Optimise an EA and get the best of the optimised ones. - page 26
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
If the account is monitored in myfxbook, then the chart is easily drawn there for any magician in manual analysis, just 3 clicks of the mouse).
In this thread - people have already complained about "one loss in the account".
MyFxBook - it shows monitoring of the whole account. It cannot divide trades by magic.
This is what is required.
The League's 270+ systems all work in parallel on the same account, all without MM, with minimal lots. If you need monitoring, you should use only one preferred trade from this account. Does MyFxBook allow this? MyFxBook does not allow that, that's why I use a script and as you can see all my charts are simple Excel charts prepared with a special script.
how are the results on the real account coming out? are expectations met?
I openeda real account quite recently.
If we are talking about "trading on the demo and on the real account" then of course it is almost identical, up to the spread. However, I have already noticed one deal that went one way on the demo and another way on the real account, precisely because at that very moment the spread was wide enough and on the demo it was lower than the current EMA, and on the real account it was higher. And the system is reversed. As a result, one account has flipped and the other one continued working on the current order. But this is still a rare case.
And if we are talking about profits, it's too early to talk about it; meanwhile, there is "chattering near zero".
And in general the question is about choosing the most stable TS which will trade some time in the same way as it appeared during testing and as it did during demo trading. Unfortunately, I have not yet formed my own criteria for evaluating stability, and the selection of systems for the real trade is more intuitive.
That's interesting - I'll have a look.
But it seems to me that this project is about a batch of TCs, so we need to analyse them together somehow - after all, the point is to ensure that the organism survives when some of the clients die before they regenerate, while it's not obvious to me.
You want to analyse them together - no problem, you're in Excel!
You take two tables and do what you want with them.
Ideally, you can write a script, which makes a joint table (here, I generate my charts with many curves), and then summarize in pairs or triples, as required. How many do you want to have in the generated TC list ? (Keep in mind, when running the script - for each will have to enter its magik, so two dozen - does not make sense, two or three - in my opinion just right)
Those who understand the pointlessness of this endeavour are silent because they don't want to offend anyone.
And some will also think they are jealous.
I, on the other hand, am quite open to criticism.
If the idea is pointless, an open account will go down the drain.
For me - the idea is just not fully developed. The first part of the idea is quite accomplished.
I said that the meaning is to exclude the question "what to do to make an expert earn money" in favor of the question "how to select the most sustainable of the few experts who are already earning money, such that they will continue to earn with the maximum probability.
So, now I'm thinking about the second part of this idea - I'm considering various possibilities of "selecting the best". So far, unfortunately, the selection is more intuitive. I don't like this. I would like to formalise it as much as possible.
Redone
There. Now it's OK.
29 regcodes.
The real account has only recently been opened.
If we are talking about "trading on the demo and on the real account", then of course it is almost identical, to the spread. However, I have already noted one trade, which went one way on the demo, and another way on the real account, precisely because at that moment the spread was wide enough, and on the demo it was lower than the current EMA, and on the real account it was higher. And the system is reversed. As a result, one account has flipped and the other one continued working on the current order. But this is still a rare case.
And if we are talking about profits, it's too early to talk about it; meanwhile, there is "chattering near zero".
And in general the question is about choosing the most stable TS which will trade for some time, just like they showed themselves in testing, and also like they did in demo trading. Unfortunately, I have not yet formed my own criteria for evaluating stability, and the selection of systems for real trading - is more intuitive.
I have not yet formed my criteria for assessing the robustness of these systems, so the selection of systems for real is more intuitive.
Maybe register a signal, I would look at the results, or monitoring
maybe just take all the successful ones, and in real time, as they fail, replace them with others?
Maybe a signal register, I'd watch the results, or monitoring
That's what I'm doing !
And, understandably, I will be replacing anyone who shows "control shots".
And the signal will be, from the real $100. I have already had a few suggestions in private with the meaning of "we do not want to bother with your proposal, open the signal - if it is good - we will sign up.
Now - in the process of registering, what will they say to me... I don't think they'll say no...
In general, of course, I am saddened by position of many, especially newbies (this is about another forum, where I made the same proposal).
I can well understand programmers who may be interested in the idea, but simply do not have time to deal with it.
But I completely do not understand the position of people who, on the one hand - sort of looking for ideas "how to make money", but while accusing me of wanting to "fool" others, while no one clearly said what the "fooling" is.
As I said before - the meaning of most of the answers is exactly: "You're a fraud, you want our money, we really don't understand how, and in general we don't understand the catch, but you won't get a penny of our money ..." You explain them - I don't need money, I need help with computing resources, for which I am ready to share with the experts, which turned out to be the best, that is exactly the same, someone has already optimized before.
