Market condition - flat or trend? Which dominates? - page 14

New comment
 
Serg_ASV:
I can say the same about most indicators, if not all - one thing is beautiful pictures on the history, and circled entry and exit points (beginning and end of a trend) - and quite another thing is real trading...

Please read carefully what was written earlier. (I hope no offence)

  • No indicators are used, but criteria (conditions) are used to get information within given conditions on historical data.
  • We are not talking about trading, we are talking about statistics.
I want to write about indicators for a long time, but I am lazy. It is the first thing I tell my students about indicators. Are any of the traders working using price charts in the form of lines? No, of course not. Either bars or candlesticks. So, why has not ANY indicator been displayed the same way as the price is displayed? After all, in the end we only see their final position. It would be the same if prices were displayed only by close points. It is high time to oblige all indicators to display the indicator at least with the values corresponding to the extreme price positions. This would eliminate a lot of questions and superdealers' desire to make another grail soup from the set of indicators.
 
Xadviser:
A similar variant with transfer to Excel is of course welcome.

I will give you a hint: the files recorded by TrendFletAnalysis_2mS open directly in Excel :).


Writing an indicator for any reason is not the most rational way to work with statistics. It is more rational to find the regularity first (by working with the data even in Excel) and only then apply it as an indicator. This is my experience.


Speaking of patterns. Has anybody except me noticed the tendency of increasing of time interval, taken by a fixed number of segments of given zigzag? At first I thought this was a reflection of changes in the server's quote filtering procedure. However, it would seem that it should only affect a small "step". But here is a picture showing data on ZZ with the thresholds of 20 and 100 points. In this interval, changing the PP threshold does not seem to affect the trend.


P.S. Actually, I have already raised the question of a trend of increasing average length of zigzag segments, but for this zigzag, the trend is quite strong. And the second point - the persistence of the trend within a wide range of zigzags parameters is a good illustration that the market is not alien to fractality.

 
lna01:

I'll give you a hint: the files recorded by TrendFletAnalysis_2mS directly open in Excel :).

I certainly guessed it, but you overestimate my capabilities :)

I consider myself more of a user, although by education I am an engineer. But I came to market from trading, where I had to deal with forex (hedge) and got result of successful application of statistics. For example I used 20/80 rule, known as "Parett's Rule". Its different interpretations are: 20% of clients give 80% of revenue (income), 20% of efforts give 80% of result, etc. The rule worked (of course not with the exact ratio, but about the same). By applying it I managed to double my revenue. I am sure that this rule is applicable to the market. (got a little distracted:))

Writing an indicator for any reason is far from being the most rational way of working with statistics. It is more reasonable to find the regularity first (by working with the data even in Excel) and only then apply it as an indicator. I am speaking from my own experience.

Of course I agree, but

  1. I am more of a visual person. For me pictures (clear) speak more than descriptions. That's why I find it difficult to express my thoughts remotely, for me it's much easier to explain in pictures.
  2. All ideas are born from visual observation of charts, as a consequence I look in the same direction for analysis.
  3. As an indicator, also because initially there were thoughts of using it as a signal.

Speaking of patterns. Has anyone besides me noticed a tendency for the time interval occupied by a fixed number of segments of a given zigzag to increase? At first I thought this was a reflection of changes in the server's quote filtering procedure. However, it would seem that it should only affect a small "step". But here is a picture showing data on ZZ with the thresholds of 20 and 100 points. In this interval, change in the PP threshold does not seem to affect the trend.

Not yet :) I have told that I do not have enough points to study and compare them properly.

Let us decide on a name. I propose to name the new indicator TFS (Trend-flat segments) not to confuse it with ZigZag. (You may suggest other variants).

With the trend yes.... though there are some ideas, and based on the data presented by you (.... but for this zigzag the trend is quite strong. And the second point is that the trend persists in a wide range...). But I want to make sure once again that the results are correct, and not only for EUR/USD.

P.S. Actually I have already brought up the question about tendencies of increase of average length of zigzag segments, but for this zigzag the tendency is too strong. And the second point - tendency persistence in a wide range of zigzags parameters is a good illustration that the market is not alien to fractality.

