'm struggling to learn in that field for all this years.. , if you don't mind would you please check this picture and give me some advice..
best regard
MTH

GBPUSD 8 point brick, shift 1 point increment renko chart with stochastic 300,1,1 as visual movement shifting acceleration

Attached Image (click to enlarge)
Click to Enlarge Name: Shifted Flip Renko with Stoch acceleration and TZ on GBPUSD 8 point brick.png Size: 229 KB
  • #10,725
  • Oct 22, 2014 2:24am Oct 22, 2014 2:24am
Quoting Kiads
{quote} Thank You Sir , I'm struggling to learn in that field for all this years.. , if you don't mind would you please check this picture and give me some advice.. best regard MTH GBPUSD 8 point brick, shift 1 point increment renko chart with stochastic 300,1,1 as visual movement shifting acceleration {image}
BIT of OUT of TOPIC

really impressive to see after all the achievements you did via mth renko system, still trying to learn new things, thats why you are great........

i am proud to be a friend of you....................
  • #10,726
  • Oct 22, 2014 2:28am Oct 22, 2014 2:28am
Quoting Eurusdd
The way you choose to display the realization of a stochastic process is important. For most of our chart, the time component is what is used to display changes in price. That s why we see so much randomness. Now, if you replace the time component with a space component, you will get a renko chart, which in most cases show less randomness. There are other ways to display your data in order to see exactly what price is doing! As far as the time component is concerned, recurrent prices rule and I will prove that rigorously soon. Little technicality....
Hi,Eurusdd，I can't send you message cos setting restricted me，could you please Subscribed me？
Just want to show you my somthing understood of your laws（100% surely win rate！）
  • #10,727
  • Oct 22, 2014 2:37am Oct 22, 2014 2:37am
Quoting Neio
{quote} Stoch at 0/100 levels is really the same as a potential TZ. It is going into areas where price has not been in h number of bars. A stoch configured with 300,1,1 is essentially the same as configuring the TZ indicator with a h of 300. Stoch obviously does not show mid-bar transients Stoch is showing the highest/lowest close within x number of bars. The TZ indicator shows places were price has not been in x number of bars. So when price hits 0/100 levels stoch cannot follow so you can expect at some time price will need to come back down/up...


The same thing is with painting the leg of the Zigzag (http://www.forexfactory.com/showthread.php?t=435149), and that (http://www.forexfactory.com/showthread.php?p=6802415#post6802415) is also why in the RecurrenceStatistic indicator you have that ugly Donchian thing.

Check here (http://www.forexfactory.com/showthread.php?p=7812956#post7812956) for the approximate determination of the "correct" settings of the Stochastic oscillator. Best to read the whole thread (another nugget by Eurusdd).

Cheers,
k
  • #10,728
  • Oct 22, 2014 2:51am Oct 22, 2014 2:51am
  •  Neio
  • Joined Oct 2012 | Status: Can't think outside?Enlarge the box | 688 Posts
Quoting kprsa
{quote} The same thing is with painting the leg of the Zigzag (http://www.forexfactory.com/showthread.php?t=435149), and that (http://www.forexfactory.com/showthread.php?p=6802415#post6802415) is also why in the RecurrenceStatistic indicator you have that ugly Donchian thing. Check here (http://www.forexfactory.com/showthread.php?p=7812956#post7812956) for the approximate determination of the "correct" settings of the Stochastic oscillator. Best to read the whole thread (another nugget by Eurusdd). Cheers,...

Yes they are all essentially the same thing. I don't think a lot of people realise that. I did not at first.

The problem with finding the right MA was that while ranging you can easily find an MA that stays flat......the problem where everything fails is in strong trends. Common theme there.

