Discussion of high-frequency trading on MT5 - page 21

[Vladimir Kustikov]

hrenfx:
This is an ancient post of one thread. At that time, the implementation of the calculations had already been posted. Further on, very specific research was continued.

I would like to ask what is your current status in these studies? Have you arrived at something worthwhile or are you still in the process of searching?

[hrenfx]

Vladix:
I would like to ask - what is your status today in this research? Have you arrived at something worthwhile or are you still in the process of finding something?

I would not like to be the hero of such scenarios:

papaklass:

Now, about trading. How do traders share information? You tell a story about a solution to a problem and suddenly you say "well, I won't tell you any more because ...." or some other similar phrase. And that's it. There will be no substantive discussion. The followers will not be able to repeat the research of the author. This information is for nothing, information for the sake of information. I understand that the author doesn't want to reveal trade secrets, and that's fine. What's not normal is another, why start saying "A" when you know that "B" is not going to happen. When you can't lay out the full methodology and results because the idea is still profitable. Why? What for?

So I will just say that for me the research process itself has been very useful and full practical use of the results has not yet been made.

[Андрей]

hrenfx:

I would not like to be the hero of such scenarios:

So I will just say that for me the research process itself has been very useful and the full practical use of the results has not yet been made.

I think Vladix meant not the results of your research in all the details and conclusions in the form of ready-made algorithms with all the settings and recommendations))) Such clearly only a sucker would post such a thing for free. I most likely meant the issues you raised. Opening such information and you do not hurt commercially and others will give food for thought, thereby increasing your credibility, if you do it has any value)))) Something like that, only for the moment.

I am personally intrigued by your arguments in some posts, and I think so are many others. Although I don't know how to apply such technologies in practice, but from a purely intellectual point of view it is very interesting.

[Heroix]

gunia:

I think Vladix didn't mean your research results in all the details and conclusions in the form of ready-made algorithms with all the settings and recommendations))). Such clearly only a sucker would post such a thing for free. I most likely meant the issues you raised. Opening such information and you do not hurt commercially and others will give food for thought, thereby increasing your credibility, if you do it has any value)))) Something like that, only for the moment.

I am personally intrigued by your arguments in some posts, and I think so are many others. Although I don't know how to apply such technologies in practice, but from a purely intellectual point of view it is very interesting.

Yandex to help. Hrenfx has already provided a lot of material. You can't dig it all up. Why would he do it again?

[Андрей]

Heroix:
Yandex to the rescue. Hrenfx has already provided a lot of material. It's a lot to dig through. Why would he do it again?

It's understandable, why turn on Captain Hindsight, I use Google. Sincethe questions have been announced in this particular branch, we can deal with them here, so that the branch has considerable informational value. For example, I provided a link to question 10. Pricing algorithms. We wish we could gather articles, qualified forum posts or any other concentrated material on each question that would thoroughly explain the subject. And there would be happiness.

It is clear that somewhere on the Internet there is all the information you need, the point is to get it quickly and accurately, and if you search for it with a search engine, you don't know that it will be like that. I'm not proposing to reveal the secrets, but just pick up open-ended information on each issue.

Although we can simply squabble, discuss equity volatility, cram like a teddy bear in a closet, and gather information that will later turn out to be trash, because the probability of one person making a mistake is greater than that of a whole group.

This is just thinking out loud... Maybe papaklass is right, a trader is a wolf for a trader and the synergistic cooperation, even on non-threatening issues is out of the question.

[Anatoli Kazharski]

gunia:

...

Well, that's just thinking out loud... Maybe papaklass is right, trader to trader is a wolf and synergistic cooperation, even on non-threatening issues, is out of the question.

Worse. At least wolves gather in packs. ))

[Server Muradasilov]

There is a joint movement of financial instruments - but each time one stops others go farther away, because of the withdrawal of funds and their filling into the market.

[Server Muradasilov]

server:
There is a joint movement of financial instruments, but each time some of them stop, others go farther away; this is due to withdrawal of funds and their distribution to the market.

Indices, in particular, are indexes. The creation of indices is the norm, as an economic indicator of some kind. But when it starts to trade, it is an artful raiding of money from other market participants). In order to rake in money from other participants someone has to sell out, big companies have big losses too and often - trading robots have no intelligence and never will have it

[Server Muradasilov]

For information - there are various formulas can and answers to our questionshttp://yadi.sk/d/f-SRV_8u2TiEFhttp://yadi.sk/d/jJdntoel2TiDb

[hrenfx]

I'd be happy to share information, but I have no idea how to do it in a forum format that won't be a burden in the future. I am not in a position to write articles.

In the end it all comes down to realising what one wants to get. And one wants to find real market correlations. Obviously, if in the sample of two BPs the first 90% of time the prices of each practically didn't change, and in 10% of the remaining time they did, then 90% cannot be taken into account when looking for correlations. Therefore before applying the mathematical methods, the initial TEVs are always transformed in different ways (the simplest one is BPs from local extrema). This also allows to remove temporal synchronization, etc. In general, the relationships found in this way are much more adequate than without prior transformation. Such transformations can be used in simple linear mathematical methods to find non-linear interrelations. I.e. there is no need to make a vegetable garden with chi-squares and other nonsense.

If we consider each FI separately, we have come to understand that we should not look for flat and trend, but for a stable LINEAR regression. Interesting results are obtained by examining, by the hour, when which FI has the most stable linear regression.

In general, the character of each FI movement is enormously dependent on the time of day and even on the day of the week. Obviously, to make your instrument work round the clock is a strong self-limitation in the search of patterns.

I have also formed the hypothesis of a symmetric distribution. For example, for linear regression there is always the issue of sample size selection. But it can be made adaptive by the following algorithm: find the outermost sample, which has the most symmetric price distribution around the linear regression. Such a sample, or rather a linear regression with such a period, will most accurately reflect the current market realities.

When you set your own limits influencing the best price in the ECN/STP, you begin to analyze how likely you are to put your own limit here. In other words, the price levels become a probabilistic view. What is the probability that one of thousands of market participants will put his limit here? So it is investigated.

In short, I could write a lot of nonsense, and it will take a long time to go into each point. I have no copyright on posts and works. Use as you see fit. The main purpose is no longer a conversation, but just to get people thinking.

Well and on pricing in a broad sense, the article cited above is a lot of water. I see this topic primitively as follows:

Задача алгоритмического отдела банка выдоить как можно больше денег из других участников рынка. Понятно, что в конце-концов все участники рынка прямо или косвенно имеют торговые счета в нескольких крупнейших ММ-банках мира. Эти банки обладают огромной статистической базой по торговым особенностям очень существенного среза рынка FOREX. И алгоритмические отделы на основании разрабатываемых ими же мат. моделей, примененных к этой инсайдерской информации, разрабатывают свои алгоритмы ценообразования. Они полностью допускают возможность заработка некоторыми участниками рынка. Но главный показатель их результативности - разница между сливающими и зарабатывающими.

Сами ММ-банки между собой конкурируют, так что какой-то однозначной монополии и согласованности между главными инсайдерами нет.