collapse

Author Topic: Visual-based strategy buiding  (Read 1148 times)

Offline FxTS

  • Jr. Member
  • **
  • Posts: 30
Visual-based strategy buiding
« on: February 06, 2019, 06:24:48 PM »
Dear community,

I would like to ask an interesting question since I have got an opportunity to be engaged in a project, connected with trading software creation of tool that allows to construct EAs without any coding required.

The idea is quite simple: you just drag and drop the needed elements and receive a coded EA in two formats (both for the backtesting and trading MT4/MT5 platforms) that is ready to be downloaded. As for me, it will make the process of strategy development easy and comfortable.

The project is at the initial MVP stage and at this moment, so I would like to hear your expert opinion about all these things. Have you ever used any tool allowing to construct strategies in a way, similar to this? If so, what were the results? Have you faced any difficulties while using it? What were the advantages or disadvantages of the tool youve used?

Feel free to share any idea that pop up in your mind

*Unfortunately, I`m not allowed to mention the name of the project, company, or any details due to the NDA, so lets just discuss the concept in general.
Education, Exploration, Experimentation

Offline FLechdrop

  • DonbonsPrivateGroup
  • Sr. Member
  • *
  • Posts: 604
Re: Visual-based strategy buiding
« Reply #1 on: February 06, 2019, 06:29:32 PM »
Yes, some of us use EA Lab: https://www.tradeview.com.au/product/ea-lab/

This sounds like more or less the same as what you are proposing. You could try to compete with them on both quality and price.

Offline FLechdrop

  • DonbonsPrivateGroup
  • Sr. Member
  • *
  • Posts: 604
Re: Visual-based strategy buiding
« Reply #2 on: February 06, 2019, 06:32:15 PM »
As for quality: it seems to me that if you know how to code you could do quite a bit more than with EA Lab currently. I would be interested in anything to do with machine learning.

As for price: they charge quite a lot, including a monthly fee for using the software (but you can get that free with IC Markets) and an additional monthly fee for support. Also quite hefty prices for courses teaching you how to actually use it.

Offline Nasdaq100

  • Full Member
  • ***
  • Posts: 400
Re: Visual-based strategy buiding
« Reply #3 on: February 06, 2019, 06:39:01 PM »
The idea is brilliant

However, prices for both using and learning is the major downside in my opinion. Because it will simply mean that there will
be no open public workshops where everyone can participate unfortunately.

Offline FxTS

  • Jr. Member
  • **
  • Posts: 30
Re: Visual-based strategy buiding
« Reply #4 on: March 01, 2019, 08:49:30 PM »
Machine learning is a great technology, but it requires experinced specialists to build it at the high level. At the same time, it would be more useful for tools purposed to backtesting or something like that, while this tool is just to turn your idea into the format the machine will understand.

The idea is to create user-friendly interface that will allow to operate with it without any previous education. The user will choose the elements of the strategy from the list, place them in the proper order and the platfrom would generate him a code expressing that idea.
For example. there would be several places for elements to fill. Let`s take, for example, MA crossover. It will have the following:
1. Element one - "fast" MA, for instance 9EMA
2. Situation - crosses upwards
3. Element two - slow MA like 200 SMA
4. Action required - Buy 1 lot, SL .... TP ....

The user will be able to drag and drop all these element like "Moving average", "RSI" and other from one list and situation from another and compose the strategy.
Education, Exploration, Experimentation

Offline Humble Trader's Fx

  • Global Moderator
  • Sr. Member
  • *****
  • Posts: 1497
Re: Visual-based strategy buiding
« Reply #5 on: March 01, 2019, 10:53:01 PM »
Machine learning is a great technology, but it requires experinced specialists to build it at the high level. At the same time, it would be more useful for tools purposed to backtesting or something like that, while this tool is just to turn your idea into the format the machine will understand.

The idea is to create user-friendly interface that will allow to operate with it without any previous education. The user will choose the elements of the strategy from the list, place them in the proper order and the platfrom would generate him a code expressing that idea.
For example. there would be several places for elements to fill. Let`s take, for example, MA crossover. It will have the following:
1. Element one - "fast" MA, for instance 9EMA
2. Situation - crosses upwards
3. Element two - slow MA like 200 SMA
4. Action required - Buy 1 lot, SL .... TP ....

The user will be able to drag and drop all these element like "Moving average", "RSI" and other from one list and situation from another and compose the strategy.

I have been following machine learning (AI) for some time now,  especially the use of Python.  I can almost tell you with some certainty, that BIG BANKS have already created algorithms which make them more millions upon millions. As with the stock market,  it's now the bots that are competing,  we are just nibbling at the crumbs with our simplistic EA'S at best. But no matter what, there is always space for creativity and persistence.

Imagination is the way.;)

Regards,
HumbleTrader
We humbly approach the Forex Market and take only what is earned through our hard work and intelligence.


Offline FLechdrop

  • DonbonsPrivateGroup
  • Sr. Member
  • *
  • Posts: 604
Re: Visual-based strategy buiding
« Reply #6 on: March 02, 2019, 11:22:55 AM »
Machine learning is a great technology, but it requires experinced specialists to build it at the high level. At the same time, it would be more useful for tools purposed to backtesting or something like that, while this tool is just to turn your idea into the format the machine will understand.

The idea is to create user-friendly interface that will allow to operate with it without any previous education. The user will choose the elements of the strategy from the list, place them in the proper order and the platfrom would generate him a code expressing that idea.
For example. there would be several places for elements to fill. Let`s take, for example, MA crossover. It will have the following:
1. Element one - "fast" MA, for instance 9EMA
2. Situation - crosses upwards
3. Element two - slow MA like 200 SMA
4. Action required - Buy 1 lot, SL .... TP ....

The user will be able to drag and drop all these element like "Moving average", "RSI" and other from one list and situation from another and compose the strategy.

I have been following machine learning (AI) for some time now,  especially the use of Python.  I can almost tell you with some certainty, that BIG BANKS have already created algorithms which make them more millions upon millions. As with the stock market,  it's now the bots that are competing,  we are just nibbling at the crumbs with our simplistic EA'S at best.

Exactly. Which I am suggesting it would be helpful to go in that direction. The current proposal sounds like another EA Lab. It could work, if you can compete with them on price or give more options. But it is not something that does not exist yet.

Offline FxTS

  • Jr. Member
  • **
  • Posts: 30
Re: Visual-based strategy buiding
« Reply #7 on: March 08, 2019, 10:36:06 AM »
As far as I know, large institutional market participants use AI-based software either for corellation-based arbittration strategies or like HFTs making profit on relatively small movements (something like scalping). But this is moslty about trading, while the purpuse of product I`ve mentioned above is create strategies.

By the way, which tools and features you would like to have during strategy building process? If I got the idea properly, there should be a possibility to use AI to create new strategies, am I right?
Education, Exploration, Experimentation

 

browse forum