Regards

Neio
An expert is a man who has made all the mistakes which can be made
  • #10,729
  • Oct 22, 2014 3:08am Oct 22, 2014 3:08am
Quoting scalper.exp
{quote} BIT of OUT of TOPIC really impressive to see after all the achievements you did via mth renko system, still trying to learn new things, thats why you are great........ i am proud to be a friend of you....................
Thank You for being my friend too...
I'll keep learning for the rest of my life because my signature 'force to do that' on daily basis.. lol
best regard
MTH
  • #10,730
  • Oct 22, 2014 3:39am Oct 22, 2014 3:39am
Quoting kprsa
{quote} The same thing is with painting the leg of the Zigzag (http://www.forexfactory.com/showthread.php?t=435149), and that (http://www.forexfactory.com/showthread.php?p=6802415#post6802415) is also why in the RecurrenceStatistic indicator you have that ugly Donchian thing. Check here (http://www.forexfactory.com/showthread.php?p=7812956#post7812956) for the approximate determination of the "correct" settings of the Stochastic oscillator. Best to read the whole thread (another nugget...
Quoting Neio
{quote} What I observed in the above settings is that mismatches often occur as follows (it may also apply to the Stoch mismatches across two tfs) 1. Stoch has a limited range so mismatches often occur when Stoch hits the 0/100 levels. This often happens during trending moves. 2. Also noticed that you often get mismatches around the 50 level on some stoch configurations: 3. When there is a high volume of tick moves The above may also apply to comparing Stochs across different timeframes but as I mentioned we moved on quite quickly from that...
Quoting Neio
{quote} Stoch at 0/100 levels is really the same as a potential TZ. It is going into areas where price has not been in h number of bars. A stoch configured with 300,1,1 is essentially the same as configuring the TZ indicator with a h of 300. Stoch obviously does not show mid-bar transients Stoch is showing the highest/lowest close within x number of bars. The TZ indicator shows places were price has not been in x number of bars. So when price hits 0/100 levels stoch cannot follow so you can expect at some time price will need to come back...
Excellent posts.
Even I completely missed the significance of the Donchian!
A pip is only worth it if you know how much you risked to earn it
  • #10,731
  • Oct 22, 2014 4:54am Oct 22, 2014 4:54am
Quoting Neio
{quote} Yes they are all essentially the same thing. I don't think a lot of people realise that.
This observation, among other things (like the Stoch/ZZ similarity strategy where the indicators say almost the same thing), gives simplifications through cross-referencing of the indicators.

One can, for example, consider using some of the stats that we have on the TZs on the ZigZag.

k
  • #10,732
  • Oct 22, 2014 5:03am Oct 22, 2014 5:03am
Quoting kprsa
{quote} This observation, among other things (like the Stoch/ZZ similarity strategy where the indicators say almost the same thing), gives simplifications through cross-referencing of the indicators. One can, for example, consider using some of the stats that we have on the TZs on the ZigZag. k
Difficult is to use stoch and moving average on strong trends. "eurusdd" had said stoch should be equal in every way with EMA1, and the period of stoch is calculated by XEMA but strong trend ....
  • #10,733
  • Oct 22, 2014 5:15am Oct 22, 2014 5:15am
Quoting burnssss
{quote} Difficult is to use stoch and moving average on strong trends. "eurusdd" had said stoch should be equal in every way with EMA1, and the period of stoch is calculated by XEMA but strong trend ....
No matter what the particular strategy is (and I am not discussing any strategy in these posts) it is wise to know what the indicators you have on the chart are actually doing, how they are interconnected and how they relate to the price structure.

It may also help people to avoid mistakes (for example the mistake of using the fixed default parameters) in the common indicators.

Cheers,
k
  • #10,734
  • Oct 22, 2014 5:23am Oct 22, 2014 5:23am
Quoting kprsa
One of the problems often encountered here is exactly the opposite: that the market is not sufficiently random, it does not behave like a Markov process, it has memory and therefore the events on two candles are not independent. Hence the probability of having two TZs on subsequent candles is not ignorable, particularly when strong breakouts occur.
Nice post, k. Does your indicator give the stats for consecutive TZs? I've never been able to use your indicator because my broker doesn't support MetaTrader, but I did code one of my own for NinjaTrader. According to my stats (over about 10,000 bars), the probability of the next bar having a TZ after a TZ bar is on average 17% for EUR.USD (averaged for timeframes M1 to H4, h optimised for 3% TZ bars). Not ignorable, as you say. Though they only occur on mid-bar TZs, which are pretty rare anyway (about 1% of all bars). Does this agree with your stats? Thanks.
You reap what you sow.
  • #10,735
  • Oct 22, 2014 5:51am Oct 22, 2014 5:51am
Quoting NorthTrader
{quote}Does your indicator give the stats for consecutive TZs?
No, it doesn't. The indicator is slow as molasses in January as it is...

