Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We鈥檒l occasionally send you account related emails.

Already on GitHub? Sign in to your account

FinRL on Synthetic Indices #357

Open
Tanakablack opened this issue May 6, 2024 · 3 comments
Open

FinRL on Synthetic Indices #357

Tanakablack opened this issue May 6, 2024 · 3 comments
Assignees

Comments

@Tanakablack
Copy link

Firstly, thank you all guys for all the hard work you are putting into this project, its really awesome and user friendly even to newbies like me. I have been running a few notebooks with ease ofc with some small errors haha 馃槃. Ok, i wanted to bring up a new idea: Can we consider synthetic indices?

Synthetic indices are simulated stock markets offered at deriv.com. They operate all day every day and mimic real market movements, but without being affected by unexpected news or world events. This makes them a good option for rl since they offer a simpler and more predictable environment for the models to learn from.

Pros
They are a bit consistent and always available 24/7 , which means the historic data is continuous without interruptions on weekends.

They are 100% technical, since the price is determined by PRNGs to simulate real life markets they have no fundamentals so they don鈥檛 have sudden movements caused by real-world events, so it's easier for our models to learn the patterns.

Cons
The data in bigger timeframes like Daily is still small compared to stock/forex markets since they all start from 2019, however the data in intraday is relatively huge.

This is not a big market compared to Forex, Stocks or Crypto so the returns will be significantly low, and they are offered at deriv.com only (which can close anyday).

Conclusion
I have been trading these synthetic indices for a while and even though the price is determined by PRNGs they are still quite predictable and follow basic market rules like trendlines, support & resistance. From my own point of view, they are quite predictable in intraday but random in long term.

Also, i have been working with some notebooks (highly appreciated) like for cryptocurrencies but it is a bit outed, and there are no forex related notebooks.

I believe what we will learn from synthetic indices could help us build a strong foundation for technical analysis before we move on to more complex markets like stocks, cryptocurrencies and forex.

Looking forward to hearing your thoughts on this.

@zhumingpassional
Copy link
Collaborator

good suggestions.
does it work in your scenario?
could you pls submit a PR (notebook or python) in finrl?

@zhumingpassional zhumingpassional self-assigned this May 7, 2024
@Tanakablack
Copy link
Author

Hie @zhumingpassional , thanks for your response

I haven't tested this idea yet, these indices are closer to forex than stocks so the Crypto notebook environment might be the closest we can use to test this out. However the notebook seem to have some bugs or outdated.

Please check this tutorial notebook:
https://github.com/AI4Finance-Foundation/FinRL-Tutorials/blob/master/3-Practical/FinRL_MultiCrypto_Trading.ipynb

AI4Finance-Foundation/FinRL-Tutorials#61

I was thinking if I manage to run this notebook without errors, I can then replace the dataset with the historic data I have for synthetic indices, I also noticed the notebook doesn't include custom indicators yet, but we can do this later.

@Tanakablack
Copy link
Author

@zhumingpassional Please advise.....

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants