Market condition - flat or trend? Which dominates? - page 5
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
You are told throughout this thread, but in different words, that every trader's TS has its own notion of a trend.
So I'm not asking you to set out the developments with "my" concepts. Let everyone have their own. Were they made in principle? Has anyone asked themselves this question? So far I have found something similar (in the sense of statistical research) only from Igor Kim (link below).
... Until that happens, you can yell at yourself and have a conversation with yourself too.
I am a proponent of constructive rather than emotional discussion. I think the "yelling" topic can be closed. I asked politely (using the word "please"), not yelled. And also apologised if that kind of treatment irritated anyone.
The links were in the threads
https://forum.mql4.com/ru/11667/page10 "Martingale is not evil at all..." Where it was suggested that price movement patterns work, not martingale
https://forum.mql4.com/ru/10963/page29#71542 "Beginning trader works...." Where the author of the thread was asked to perform statistical measurements to find out how well the system works. In addition the author was also interested in martingale
https://forum.mql4.com/ru/11548/page5#71565 "Breaking through the morning flat....". The participant of this topic had made statistical researches in another field and wanted to hear his opinion.
KimIV "My statistical research has shown that there are quite distinguishable correlations between the days of the week, on which one can make real money....."http://kimiv.ru/forum/viewtopic.php?t=13&postdays=0&postorder=asc&start=0&sid=b9c96ff77d0aa74f08022f6068c55c16
As you can see everything is on topic. Moreover, I did not develop my questions in those threads not to litter them, but gave a link here
to Xadviser
Я сторонник конструктивной, а не эмоциональной дискуссии.
That's great!
The notion of a trend is quite clearly defined by mathematical statistics for time series with the required characteristics (otherwise, with certain limitations on statistical parameters). But these criteria do not work for quotes due to violation of these very limitations. For this and other reasons, the set problem simply does not have one formal, reasonable solution in principle. And it should only be sought within some, specifically set PURPOSE.
I, for example, had the task of finding an empirical dependence of the lifetime of channels constructed on the basis of linear regression. And that dependence should have a statistical advantage. Why? That's a whole other story.
But all is not smooth for such channels either. "Authorities" claiming that a flat prevails are easy to disprove. Not enough trends? It is enough to start an iterative process of searching for these very trends. For example, at each reference point in some historical range we need to enumerate the linear regressions (at a fixed current datum) and leave for each reference point only the LR that has the maximum duration in the future. Similarly, one can refute those who argue that the trend prevails. You need a 50/50 ratio - just as easy to arrange :o) But there is another subtlety, the point is that a linear regression or a channel can be built, literally, on any data, but formally (from a mathematical point of view) the built channel will not be a trend. I came up with the following as a pseudo-trend criterion: if the channel "lived", i.e. the price did not go out of its borders one more initial length, then the trend was rather a trend. This is due to speculative calculations of the probability of occurrence of a trend of a certain length in a "random" time series. If you set the historical range for finding channels on the clock, from 300 counts and above, then trends will be everywhere.
In this regard, I recommend reconsidering the set objectives, I repeat, they have no solution in principle. Specify the intended criterion, which is probably related to trend and non-trend.
PS: And still treat the participants more accurately, this is not a military unit and no one will march in formation. Otherwise we might complain to Prival and he'll be an old colonel... :o)
To Neutron. IMPORTANT!
https://www.mql5.com/ru/forum/50458 post "grasn 11.01.07 16:16".
Seryoga, for you with some delay telling (just reread from the place indicated further and realized that his promise was not fulfilled). answer your question "how it all works" - it's simple. So, let's say we have a formula for calculation of the linear regression existence time on the basis of some channel parameters and this formula has a statistical advantage (very important). From the current datum (it is fixed), we iteratively look through the past (historical) datums and build a linear regression for each sample. As Vladislav wrote, we get a fan "going" into the future. Now for each such channel we calculate its probable length. Now we obtain neither a fan, nor a straight channel (reversal zones appear) where the price will develop. And if we take into account the price position in the formula, we can calculate the zone of price movement more accurately. Just when collecting statistics, we should not forget that the channel breaks when the price leaves its borders, respectively goes up or down and this defines the price position quite precisely, i.e. if we get the calculated length of the channel, the price will leave it either up or down, this can be reasoned. Serega, well, I haven't promised to tell you about the contrivances (waves, diffusions ...) :o))
to Prival
... Everybody praises these NS, but I lost my confidence in them about 20 years ago, I used to program similar things, may be I'm mistaken and my knowledge is outdated, like that of a dinosaur...
Nothing has changed, if there is not even the slightest regularity in "behavior" no NS can predict this very behavior. And it is necessary to look for these very laws-dimensions, while NS is only a tool, though a very good one :o)
Do I understand correctly that you have a "definition" of trend/float, but since this definition is actually the essence of a breakout TS you don't want to publish it? But for other "definitions" the statistics will, in general, be different. So the point of this thread personally remains unclear to me. If the problem is in the program for collecting statistics, you do have to share your "definition" with at least one of the programmers. But you are unlikely to find the definitions that allow you to build a reliable and profitable TS. Even if someone has such a definition. Which of course is unlikely :)
1. the definition exists, but it does not imply in any way the possibility of opening trades with a favorable outcome in the future. Because we are talking about gathering statistical information in the past. I.e. we assume (for our input parameters) that at that period of time (T1) price was within the specified range, while at that period (T2) it wasn't. We gather all T1 and T2 on the selected measurement interval. Therefore, there is no TC in this case.
2. There are no secrets. I will post my soaps a little later (I will try to elaborate)
3. The point - to get information about the market. And to use (take into account) everyone in your TS (or not). The example of such researches and their consideration in trading
тут - http://kimiv.ru/forum/viewtopic.php?t=13&postdays=0&postorder=asc&start=0&sid=b9c96ff77d0aa74f08022f6068c55c16
KimIV "My statistical research has shown that there are quite distinguishable dependencies between days of the week on which you can make real money..... "
But I was "pushed" to use them by the primary source
It is generally accepted that the market is mostly sideways and trends about 15-20% of the time. https://book.mql4.com/ru/samples/expert
Draw your own conclusions, or check the opinions of "authorities".
Xadviser wrote (a):
It would be more correct to use the term "Price Range". And as mentioned above it (price range = flat) can simply be set, not solved.
I argue that the market is in a flat 100% of the time:
So much for the "price range"...
Xadviser wrote (a):
It's just that for me this "philosophical question" has been solved, and so I've moved on to the next one, not taking into account the lack of answer many have to the first one.
...
If you are willing to assist me in solving this problem, I would be very grateful. As promised, I will post the results here.
And now we have the now-standard situation: "I'm asking you to give me, but you won't get anything in return".
Make the first step - propose the criteria for separation of trend/flat, get people interested.
Writing a script that collects statistics is not a problem. If it's really interesting, I'll be the first to do it.
to Neutron. IMPORTANT!
https://www.mql5.com/ru/forum/50458 post "grasn 11.01.07 16:16".
Seryoga, I am telling you about it with some delay (I just reread it further from the designated place and realized that I have not fulfilled my promise). I am answering your question "how it all works" - everything is simple. So, let's say we have a formula for calculating the lifetime of a linear regression based on, some channel parameters and this formula has a statistical advantage (very important). From the current datum (it's fixed), we iteratively look through the previous (historical) datums and build a linear regression for each sample. As Vladislav wrote, we get a fan "going" into the future. Now for each such channel we calculate its probable length. Now we obtain neither a fan, nor a straight channel (reversal zones appear) where the price will develop. And if we consider the price position in the formula, we will be able to calculate the zone of price movement more accurately. Just when collecting statistics, we should not forget that the channel breaks when the price leaves its borders, respectively goes up or down and this defines the price position quite precisely, i.e. getting the calculated channel length, the price leaves it either up or down, this can be worked out. Serega, well, I haven't promised to tell you about the tricky part (waves, diffusions ...)).
I apologise for the intrusion. Is there anything you can look at? And generally speaking, I'm interested in this topic, but I haven't understood everything from the middle of the conversation. If you don't mind, you could start a new thread on the subject.
Citing an authority is no excuse. It would be funny if I (or anyone else) would say "but MACD_Sample says so!
You draw your own conclusions or verify the opinions of "authorities".
Andrew, please don't be offended, but I'm just trying to get my point across this thread (maybe I'm bad at it), that my doubts came after this very statement
So much for the "price range"...
You only have one sample on the study interval. That's why the statistics won't work, although you're formally right (I've pointed out a similar variant with reference to the author of that very statement about 20/80).
Take the first step - offer criteria for separating trend/float, get people interested.
Writing a script that collects statistics is not a problem. If it's really interesting, I'll be the first to do it.
I am not asking for anything yet. I wanted to find out before something to ask or offer whether this question was asked by someone and how it was solved.
My vision and solution is set out below
A bit of backstory.
I have come across similar statements many times: " It is generally accepted that the market is mostly sideways,
and trends in the market take about 15-20% of the time. https://book.mql4.com/ru/samples/expertUnfortunately, the author did not specify criteria for evaluating these values.
Taking this statement as a base it is obvious, that we have a statistical advantage. Why nobody uses it? Apparently it is not entirely true.
According to my superficial observations, time of price staying in a conditional range (Flat) and time of transitions between these ranges (Trends)
are roughly equal on average.
Last year I started manual testing of my TS. One of the aspects taken into consideration when making the TS was the assumption of the equal ratio
trend and flat areas over a long period of time.
The test ended positively. However, a doubt crept in that the result may not depend on the conditions adopted at all, but was a random one.
The manual test did not fulfill all entry conditions due to physical impossibility.
It is necessary to check the principles implemented in the TS. And first of all, the assumption of the flat/trend = 50/50.
How to do it? One of the possible ways is to set necessary parameters required for the evaluation.
It is quite possible that the author of the 20/80 statement has taken the calculated range of some thousands of points, then the ratio can be considered as close to the truth.
However, in order to get more reliable statistical data the interval under study should be much larger in relation to its components.
In general, any statistical information in any aspect of activity gives it some estimate and possible starting points for analysis.
Unfortunately, I almost did not manage to find open statistical data about Forex, except for some particular researches. I have given the links to them above.