Statistical Analysis of TZs | Page 7 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Voodoo Science-Statistical Analysis 45 replies

Statistical analysis of trading (PA, MM and more...) 3 replies

Best tools for statistical analysis of tick data and news events? 8 replies

Statistical Analysis of Trading History 9 replies

Statistical analysis fallacy 33 replies

Attachments: Statistical Analysis of TZs
Exit Attachments

Statistical Analysis of TZs

  • #121
  • Jun 9, 2015 6:41pm Jun 9, 2015 6:41pm
Quoting rrrem
{quote} Hello NT, first, great and invaluable work. What i understand from initial proposition, TZ concept should work with same probability on all TF (M1 or H4). Your comment?
Thanks rrrem. We initially found that H4 (M240) was the best timeframe for the First Opposite system, and later tests I ran seemed to confirm this. I haven't posted the latest results for other timeframes yet, but will do soon.

It may be conincidence that M240 yields the best results, or it may be because the parameters (h values and so on) were optimised for that timeframe. You'd have thought that all timeframes would yield very similar results over large data samples, due to the fractal nature of the market.

I'll post more about this later.
You reap what you sow.
  • #122
  • Edited Jun 10, 2015 1:23am Jun 9, 2015 10:36pm | Edited Jun 10, 2015 1:23am
Quoting pedma
{quote} I use h that is the smallest one to get 3%. This can be our own choice or based on Jay's trade examples. {quote}
Hi NT, sorry for posting this idea below. It has no relationship to statistical analysis (yet), but it has relationship to TZ (need some time to create statistic on this). I like scalping, so trying to trade on TF M15 using TZ. One of my favorite h is 8. It is in testing phase, so any idea to make it better is welcome. Image can be found here: http://www.forexfactory.com/showthre...69#post8315469
Good friends are more valuable than gold ...
  • #123
  • Jun 10, 2015 1:24am Jun 10, 2015 1:24am
Quoting kprsa
{quote} In the RecurrenceStatistic code there is no mention of the opposite TZs. While there may have been people adding other stuff to it, the original code I wrote dealt only with single TZs. Cheers, k edit: the mql code is here: {file}
Thanks kprsa

I knew if I threw something out there from one of what I like to call the Originals( the Founders if you are a Trekkie) one of you would show up and clear up what I was missing.
If you are good at something, never do it for free--Joker
  • #124
  • Jun 10, 2015 1:25am Jun 10, 2015 1:25am
Quoting NorthTrader
{quote} Yes, all the indicators I post are for NinjaTrader, sorry.
Thanks, your ideas and concepts are still very much appreciated.
If you are good at something, never do it for free--Joker
  • #125
  • Jun 10, 2015 4:53am Jun 10, 2015 4:53am
Hi NT,
Last week i have output window at the time i test your code, but today there is no window. May i know what the problem is, and how to solve it ?
Thanks in advance.
Good friends are more valuable than gold ...
  • #126
  • Jun 10, 2015 6:55am Jun 10, 2015 6:55am
Quoting pedma
Hi NT, Last week i have output window at the time i test your code, but today there is no window. May i know what the problem is, and how to solve it ? Thanks in advance.
Not sure pedma, but maybe it was closed by mistake? If you select "Tools > Output Window ..." from the Control Center menu then it should come back.
You reap what you sow.
  • #127
  • Jun 10, 2015 7:45am Jun 10, 2015 7:45am
Quoting NorthTrader
{quote} Not sure pedma, but maybe it was closed by mistake? If you select "Tools > Output Window ..." from the Control Center menu then it should come back.
Thanks a lot NT. Yes, i remember now that i have closed it. It is ok now ... thank you.
Good friends are more valuable than gold ...
  • #128
  • Jun 12, 2015 11:34pm Jun 12, 2015 11:34pm
Hi NorthTrader,

This is my first post here, but I have been following this thread from the beginning.