Well, I am knocking :) I started reading this branch, but it seemed to me that people went too deep into the wilds, and I didn't get to your messages. Well, the results of your measurements are of course very interesting. (I mentioned them, didn't I?) If you can, please reply.

  • What goals have been set and have they been achieved?
  • Have the expectations been met?
  • Did it lead to any decisions, thoughts, etc.?
  • what parameters were measured? with what parameters of ZZ?
  • You also mentioned ".... I plotted the duration of zigzag intervals for my own ZZ." What do you mean for your own? How is it noteworthy or different from others?
 
Xadviser:
lna01:

I'll give you a hint: the files recorded by TrendFletAnalysis_2mS open directly in Excel :).

I certainly guessed it, but you overestimate my capabilities :)

I'm afraid to seriously engage in market statistics without working with the data myself and, moreover, relying solely on volunteer programmers will not work.

I agree of course, but

  1. I am more of a visual person. And pictures (clear) tell me more than descriptions.
In terms of visualisation, MT is exclusively time series oriented. A specialised program will allow to build a much more diverse picture :).

Let's still decide on the name. I propose to name the new indicator TFS (Trend-flat segments), not to confuse it with ZigZag. (other variants may be suggested).

You can do it even if you're a rude man ... :). All the more it is not ZZ, it is only based on ZZ.

With the trend yes.... still need to be sorted out, although there are some thoughts, and based on the data presented by you (.... but for this zigzag, the trend is quite strong. And the second point is that the trend persists in a wide range...). But I would like to make sure once again that the results are correct and evaluate not only EUR/USD.

Thoughts are the most interesting thing. By the way, raw - they can be ... er... more nutritious :)
If you have a possibility to answer, please ...
No goals were set in regard to this effect. If it turned out to be a property of a particular zigzag, the zigzag would have to be abandoned. But it seems to be a property of the market, and I don't see any benefit other than harm from it yet. I still do not want to talk about my zigzag publicly, and in short, as experience shows, it does not work.
 
lna01:

I'm afraid you can't do serious market statistics without working with the data yourself and relying exclusively on volunteer programmers.

Yeah, I guess I'll have to look into it. Although it seems that you have to do it once (the thought was comforting), but unlikely.

In terms of visualisation, MT is focused exclusively on time series. A specialized program will allow to build a much more diverse pictures :).

Which one would you recommend for mastering? Or if I can unload the data into Excel, I can draw pictures (charts) there too? By the way, how can it be done (the files recorded by TrendFletAnalysis_2mS can be directly opened in Excel)? If it is not a problem...

Thoughts are the most interesting thing. By the way, raw - they can be ... er... more nutritious :)

Well, for example the most delicious :)

Since we have a trend in some (fairly wide) range, i.e. at different channel widths, we can

  1. run several independent Expert Advisors with non-multiple ranges (e.g. 25, 55, 130 or 30, 50, 70 something like that combination) and see how they work together. Since on the average they are all on the plus side (we are talking about Euro and channel break), when working together, the combinations will be of the opposite direction, which should reduce the drawdown, compared to the work of one expert.
  2. strengthen the trend in any other way.
  • using the same channels, but this time "playing along" with the lower channel (of a smaller range) on the priority of the higher one.
  • or vice versa, evaluating a move inside the higher channel as a flat one, open on the lower one from the border inwards.
  • Using a (still need to think over the optimal) algorithm for increasing positions in the direction specified by the criterion.

Everything will be seen more clearly when Mr. Komposter finishes his indicator. It will be possible to more clearly estimate and illustrate in the figures. For now it is only a guess. These are the raw thoughts that arose in connection with the data obtained. But there are also less "raw" ones :)

I also consider not only the distribution but combinations of sequences to be important data. For example, transitions from the uptrend to the downtrend. What is the ratio of flop and non-flop transitions, etc.? It will be possible to correct the trading conditions depending on these data.

No goals were set for this effect. If it turned out to be a property of a particular zigzag, the zigzag would have to be abandoned. But it seems to be a property of the market, and I don't see any benefit but harm from it so far. I still don't want to talk about my zigzag, and in short, as experience shows, it does not work.

