Topic: Inconsistency of results!!!

I created a portfolio of 15 robots!! And the results in FSB Professional were very positive!!! NOT one losing day!!! I put these robots on a real account and they have been losing for 5 days. What's the matter?

But there is one thing: these robots did not pass the Monte Carlo test.

How can I create robots so that the results of real trading coincide with the results of FSB.

Post's attachments

Снимок экрана (527).png 97.14 kb, 2 downloads since 2023-10-29 

You don't have the permssions to download the attachments of this post.

Re: Inconsistency of results!!!

The robots were created for M1. And the history of 20,000 BAR premium FSB quotes was used!

3 (edited by Naya 2023-10-29 16:26:05)

Re: Inconsistency of results!!!

Be sure to study the software quite well.
Achieving excellent results with fsb pro in Backtest does not require much skill.
I will suggest

1)  create strategies for H1 and higher time-frames only...but preferably H1. Ensure to use at least 20000bars

2) if you will use m1 timeframe, then use at least 300000bars or more.

3) in any case make sure you do many curve fitting checks like Monte-Carlo and multimarket test. Search through this forum and other websites for curve fitting in trading strategy development and how to reduce it(eliminating it is impossible)
4) learn, test, correct, relearn and keep working till you find something stable and profitable and stick with it.

NB: the most stable systems are the simplest!!!

Re: Inconsistency of results!!!

Hello!!!Thank you for your answer!!!I am very pleased!!!What are your achievements!!!Do you have profitable trading systems?

Naya wrote:

Be sure to study the software quite well.
Achieving excellent results with fsb pro in Backtest does not require much skill.
I will suggest

1)  create strategies for H1 and higher time-frames only...but preferably H1. Ensure to use at least 20000bars

2) if you will use m1 timeframe, then use at least 300000bars or more.

3) in any case make sure you do many curve fitting checks like Monte-Carlo and multimarket test. Search through this forum and other websites for curve fitting in trading strategy development and how to reduce it(eliminating it is impossible)
4) learn, test, correct, relearn and keep working till you find something stable and profitable and stick with it.

NB: the most stable systems are the simplest!!!

5 (edited by Naya 2023-10-29 16:38:30)

Re: Inconsistency of results!!!

I do have some 'stable' systems but still in evaluation phase. I HAVE NO INTENTION OF SHARING YET.
Created them with my custom indicators.
The only optimizable parameters in my systems are stops or profit levels to adapt to volatility.
I noticed that generating or creating strategies with many numerical parameters amplifies exposure to curve fitting.

Re: Inconsistency of results!!!

Do you have a positive experience in trading robots! I'm trying to create robots, but greatness does not come through Monte Carlo

Naya wrote:

I do have some 'stable' systems but still in evaluation phase. I HAVE NO INTENTION OF SHARING YET.
Created them with my custom indicators.
The only optimizable parameters in my systems are stops or profit levels to adapt to volatility.
I noticed that generating or creating strategies with many numerical parameters amplifies exposure to curve fitting.

Re: Inconsistency of results!!!

There is no fool proof tool or test as far as I know.
Monte-Carlo has it's use and some Monte Carlo tests work for some strategies and fail on some.(I mean some Monte Carlo tests are useless for certain strategies and those strategies may pass all tests or fail all test and it will mean absolutely nothing).
I can't point you to a perfect system or a perfect way of developing one. Just learn as much as you can and practice more. There is so much you can learn by doing a word search through this forum and you will learn from the wealth of experience of other users.

With time, With study and  practice and you may stumble on systems that have over 20years stability and will pass most robustness tests...you may just stick with it and merely optimize from time to time.
An ideal system should perform acceptably going forward after optimization and yet not fall apart in backward out of sample.

Re: Inconsistency of results!!!

Kamil Shakirov wrote:

I created a portfolio of 15 robots!! And the results in FSB Professional were very positive!!! NOT one losing day!!! I put these robots on a real account and they have been losing for 5 days. What's the matter?

But there is one thing: these robots did not pass the Monte Carlo test.

How can I create robots so that the results of real trading coincide with the results of FSB.

Eu sugiro que você ao Export o ia use o próprio testador do mt5 ou mt4 para refinar de acordo com as condições da corretora funciona bem pra min

Re: Inconsistency of results!!!

Kamil Shakirov wrote:

Do you have a positive experience in trading robots! I'm trying to create robots, but greatness does not come through Monte Carlo

Naya wrote:

I do have some 'stable' systems but still in evaluation phase. I HAVE NO INTENTION OF SHARING YET.
Created them with my custom indicators.
The only optimizable parameters in my systems are stops or profit levels to adapt to volatility.
I noticed that generating or creating strategies with many numerical parameters amplifies exposure to curve fitting.


Tente sempre usar a otimização walk foward os robôs criados com ele com um bom desempenho normalmente passa em Monte Carlo tente usar pelos menos 3 meses de dados e um timeframe maior como m30 ou h1