Stock trading in Elixir

I am interested. I am learning this language to build personal stock trading platform.

I think that @CinderellaMan will also be interested

@PJUllrich Indeed, I would be very interested - especially in those indicators - I would totally steal them

@rm-rf-etc TLDR: Please release it, many people will benefit from it but it’s hard work to get contributions.

On a more serious note, I released a cryptocurrency trading “environment” last year and I believe that some people are/were using it but you need tons of marketing, introduction materials and tutorials, and “stuff” to get people excited to the extent of contribution. At least for me, it didn’t work out this way (and I had similar hopes that people will contribute as they will be using it and we will make it better together).

I think there are multiple reasons for it:

• some people just want to use a product (they don’t care about the code so much, Elixir makes it easier for them to modify it a bit and that’s it)
• some people are overwhelmed by the scale of the project (they need more dev docs etc - do you have any of those?)
• some people aren’t interested in the product, but they want to understand it for learning purposes - again you need even more dev docs

I would love to see more algo trading related stuff being released to the community as it would enforce the idea that this is what Elixir is good at, but it’s probably too early to expect people to contribute (it requires lots of non-coding related effort from you).

For me, to get people interested I gave up on docs etc - I started to record youtube videos and I can see people being interested far more than in “finished” code - maybe that’s the way to get people excited to do it - through videos develop system up to the level when people feel comfortable to contribute as they fully understand the design decision behind it.

I hope that this didn’t come across negative - I think we have a great Elixir community and chance here are much better to extend your project than in any other language.

First of all, thank you for all the effort and wonderful videos.

Personally I second creating video tutorials. Some people find it a lot easier to see things happening, than to read a part of a document/book.

This is just my opinion of course but my students have similar opinions. (Or I’m just a really bad writer, that’s also possible)

Thanks for the replies everyone. @CinderellaMan, I don’t take it as negative. Just sounds realistic.

I’ll clean up a few things and post it here.

I haven’t added any docs yet, there’s just unit tests.

@subhrashisdas I haven’t added any sort of frontend yet, but I do want one eventually. I would probably build it as a client app so it could run independently. And I’ll probably use Scenic for that (it seems amazing).

Here’s just the indicators.
https://github.com/rm-rf-etc/trade-indicators

Please feel free to provide suggestions for improvements/changes/etc. I’m still fairly green with Elixir.

I also have code for downloading data from Webull and Alpaca/Polygon, and I hope to eventually have everything needed for backtesting, forward testing, scanning, live trading, and driving a client frontend. Those parts are still half-baked at the moment.

Right now I’m working on a module to do pattern detection based on reversal points.

I’m working on something similar right now, it’s a portfolio analytics app which offers backtesting for custom investment portfolios against 20yrs historical data. I could definitely help you out with the backtesting and frontend work if you made it opensource.

Hi @GunnarPDX. Great! I will. Do you have anything public yet I could review?

@CinderellaMan I was looking at igthorn just now. Seems like there’s a good amount of overlap of objectives between my project, your project, and GunnarPDX’s project. If we built a common core, we could probably all collaborate to help us each other realize our individual projects faster.

For my project, I’m not focused on crypto, but I’m also not opposed to it. I’d propose a common core being an engine for ingesting and streaming back OHLC data to support any variety of real-time use, whether that’s a trading platform GUI or a trading bot. The engine would have an API so any variety of client could be built for it. I’m planning on using an approach like this anyway since I’d like to run a client app on a separate machine using scenic. So for my project, the core would be running in a bot on a server, and the client would connect remotely from my laptop. The client would also support multiple connections so I could manage multiple bots.

I imagine the engine would have a state struct for everything that’s in memory, and for the purpose of either rendering a longer term chart, or for performing backtesting, we would stream from the DB… I could actually use some help on this, I don’t have a lot of experience in the area of managing large time series datasets. I’ve been using timescaleDB BTW.

For supporting crypto, forex, stocks, commodities, what have you, that would all be built as abstractions… somehow

Well that’s my proposal anyway. Let me know if that would align with your approach. Thanks.

Just sent you a PM with some stuff to check out. It sounds like I may have ran into similar issues with the time-series data, I just used Postgres and managed to make things work pretty well, though TimescaleDB should make things easier by the looks of it. Let me know what you think, I’d love to contribute.