Quoting NorthTrader
{quote} But if you enter and place stops dynamically, at different levels each time, then you will win (according to the analysis). And if you use even more discretion - adjusting entries, stops and targets according to the current market environment, or skipping trades entirely if something looks wrong - then you will win even more.
I am afraid that I cannot understand your strategy completely (because of my poor english skill), but I agree to the above statement.
I have also experienced that the dynamic entry&exit can produce profitable backtest.


But then,

Quoting NorthTrader
{quote} But if you know how to trade in the first place, then why bother trading TZs and not some other system?
I also thought about that and currently I am not working on TZ strategy.
But I hope your analysis will go well and look forward to hearing good news from you.

Thank you,
  • #129
  • Jun 13, 2015 3:28am Jun 13, 2015 3:28am
Quoting jimsterk
But I hope your analysis will go well and look forward to hearing good news from you. Thank you,
Hey Jim,

O-hisashi buri. You've been very quiet recently! Thanks for your kind words.

By the way, the system is very simple. You just trade towards the first opposite PTZ (the "Target TZ") that appears after a PTZ/CTZ (the "Base TZ"), within a range of h values (30 to 120 works best). So the Base TZ can be the same base for several Target TZs, so long as the h values don't exceed 120. Once they exceed 120, you switch sides and wait for an opposite-direction setup to appear.

This probably isn't the best way to trade TZs, but it is pretty simple. And with the extra parameters I added it's profitable over a long backtest period and many currency pairs, even after trading costs.

I'm currently forward-testing it, with disrection, as well as thinking about new ideas. So I'm a bit quiet on this thread at the moment.

Bye for now,
NT
You reap what you sow.
  • #130
  • Jun 14, 2015 7:45am Jun 14, 2015 7:45am
Quoting jmn5611
Do you fellas have all the indies from this venture? There is one indie that gives the statistical probability of price coming back to a TZ for a given pair. The TZ project was very nice. I do not think you can trade them because no one has been able to quantify just how far price should move away from a PTZ before you take the trade betting that it will come back to it. I think this is new problem. So far, no has got that right. Anyone that has will most likely never tell.

jmn if you have any experience programming you make some modifications that will show you everything you need to know.
Attached Image

--Top Line = SL--
--Second Line = Expirty between TP and SL--
--Third Line = TP --
--Fourth Line = # of Setups/WR--

**Note this is not the First Opposite TZ method as described by NorthTrader**

What I have found to be the key to making my application of the proposition work is finding an inefficiency in the market. For the stats I just posted, any pullbacks further than 88% from the target to SL is an inefficiency that you can profit off of. The marking pulling back between 0-12% during the overlaps is an efficient market that you will not be able to profit from. This is what makes trading these TZ systems hard. You need to wait for the inefficiency to be created and you need to be there when it happens which equates to a lot of god dam screen time

There is no set pullback that will work. Trading this system is like taking a turn that you yiled on green. No one pattern will ensure you are not going to get hit when you turn, you need to wait till the road is clear for you to turn. Just like you need to wait for the market to create the inefficient in order to make a profit.

Quoting kprsa
{quote} In the RecurrenceStatistic code there is no mention of the opposite TZs. While there may have been people adding other stuff to it, the original code I wrote dealt only with single TZs. Cheers, k edit: the mql code is here: {file}
kprsa, thank you for your work with the indicator. I hope you do not mind but I have modified it to my own liking to fit with two TZ. It works very nicely. The only thing I have had trouble doing is converting it to only work with closing prices. So the close of a candle needs to be H transient or recurrent. Do you have any tips on how to properly do this? I was going to inbox you to ask about it but figured you probably get a lot.
  • #131
  • Jun 14, 2015 10:41pm Jun 14, 2015 10:41pm
Quoting Sis.yphus
What I have found to be the key to making my application of the proposition work is finding an inefficiency in the market.
Hey Sis., thanks for posting this but I'm afraid you've lost me a bit here (Monday morning and all that). Are you talking about an option strategy? OTM = Out of the money, right? And the second line ("Probability of resolving 2nd PTZ AFTER 1st gets Confirmed : 88.5714%") means that the expiry period is between the last bar of the CTZ (1st PTZ) and the last bar of the PTZ (2nd PTZ)?