What is the harm?

 
Xadviser:

Which one would you recommend for mastering? Or if the data can be unloaded into Excel, can I draw pictures (diagrams) there too? By the way, how can it be done (the files recorded by TrendFletAnalysis_2mS can be directly opened in Excel)? If you don't mind...

I use matlab. Many people prefer matcad. Excel seems to be popular too. Files are opened using the "File" menu item of the same name, only the format of the cells should be set to "number". But in general, such questions are not for this forum, it is necessary or on their own or with a tutor.

Thoughts are the most interesting thing. By the way, raw - they can be ... er ... more nutritious :)

Like this is the most delicious :)

...
Well, at this stage that should have occurred to me . There's probably no other way but to work it out honestly :)

What's the harm?

The harm is that the distributions get smeared
 
lna01 писал (а): But in general such questions are not for this forum, it needs either on your own or with a tutor.

I won't pester you with childish questions any more

Well, that's about as far as it should have come at this stage. I guess there's no other way but to work it out honestly :)

The list was completed with another crude idea in this direction

  • Using the same channels, but this time "playing along" with the younger one (of a smaller range) by the priority of the higher one. It means to open on the lower one only in the direction where the higher one was pointing.
  • or vice versa, evaluating a move inside the higher channel as a flat one, open on the lower one from the border inwards.
  • using some (still to be thought over) algorithm to increase positions in the direction specified by the criterion.
  • Run it on unrelated pairs with the same trend (unrelated - those which do not have the same values (names) in pairs).

The other came to mind earlier, but we need to see where this road leads as well. Will you test it?

I read the thread you linked to. I get the feeling it's about the same thing, but people have been coming at it from the other side and it's very confusing (here https://forum.mql4.com/ru/9358/page12 and a couple of pages back, if you're interested).

 
Xadviser писал (а): I suggest you delete unnecessary (off-topic correspondence) from the threads.

Will you stop doing that to me! There's a lot of advanced people moving their lips trying to get into the subject and you delete it!

Do you think that if no one writes, no one reads?

 
granit77:
Xadviser wrote (a): I suggest you delete unnecessary (off-topic correspondence) from the threads

You guys stop it! There's a bunch of advanced people moving their lips around trying to get into the subject, and you delete!

Do you think that if no one writes, no one reads?

Thanks for the insight :-) I thought only Composter and Candide were interested (and even questionably so). But don't worry, I'll only delete irrelevant posts (this one too :-))


2 Prival 47th (4th faculty the best !), last graduation 1992. Our people are everywhere.
 
Xadviser:

The list has been supplemented by another crude idea in this direction

  • Using the same channels, but "playing along" with the lower channel (of a smaller range) by the priority of the upper one.
  • or vice versa, evaluating a move inside the higher channel as a flat one, open on the lower one from the border inwards.
  • using some (still to be thought over) algorithm to increase positions in the direction specified by the criterion.
And how is it supposed to get rid of the arbitrariness of their choice? I mean let's say the lowest one can have e.g. a width of 20, 30, 50 etc. points, the highest 100,121, 150, etc. So far I don't see any criteria in this approach to determine the parameters of the senior-younger channels.
  • Run it on unrelated pairs with the same trend (unrelated - those which do not have the same values (names) in pairs).
Can you elaborate on what this might give?

The other came to mind earlier, but we need to see where this road leads as well. Will you test it?

It's very easy to build up a sea of all sorts of data and drown in it.

I read the thread you linked to. One gets the feeling it's about the same thing, but people have come at it from the other side and it's very tricky (here https://forum.mql4.com/ru/9358/page12 and a couple of pages back, if you're interested).

Perhaps any approach may eventually result in the breakout and/or rebound game, and in this sense the Forum is no different. Perhaps other formal analogies are appropriate for that thread, but the devil is in the details, as you know. As for being cumbersome - these people started dealing in the market much earlier than that thread arose. I think if you have been using your approach for a long time you will also become muddled in it :). By the way, it is important to understand that the true purpose of special terms is not to complicate, but to simplify the understanding of ideas.
1...78910111213141516171819
New comment