An effective trading strategy based on multi-currency analysis of multiple DCs - page 7

New comment
 
Right, so I enter and exit the market when it is calm, but if you look deeper, the main advantage of this is to predict in time, suppose there is an order, which was set a couple of days ago, we feel the change of situation and think when to withdraw in time, suppose that such an analysis will save the situation, for example we will not lose or win even more, without any apparent danger, that is roulette as a visible factor, will step aside:) It is before the pit, not when it appears or after it. It is before the pit, not at the time of its emergence or after it, I do not believe that this is pure chance or luck, because I'm one of those people who are not lucky if they have not put their own work, and it is not empty words, it's just there.

And the funnels do not go back, just not within a day, these kinds of drops as a rule, either continue in the same direction, or talk about a reversal, the difference in this I have not yet grasped for sure, but I think it also has its character, because I have so far limited to identify the entrance to the funnel, rather than out of it, for me it will be the next step.
 
I assume that before the funnel you register the frequency of tick arrivals or on the contrary the calm before the storm and how prices at these moments from tick to tick change - often it happens just a minute or two before the jump if it is before the news. Well, if the funnel is not a news one, I do not know how. Maybe you can see something. I haven't investigated it before.
 
It is the calm before the storm, the ticks literally walk in one place and sort of freeze, that is why I call it a tornado funnel, not a bear or bull candle, but this calm is very characteristic, I do not even know how to describe it properly, because it depends on the story, smooth flows, not sure or contrary, but it comes like a complete calm in the market and then something happens that did notI don't think so, I may have got a margin call on a real account, or in other words I lost it, about one year ago and I have been trying to describe these unpleasant phenomena since then.The matter was about pennies in the relative sense, but I can still see that chart in front of my eyes and it is not about the lost money, it is a matter of principle to find out what has so unpleasantly discredited my assumptions and actions.
 
Renat:
xnsnet:

Renate, have you ever considered that the same tics can be seen in different ways

Been wondering, stepped on a rake. Brokers have automatically adjusted and manual filters which are periodically adjusted. At some point they change one of the parameters and that's it - the whole tick picture is gone. Add an extra 5 second delay and requotes (this is the fate of many pips traders) on the real account and the situation is even worse. Who should we blame? Only myself.

So all the same - "Do not get into ticks, consider the noise, trade less frequently".
I mean: possibility to correct quotes, how they are filtered and adhered to, possibility to combine quotes from several sources, etc.? This is not a question of idle curiosity, I believe it bothers a lot of people, and the answer to this question is very important for the creation of a proper strategy that excludes a conflict of interest between a brokerage company and a trader.
All brokerage companies invite traders to their clients and call us desired clients, but the client differs from the partner in that the client is used, and with the partner they share the profit, but traders have to sacrifice, because there is no choice and no alternative. But I want to avoid unnecessary conflicts of interest and not to strain already complicated relations with brokerage companies.

And what do you call noise? Looking at the links you gave, I came across rather vague definitions, and some definitions say that the noise is a struggle between bulls and bears in a new trend formation zone. By definition, the noise is noise and distortion from the outside and superimposed on the main signal, the struggle is a signal formation, it is not random and depends on the market laws. The earlier xnsnet pictures are not indicative of noise, if you separate the trend from the first two pictures from different DCs, I am sure the difference will be not only in ticks, but also in the trend. There is no uniform centre of quotes formation on the market, and when we receive data from different sources, we will have a different picture, but it is not due to noise at all.

I cannot agree with you about uselessness of ticks analysis. Its analysis allows to reduce the influence of "noise" made by brokerage companies through manipulations with quotes. And the question of pipsing and scalping has nothing to do with this topic. No one says that we should trade trying to catch the pit shown in the picture. But if an expert system is available that can carry out efficient analysis and make forecasts, having forecasted such pits, you can make a correct decision about opening or closing positions before entering the pit, considering the time needed for order processing.
Also it is not correct to suppose that a change of quotes source by the brokerage company will change the entire picture, it depends on the Expert Advisor system stability. In my system I changed brokerage companies with significant differences from the previous quotes without retraining using new data and some tools for analysis because they were present in one brokerage company and not in another. But even such changes didn't influence the accuracy of the forecasts and the system didn't need retraining. So you are not quite right in your assertions.

411
Mathemat 04.05.2007 17:05

I think it's self-destructive to work with gold on ticks. I think it's better to work with it on larger timeframes. It likes very much to draw huge studs, figures of about 20 (gold).

