Optimise an EA and get the best of the optimised ones. - page 7
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
As a result, the question from "what to do to make my TS work steadily" is transformed into the question "how to choose a TS that already works steadily".
Interesting approach. But how do you determine that the TS is already working stably? Based on the results of back-testing? But most likely it may be a good fit, coincidence. Forward testing? It is no different from the back-test.
Is there a possibility to see the results of live trading on the account? A signal, monitoring, or at the very least a Statment?
That's an interesting approach. But how do you determine that the TS is already working steadily? By the results of back-testing? But that's very likely to be a fit, a coincidence. Forward testing? It is no different from the back-test.
Is there a possibility to see the results of live trading on the account? A signal, monitoring, stats, at the very least?
Let's keep it on a need-to-know basis.
I'm telling you - here, the charts provided are live trading on the demo, in this case from December last year. You can see that these charts are not from the strategy tester, but from Excel. Constructed by analysis of trades.
All the TS have been tested on a yearly interval (although the evaluation methodology has changed slightly since then) Since this is not scalping, small slippages do not play a role. Even the recent awful slippage on CAD pairs - it just increased drawdown of several TPs with "Canadian", made profit on one of them and in a week already these deals have lost their results.
Watching the "total result" - makes no sense - because, as I have repeatedly said - periods of profit in TCs is shorter than periods of loss, and the number of losing TCs - also more profitable. As a result total deposit is constantly decreasing, and this is normal. I have a separate demo account with several TS, put at the end of last year, when I just formed the selection criteria. They are not losing, but they have nothing to brag about yet.
If you are paranoid and have doubts I can provide you with demo account number and its investment password for MT4:
Login: 9968945
Invest password: TCxG16r9
Server: Alpari-ECN-Demo
But I have to warn you right away, there are more than 250 TS working there now, the "outsiders" are constantly re-optimized, and it's impossible to figure out which one works how - manually. One can only make sure that charts, which are presented above - correspond to trades on magiks.
But, you can take any deal, watch the magik - and in the tester run EALeague on this magik.
Tell me, why don't you like the period when the Expert Advisor starts to lose?
After all, by reversing the trades, you will get almost as much profit as the System will "lose"...
That's an interesting approach. But how do you determine that the TC is already stable?
By the way, yes, suggestions on evaluating the stability of the TS are accepted.
Here, the TS is optimized, it is standing on a demo and trading with minimal lots.
How to choose the best of them to install on a real account with MM included ? Useful Suggestions - I can also thank you with a three-month restrcod on one of the TC-favorites.
I consider genetic optimisation over a year to be optimal. Backtest 5 months, forward 7, OHLC mode at 1M.
Such optimization on quad-core i5 requires two to four hours, depending on number of parameters. On single-core AMD Sempron LE-1200 it takes 20-40 hours.
The total counter of settings is not required since MT5 allows you to stop optimization and further start it from the point where it was stopped. I use it quite often.
2-4 hours is tolerable, of course.
Stopping and starting further is not acceptable, because it may be done to start optimization of your own TS.
Looked at the files at the link, without set files do not even want to spend time on this fiddling (to deal with settings for optimization).
Let me put it this way, I don't need TCs that are not clear how they work, I just have a chance to run optimization when computers are idle, on gratuitous terms, so to speak, but it's now, not forever. So if there is a specific set, clear optimization time and tool, there is certainty with TF, then I can run optimization. Yes, and specify the api of the server from which to take the story. By the way, why don't you just write the necessary information from optimization to a file, so you don't have to deal with saving optimization results?
2-4 hours is tolerable, of course...
Stopping and starting further is not acceptable, as it can be done to start optimising your TC.
Looked at the files at the link, without set files do not even want to spend time on this fiddling (to deal with settings for optimization).
Let me put it this way, I don't need TSs, which are not clear how they work.
Sets will be, and more detailed instructions for optimization will be. Just "League of Trading Systems" I deal with on the residual principle, and I don't always have time for it.
And about "it's not clear how they work", could you be more specific? On the previous page, I clearly indicated all the techniques used in the "League" and explained why there are 16 of them per symbol.
