Using artificial intelligence at MTS - page 25
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
Folks, please advise how.... during optimization for a certain time interval different results are obtained.... it means that the genetic algorithm each time selects a new path of genetic development :-). Anybody faced such a nonsense? Because of these problems, it is impossible to get statistics on strategy success....
I don't understand how you can seriously discuss a strategy that yields 44 deals in 2 years... Too few statistics!
There is a lot more that I don't understand.
It's not about the trades, it's about the principle. As far as I have heard, NS works on stationary markets, but on stationary markets there is no problem making a TS with a good forecast. But the market is not stationary. For this market NS loses a lot to any statistical approach. In the latter you can always tell what is decided, what can be decided and what cannot be decided yet. NS is a black box, there is no market model, there is no possibility in principle to at least try to answer the question of confidence intervals of the results obtained. Maybe I don't understand something? What is the beauty of NS? Here NS has learned something. What did it learn? To get the profit factor? Bullshit. It's not the profit factor that's the problem, it's the credibility of the result. After all, this is not the first forum on NS, all waiting for NS apologists to give a link to '59 - NS is hardwired to history, and the marketplace does not see the future well from history.
... and on the market the future is poorly visible from history.
But what about this "lure" :)
Technical analysis is based on three assumptions:
1. Market movements take everything into account.
2. Prices move in a directional way.
3. History repeats itself.
I myself have started to look into NS, it's interesting, but I don't see any practical application yet.
Well, what about this "lure" :)
Technical analysis is based on three assumptions:
1. The market movement takes into account everything.
2. Prices move in a directional way.
3. History repeats itself.
I myself have started to look into NS, it's an interesting thing, but I can't see any practical use for it yet
I agree. Essentially the same TA, but TA has at least some detail, you can go either way, but the level is the same - there is no assessment of confidence in the result.
I agree. Essentially the same TA, but TA has at least some details, you can do it both ways, but the level is the same - there is no assessment of confidence in the result obtained.
As I have been dealing with Forex for a short time and have been testing my own strategies for even less time, I realize that the basis for profitable trading is money management, namely MM (Money Management) and not MR (Management Risk), i.e. when to stop losses or take profit, in the Kodobase forum they mostly use MR instead of MM - i.e. which part of the deposit to risk in order to take a possible profit
But it is important to predict not the price (many people want to see it in the NS), but the effective SL levels or effective time of transaction (time for taking a decision), this is what I would like to have with the help of the NS.
I've been trading not so long, I've been testing my strategies even less, but I get more and more sure that the basis for a profitable trading is MM (Money Management), not MR (Management Risk), i.e. when to stop losses or take profit, on Kodobase forum they mostly use MR instead of MM - that is which part of the deposit to risk in order to get a possible profit
I can conclude, in addition to the above mentioned, that the selection of indicators used for trading is not so important, it is important to detect a signal to enter the market, was it true or false, it is clear that the account equity will show what we have, but it is important to forecast not the price (many people want to see it in VS), but the effective SL levels or effective time of transaction (time for making a decision), that is what I would like to use VS.
According to my understanding, it is the forecast in NS that is basically impossible.
According to my understanding, it is impossible to make a forecast in NS.
I am not looking for a forecast, but online performance evaluation/calculation may be possible for NS to show, say, an online strategy tester
look at http://alife-soft.narod.ru/programs.html, there are many ready-made programmes
...
1) Well, I would like to conclude that the selection of indicators used for trading is not so important, it's important to determine the signal for market entry to be true or false, it is clear that the equity of the account will show what we already have,
2) But it is not the price that should be predicted (many people want to see it in the NS), but effective SL levels or effective time of a deal (time for making a decision), it would be better to use the NS.
1) You can enter the market at any time. It is not important to know the entry point, but the exit level (not the exit point!). I will tell about time in step 2
2) Predicting the price is probably the most useless thing on the market that can be created (i.e. predicting the price at a predetermined moment in time).
But it is quite possible to set the SL level (keeping in mind that nobody forbids to enter the market at any point in time, besides, all points in time for entry are absolutely equal). Based on the decision to set the SL, it remains to calculate the lot size of trading operations, and set TP, depending on the strength of the signal. Do not forget about the fact that any "Pronosis" should have an expiration date, hence the position must be forcibly closed at the end of its "lifetime". With this approach to trading - ANN is the best and most effective tool that I know.
There is no need to demand from networks what they cannot do. There is no need to ascribe mystical properties to them. You need to set adequate objectives for the networks (as well as for yourself), and it has nothing to do with non-stationary markets, that would expect a return in the form of bread with butter and caviar.
PS IgorM, don't take my post personally. It has a general character.
PS IgorM, don't take my post personally. It is of a general nature.
On what personal account? Thank you very much, you have in a few lines confirmed, even more - directed me in the right direction of my research in terms of the construction of my TS
1) You can enter the market at any time. What matters is not the entry point, but knowing the exit level in advance (not the exit point!). I will tell you about time in point 2.
2) Predicting the price is probably the most stupid thing you can do in the market (i.e. predicting the price at a predetermined moment in time), whether using networks or any other method/instrument.
But it is quite possible to set the SL level (keeping in mind that nobody forbids to enter the market at any point in time, besides, all points in time for entry are absolutely equal). Based on the decision to set the SL, it remains to calculate the lot size of trading operations, and set TP, depending on the strength of the signal. Do not forget about the fact that any "Pronosis" should have an expiration date, hence the position must be forcibly closed at the end of its "lifetime". With this approach to trading - ANN is the best and most effective tool that I know.
There is no need to demand from networks what they cannot do. There is no need to ascribe mystical properties to them. You need to set adequate objectives for the networks (as well as for yourself), and it has nothing to do with non-stationary markets, that would expect a return in the form of bread with butter and caviar.
PS IgorM, don't take my post personally. It has a general character.
+100