You are also wrong. Of course pipsipsips are out of the question, but it is possible to do normal trading and very effectively. And the statement, that I cited at the beginning of the thread, shows it. I tried to use my expert system with ticks, but it works fine on 1 minute gold chart.
When working on a larger timeframe you will always lag behind the market, though the market is inertia and it takes time to restructure even on the news, but working even on a 5 min timeframe means to follow it, not to predict its behavior and therefore always miss possible profit and often get losses.
 
xnsnet - I wrote you a letter, asking you to send me tick data in text format of the first two charts from different brokerage companies you posted in my thread. I want to highlight the trend and show that the difference is not caused by noise, but in the trend itself. But according to your link xnsnet _AT_ cln _DOT_ ru, - does not reach, please specify where it should be - @?
 

Dot where DOT dog where AT underscores as excessive security as these words define as well, spam machines... In the end write on my forum in the PM, I certainly understand not want to register too many times, I understand myself have the same problem, there are many forums:) Ticks are not fixing, and why, in my opinion it's enough, it's the constancy with a fraction of periodicity:). Record of ticks will make if necessary full-fledged, for the time being I do not care much:))

I thought, how to show better the time intervals between ticks, because I have shown charts with funnels, and such space as time is left over the horizon, it cannot be seen in pictures, as it should be seen. As a result, I add two lines to the ticks at the top and bottom, one server tick time, the other triggered tick time, that is, the time recorded by my program on the facts of the receipt, the difference in seconds between the ticks, I think I'll make display a difference in different scales, up to milliseconds, at least in the fact of the call exactly.

Bottom server time, top actual time, on this graph, one pixel == one second, distance from the leading tick. Now it will be easier to show what funnels are. Of course if I had done a walk through history, then I would have shown the funnel on the screenshot as well, so far I haven't done it and I don't write files for the same reason. Worried about this kind of output details, the story less so.


 
Clarify, what do the grey lines represent?
 

The distance from the lead tick in seconds is represented. One hundred pixels in height from the tick location, one hundred seconds, I used to stretch ticks by time distance but it turned out to be very inconvenient and it was not realistic to show two time values. There are no grey lines, so the interval is less than a second.

I first thought of drawing the time in sine wave between ticks, but again, that takes away the scalability. And here you have all the stored data, two time coordinates and one price coordinate by bid per tick. Ideas are implemented in minutes to hours, and thinking about them stretches over hours to days of weeks, until you find something that fits the essence of the issue.

Here is another screenshot

 

Judging by the last screenshot it seems to me that the difference in comparison of two brokerage companies will be absorbed by time intervals, because even on one chart time intervals are compiled in server time noticeably, relatively to the actual time. But the number and accuracy of ticks will not decrease or increase:) It starts to occur to me thoughts like: "How much will you give me if I prove, that analysis between DC is not worth a dime, well, speaking of thoughts:)

 
xnsnet:

Judging by the last screenshot it seems to me that the difference in comparison of two brokerage companies will be absorbed by time intervals, because even on one chart the time intervals are compiled in server time noticeably, relatively to the actual time. But the number and accuracy of ticks will not decrease or increase:) It starts to occur to me thoughts like: "How much will you give me if I prove that analysis between brokerage companies is not worth a dime?

Analysis may or may not be worth it - depending on the context in which you use it. A single person looking at your charts, will not see anything informative in them, while you can use them to some extent to predict the behavior of the market. And how are you actually going to prove it, what you show are just input signals for the expert system, how it will analyze them and what conclusions it will draw, you cannot model it without knowing the system and its operating principles. To complete the picture, try to display the quotes of different brokerage companies on one chart, or at least different symbols in one scale of one brokerage company.
As an example, I will show a graph for two instruments on М5: pink line - close USDJPY, blue line - the trend of close USDJPY, black line - close USDSEK, red line - the trend of close USDSEK. In principle, the trends can be replaced by moving averages, the picture will be similar. This example shows how multicurrency analysis can be used by those who work only with standard indicators, and who do not use more complex systems for analysis. If you trade on the USDJPY, using the second instrument as an auxiliary one, you can see the reversal points much earlier and more clearly than when using different indicators on one instrument. Here USDSEK is shown at the same scale as USDJPY.



However, even without any processing the charts themselves give a clear picture, if they are displayed on the same scale in the same window.
1234567891011121314
New comment