Quoting NorthTrader
{quote}I've never been able to use your indicator because my broker doesn't support MetaTrader, but I did code one of my own for NinjaTrader.
If you have invested several hours of your time to code an indicator, you could have also invested five minutes of your time to open a demo account at any of the brokers that do support Metatrader. I mean, I am sorry that your broker of choice doesn't support MT4, but I wouldn't exactly let that stop me (if I were you).

k
  • #10,736
  • Oct 22, 2014 6:34am Oct 22, 2014 6:34am
Quoting NorthTrader
{quote} Nice post, k. Does your indicator give the stats for consecutive TZs? I've never been able to use your indicator because my broker doesn't support MetaTrader, but I did code one of my own for NinjaTrader. According to my stats (over about 10,000 bars), the probability of the next bar having a TZ after a TZ bar is on average 17% for EUR.USD (averaged for timeframes M1 to H4, h optimised for 3% TZ bars). Not ignorable, as you say. Though they only occur on mid-bar TZs, which are pretty rare anyway (about 1% of all bars). Does this agree with...
Are you sure you mean TZ bars and not potential TZ bars? In my stats, I found the probability for 2 tz bars to occur consecutively to be very rare. I came up with something like .260% for h=30 and well over 10000 bars as well, although I only did it on one time frame.

Quoting Eurusdd
The way you choose to display the realization of a stochastic process is important. For most of our chart, the time component is what is used to display changes in price. That s why we see so much randomness. Now, if you replace the time component with a space component, you will get a renko chart, which in most cases show less randomness. There are other ways to display your data in order to see exactly what price is doing! As far as the time component is concerned, recurrent prices rule and I will prove that rigorously soon. Little technicality....
A third type of chart?? I can't wait to hear more about it.
"Mrrraggglhlhghghlgh"
  • #10,737
  • Oct 22, 2014 6:51am Oct 22, 2014 6:51am
Quoting LiquidGenius
{quote} In my stats, I found the probability for 2 tz bars to occur consecutively to be very rare. I came up with something like .260% for h=30 and well over 10000 bars as well, although I only did it on one time frame.
I think in order to compare your numbers with NorthTrader's, you need to divide your number (0.0026) by the probability of having a TZ on one bar. For example if that probability is 2%=0.02, the probability of having a second TZ after the first one is 0.0026/0.02=0.13=13%.

k
  • #10,738
  • Oct 22, 2014 6:59am Oct 22, 2014 6:59am
Quoting kprsa
If you have invested several hours of your time to code an indicator, you could have also invested five minutes of your time to open a demo account at any of the brokers that do support Metatrader. I mean, I am sorry that your broker of choice doesn't support MT4, but I wouldn't exactly let that stop me (if I were you). k
Yeah, good point. If it only takes 5 mins then maybe I should have! But I'm a software engineer by trade so it wasn't too bad.
You reap what you sow.
  • #10,739
  • Oct 22, 2014 7:01am Oct 22, 2014 7:01am
Quoting kprsa
{quote} I think in order to compare your numbers with NorthTrader's, you need to divide your number (0.0026) by the probability of having a TZ on one bar. For example if that probability is 2%=0.02, the probability of having a second TZ after the first one is 0.0026/0.02=0.13=13%. k
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.
You reap what you sow.
  • #10,740
  • Oct 22, 2014 7:06am Oct 22, 2014 7:06am
Quoting kprsa
{quote} For example if that probability is 2%=0.02, the probability of having a second TZ after the first one is 0.0026/0.02=0.13=13%. k
Sure?
When the probability would be even lower like 1% then the probability of having a second TZ would be 26% ?
6 traders viewing now