Or are you simply saying that when trading towards a PTZ following a CTZ, it only works over time if you enter > 88% from PTZ to CTZ? Which is a relatively rare event, as you mentioned.

If you're able to explain this a bit further here, maybe with a chart, that would be great. If not, no problem. Thanks
You reap what you sow.
  • #132
  • Jun 15, 2015 7:24am Jun 15, 2015 7:24am
Quoting NorthTrader
{quote} Hey Sis., thanks for posting this but I'm afraid you've lost me a bit here (Monday morning and all that). Are you talking about an option strategy? OTM = Out of the money, right? And the second line ("Probability of resolving 2nd PTZ AFTER 1st gets Confirmed : 88.5714%") means that the expiry period is between the last bar of the CTZ (1st PTZ) and the last bar of the PTZ (2nd PTZ)? Or are you simply saying that when trading towards a PTZ following a CTZ, it only works over time if you enter > 88% from PTZ to CTZ? Which is a relatively rare...
Ahh Mondays... haha I apologize for the confusing post about it, let me clear it up.

--OTM = Out of the money, right?

No I am not trading it as options, I know the OTM label is confusing but I found it easier to use stops and targets. OTM does mean out of the money but what I meant by it is the trade hit neither the Stop or TP(considering them my strike prices).

-- And the second line ("Probability of resolving 2nd PTZ AFTER 1st gets Confirmed : 88.5714%") means that the expiry period is between the last bar of the CTZ (1st PTZ) and the last bar of the PTZ (2nd PTZ)?

Yep that is right. As long as the H left of the 2nd TZ overlaps the H right of the first the trade is doesnt expire until the 2nd H right is resolved or all the H time passes.


--Or are you simply saying that when trading towards a PTZ following a CTZ, it only works over time if you enter > 88% from PTZ to CTZ? Which is a relatively rare event, as you mentioned.

What I was saying with that is, since when the PTZ setup occurs, there is an 88% chance price will hit the 2nd PTZ over expiring or hitting the second, for that trade to be profitable, price would need to pull back past the 88% retracement level(market needs to pull back atleast 12%+) between the two TZ in order for the market to provide a R:R ratio that would be profitable to trade.

The setup occurs many times and it often does not pull back more than 12% to that 88% level., and if you took entries above the 88% level it would be a losing system.

So if the hit rate is 95%(some one the 1 minute are) we would need a retrace of atleast 6% past the 95% level for an entry that is profitable.

I hope that clears it up a bit
  • #133
  • Jun 15, 2015 11:15pm Jun 15, 2015 11:15pm
Quoting Sis.yphus
{quote} kprsa, thank you for your work with the indicator. I hope you do not mind but I have modified it to my own liking to fit with two TZ. It works very nicely. The only thing I have had trouble doing is converting it to only work with closing prices. So the close of a candle needs to be H transient or recurrent. Do you have any tips on how to properly do this? I was going to inbox you to ask about it but figured you probably get a lot.
Hi, I have made the code free for precisely the reason that people could modify it for their purposes.
As for your question, I guess you want to in principle restrict the recurrence to the body of the candle. If my understanding is correct I think the easiest way to modify the RecurrenceStatistic indicator is to use auxiliary variables in two steps:

1) change the extrema array elements for example High[i] to an auxiliary variable, say hii. Define hii as High[i] beforehand in the code. The result of the code should not change by this.
2) change the definition of auxiliary variables, say hii to Math.Max(Close[i+1],Close[i]).

I think that would yield the result you seek.
Cheers,
k
  • #134
  • Last Post: Jun 16, 2015 12:16am Jun 16, 2015 12:16am
Quoting Sis.yphus
I hope that clears it up a bit
Thanks Sis. Yes, this clears it up a bit. By the way, is this the system you were having so much success with last month on lower TFs? And the main drawback is too much screen time waiting for the rare opportunities? How about if we code an indi to alert us somehow (audibly or by email) when an opportunity approaches? Do you think that would help? Cheers
You reap what you sow.
0 traders viewing now