If you (let's refer to "you") want to have a look at TC code, you can try to do it through "joint projects". But I'm afraid it will not go further than look - I sometimes post pieces of my code, have you seen it? Can you figure it out? Do you need it like this?
Yes, and specify the IP of the server from which to get the history. By the way, why don't you just write the necessary information from optimization to a file, so you don't have to bother saving the results of optimization?
Above is the account number, server and investment password. It is on this account that all 260 (at the moment) of "League" TS are working.
And about "write to file" - why do I need to do that if the optimization results are exactly what I need?
There is a field "back", there is "forward" - I take the minimum of these two fields in Excel, and sorted in descending order. The TC, in which the minimum is the biggest (maximin) - this combination of parameters I "score" hard in the code of the TC.
In addition - the presence of standard XML-file for MT5 allows to understand, whether a person launched optimization correctly, I already feel that my project is interested primarily by newcomers, and I'll have to explain often how and what to do.
Yes, plus, commission and spread :) Reverse trades and the EA is still losing :) And why is that? Because the shift in the probability of winning +/- is approaching 50% :)
Wo!... And so, attention - a question: if your TS STABILIZES the probability around 50%, can't you make money out of it...?
There will be sets, and there will be more detailed instructions for optimisation. It's just that I deal with the Trading Systems League on a residual basis, and I don't always have time for it.
And about "it's not clear how they work" - can you be more specific? On the previous page, I clearly indicated all the techniques used in the "League" and explained why there are 16 of them per symbol.
If you (let's on "you") want to look at TC code - you can try to do it through "joint projects". But I'm afraid, it will not go further than look - I sometimes lay out pieces of my code, have you seen it? Can you figure it out? Do you need it like that?
When we have sets, then we'll talk :) My point is that everyone has work to do and if you run such a complicated project involving a team, you need to think carefully about its organization, including distributing between people what and when they should optimize. As for the reward, I would simply open a signal in a cent account and let the project participants copy it into their accounts. I have even decided not to remove my signals after the withdrawal of reward, although I have spent a lot of time and effort on TS.
I do not know how it works, and it's not even about the code - you can not tell at a glance what's what. If I have not understood the code it is not even the code - in this case I cannot immediately understand what is going on.
Above is the account number, server and investment password. It is on this account that all 260 (at the moment) of "League's" TS work.
And about "write to file" - why would I do that if the optimization results are exactly what I need?
There is a field "back", there is "forward" - I take the minimum of these two fields in Excel, and sorted in descending order. TC, in which the minimum is the highest (maximum) - this combination of parameters I "score" hard in the TC code.
Besides - the availability of standard XML-file for MT5 allows to understand, whether one has launched optimization correctly or not, I already feel that my project is interesting first of all for beginners, and I will have to explain how and what to do.
Without ipi server I can not select the specified server, I just do not have it in the lists, and the name will not be found, this is the message in the log
"
2018.03.17 16:50:43.675 Network '9968945': no connection to Alpari-ECN-Demo
"
About writing to file - it's just more convenient, purely for me, less gestures. I am optimizing on 4, I gather such files, and then a separate program processes them and at the output I get the result, from which I can make an analysis of the efficiency of the idea (a good idea is good for any basic settings). True, I also save the set file at once, so as not to get confused, what exactly has been optimized.
About explanations for beginners - make a video, where you show how to become a member of the project, what to configure and where to attach.
When we have the sets, then we'll talk :) I mean, everyone has something to do and if you run such a complex project involving a team, you need to think carefully about its organization, including the distribution of what and when people need to optimize. As for the reward, I would simply open a signal in a cent account and let the project participants copy it into their accounts. I even decided not to remove my signals after the withdrawal of reward, although I've spent a lot of time and effort on TS.
So "... people want to understand what's what, this is a new, untested business for us...".
An idea came up - so I put it out there. Now I'm solving "procedural issues".
About the signals - not quite clear. Here, look - now I have a number of favourites. Should I open a signal for each of them? At any time, any favorite may show a "control joint", and will be expelled from the "League". And the signal to close?
There are some best TS that work for me on pre-real - here, it will be possible to put this score as a signal. But then again - first we need to work out a technique for selecting the most stable TS. And so far I spend all my time on rework and over-optimization.