https://github.com/0xnu/drl-ts
It uses deep reinforcement learning to learn market patterns and automatically make buy, sell, or hold decisions, training itself by testing strategies and learning from outcomes.
https://github.com/0xnu/drl-ts
deep-learning deep-reinforcement-learning trading trading-algorithms trading-strategies trading-systems
Last synced: about 1 month ago
JSON representation
It uses deep reinforcement learning to learn market patterns and automatically make buy, sell, or hold decisions, training itself by testing strategies and learning from outcomes.
- Host: GitHub
- URL: https://github.com/0xnu/drl-ts
- Owner: 0xnu
- License: other
- Created: 2025-08-28T09:58:52.000Z (about 1 month ago)
- Default Branch: main
- Last Pushed: 2025-09-04T11:05:31.000Z (about 1 month ago)
- Last Synced: 2025-09-06T10:54:51.859Z (about 1 month ago)
- Topics: deep-learning, deep-reinforcement-learning, trading, trading-algorithms, trading-strategies, trading-systems
- Language: Jupyter Notebook
- Homepage:
- Size: 2.3 MB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- Changelog: CHANGELOG.md
- Contributing: CONTRIBUTING.md
- License: LICENSE
- Code of conduct: CODE_OF_CONDUCT.md
- Codeowners: CODEOWNERS
- Security: SECURITY.md
Awesome Lists containing this project
README
## Deep Reinforcement Learning (DRL) Trading System
[](https://github.com/0xnu/drl-ts/releases/latest)
[](/LICENSE)
It uses [deep reinforcement learning](https://en.wikipedia.org/wiki/Deep_reinforcement_learning) to learn market patterns and automatically make buy, sell, or hold decisions, training itself by testing strategies and learning from outcomes.
> [!WARNING]
> Disclaimer: It is an experimental system; do not use it for actual investment decisions.
> Checkout this [repository](https://github.com/0xnu/deep-reinforcement-learning) if you want to learn more about Deep Reinforcement Learning.
### Interactive Brokers
Here's a guide for setting up Interactive Brokers for the Deep Reinforcement Learning Trading System:
#### Create a Paper Trading Account
+ **Visit Interactive Brokers**: Go to [www.interactivebrokers.co.uk](https://ibkr.com/referral/abiodun557) (for UK)
+ **Open Account**: Click "Open Account" and select "Individual"
+ **Choose Paper Trading**: During setup, ensure you enable paper trading (demo account)
+ **Complete Application**: Fill in all required information
+ **Fund Account**: Paper trading accounts come with virtual money (usually $1M USD)
#### Download and Install Trader Workstation (TWS)
+ **Download TWS**: Go to Client Portal → Trading Platforms → Download TWS
+ **Install**: Run the installer (available for Windows, Mac, Linux)
+ **Alternative**: You can also use IB Gateway (lighter version) instead of full TWS
#### IB Gateway (Recommended for API)
+ **Download**: Client Portal → Trading Platforms → Download IB Gateway
+ **Lighter Option**: Gateway is better for API-only usage (no GUI trading interface)
#### Configure API Settings
##### Enable API Access
+ Login to TWS/Gateway
+ Go to Configuration:
+ TWS: File → Global Configuration → API → Settings
+ Gateway: Configure → Settings → API
+ Enable API:
+ Enable ActiveX and Socket Clients
+ Allow connections from localhost only (127.0.0.1)
+ Read-Only API (for safety during testing)
+ Set Socket Port:
+ Paper Trading: 7497 (default)
+ Live Trading: 7496 (use only when ready for live trading)
+ Client ID: Set to 0 or any number (match this in your Python code)
#### Important Security Settings
```sh
✅ Enable API
✅ Read-Only API (initially)
✅ Allow connections from localhost (127.0.0.1) only
❌ Don't enable "Download open orders on connection"
❌ Don't enable "Allow connections from anywhere"
```
#### Market Data Permissions
##### Enable Required Data
+ **Go to**: Account Management → Market Data Subscriptions
+ Enable:
+ US Securities Snapshot and Futures Value Bundle (if trading US stocks)
+ UK Securities (if trading UK stocks)
+ Real-time quotes for your target markets
+ **Note**: Some data feeds have monthly costs, but basic delayed data is often free
#### Workflow for Using with DRL Trading System
Daily setup process:
+ Start TWS/Gateway:
```sh
# Make sure TWS or Gateway is running
# Login with your paper trading credentials
```
+ Verify API settings:
+ Port: 7497 (paper trading)
+ API enabled
+ Correct client ID
+ Run DRL Trading System:
```sh
# In the notebook/script
if ib_conn.connect():
print("✅ Ready for trading!")