Consider, I offer you - GUARANTEED good result of optimization. What you optimize - not the fact that it will not immediately go down. But at the same time - you will take an expert who has more than one day to show good results! We do not know what will happen next. But the thing that you have just optimized is unclear how it will work, and the thing that I suggest has shown good results not only on optimization, but also on a demo account for several days. Is it a bad suggestion ?
In response, as a rule, you start speculating about "income guarantees". Like "if you offered us a silver bullet on the security of real estate - we would still try to optimise ... but as it is - no desire".
Well... With such sentiments - is it possible to get something ?...
That's what I'm doing !
And, understandably, I'll be replacing anyone who shows 'control shots'.
And there will be a signal, from the real $100. I have already had a few suggestions in private with the meaning of "we do not want to bother with your proposal, open the signal - if it's good - we will sign up.
I'm in the process of registering, see what the answer is... I don't think they'll say no...
In general, of course, I am saddened by position of many, especially newbies (this is about another forum, where I made the same proposal).
I can well understand programmers who may be interested in the idea, but simply do not have time to deal with it.
But I completely do not understand the position of people who, on the one hand - sort of looking for ideas "how to make money", but while accusing me of wanting to "fool" others, while no one clearly said what the "fooling" is.
As I said before - the meaning of most of the answers is exactly: "You're a fraud, you want our money, we really don't understand how, and in general we don't understand the catch, but you won't get a penny of our money ..." You explain them - I don't need money, I need help with computing resources, for which I am ready to share with the experts, which turned out to be the best, that is exactly the same, someone has already optimized before.
Consider, I am offering you - GUARANTEED good optimization result. What you optimize - not the fact that it will not immediately go down. But at the same time - you will take an expert who has more than one day to show good results! We do not know what will happen next. But the thing that you have just optimized is unclear how it will work, and the thing that I suggest has shown good results not only on optimization, but also on a demo account for several days. Is it a bad suggestion ?
In response, as a rule, you start speculating about "income guarantees". Like, "if you offered us a silver bullet on the security of real estate - we would still try to optimise ... and so there is no desire.
Well... With such sentiments - is it possible to get something ?...
I used to develop a similar technology, but then abandoned it, because it was better. That is why I am interested in what you may get. Perhaps it would be better to automate the process of searching for strategies and combine all of them in one Expert Advisor? Trade without optimization, in real time. That is, a system is taken, its random parameters are set, and there are hundreds of such systems. Further, we eliminate the systems and parameters that bring losses, and those that brought a profit create their copies with slightly changed parameters +- from the current ones. And so on we continue. Train them on a small interval of history and add them to the account as is. From then on they will breed on their own.
I did the same thing with wagons for simplicity. I had 1000 robots with trading robots with random parameters like averaging period, profit and loss size, trailing percentage parameters for profit and trailing loss. It allows making the wagons turn profit on the 4-year history. Although with one caveat, with a 1 pip spread. But the robot on the wipers without self-learning mode failed to turn profit at all on a 4 year interval.
I used to develop a similar technology, but then I gave it up because it was better. That is why I am interested in what you will get. Perhaps it would be better to automate the process of searching for strategies, to combine them all into one Expert Advisor? Trade without optimization, in real time. That is, a system is taken, its random parameters are set, and there are hundreds of such systems. Further, we eliminate the systems and parameters that bring losses, and those that brought some profit create their copies with slightly changed parameters +- from the current ones. And so on we continue. Train them on a small interval of history and add them to the account as is. From then on they will breed on their own.
I did the same thing with wagons for simplicity. I had 1000 robots with trading robots with random parameters like averaging period, profit and loss size, trailing percentage parameters for profit and trailing loss. It allows making the wagons turn profit on the 4-year history. Although with one caveat, with a 1 point spread. But the robot on the wipers without self-learning mode never managed to get it to profit within 4 years.
Basically, that's what my system is.
Only I doubt the effectiveness of random selection. There are too many variants. Even for Expert Advisors with the minimum parameters I get several thousand variants. And for the majority of them - the number of possible combinations of parameters is in the billion, and I reduce their possible values (for example, I have periods from 3 to 330 divided into 25 points located logarithmically at the range of values). And I can trade even with ten thousand bots - it is not certain that a bot that will be close to the optimum will appear among them.
In addition, the result of the work will not be so quickly visible.
Genetic selection is much more productive. Therefore, the general idea is the same, but the selection of the best parameters - I still spend in the tester.
Hmmm... Not one TC has died... That's suspicious...
Today - I'll post another stats on the "favourites".