# Trading code here
```
#### Testing Checklist
Before running the DRL trading system:
- [x] TWS/Gateway running and logged in
- [x] API settings configured correctly
- [x] Market data permissions enabled
- [x] Test connection script works
- [x] Can retrieve historical data
- [x] Can place paper orders (test manually first)
#### Common Issues and Solutions
Connection problems:
```sh
# If connection fails, check:
# 1. Port number
ib.connect('127.0.0.1', 7497, clientId=1) # Paper trading
# NOT 7496 (live trading)
# 2. Client ID conflicts
# Try different client IDs: 1, 2, 3, etc.
# 3. API not enabled
# Check TWS: File → Global Configuration → API → Settings
```
#### Market Data Issues
+ **No real-time data**: Enable market data subscriptions
+ **Delayed data**: Normal for free accounts, 15-20 minute delay
+ **Weekend/holidays**: Markets closed, no live data available
#### Interactive Brokers Market Data
Interactive Brokers (IBKR) requires you to maintain a minimum balance of $500 USD (£370 GBP) in your funded account to [subscribe to and receive real-time market data](https://www.interactivebrokers.com/en/pricing/market-data-pricing.php?p=mktDataPricing), not as a one-time fee for data access. This $500 (£370 GBP) balance is a maintenance requirement, in addition to the monthly fees for the specific market data feeds you choose to subscribe to. Whilst basic delayed data may be available, you must have a funded account and select specific data subscriptions to receive live, timely information for trading.
You need to subscribe to these market data for the Trading Systems (DRL and Non-DRL) to work properly:
+ NYSE (Network A/CTA)(L1) - $1.50 / month (£1.11 / month)
+ NASDAQ (Network C/UTP)(L1) - $1.50 / month (£1.11 / month)
#### Paper Trading vs Live Trading
```sh
# Paper Trading (Safe for testing)
PAPER_TRADING_PORT = 7497
# Live Trading (Only when ready!)
LIVE_TRADING_PORT = 7496
# Always start with paper trading!
ib.connect('127.0.0.1', PAPER_TRADING_PORT, clientId=1)
```
#### Safety First
+ Always test with paper trading first
+ Never use live trading until system is thoroughly tested
+ Monitor all trades manually initially
+ Start with small position sizes
Once you complete this setup, **DRL Trading System** will be able to connect to Interactive Brokers and execute paper trades safely for testing and validation.
### Install Dependencies
Use the package manager [pip](https://pip.pypa.io/en/stable/) to install required libraries:
```sh
# Prerequisites
python3 -m venv .venv
source .venv/bin/activate
uv pip install -r requirements.txt
```
### Run the System
+ **DRL Trading System**:
+ Click the `Run` button to run the individual cell of the [interactive brokers deep reinforcement learning](./notebooks/drl-ts.ipynb) notebook.
+ **Day Trading System**:
+ Click the `Run` button to run the individual cell of the [interactive brokers day trading](./notebooks/ib-tb-without-drl.ipynb) notebook.
+ **Forex Trading System**:
+ Click the `Run` button to run the individual cell of the [interactive brokers forex](./notebooks/ib-forex.ipynb) notebook.
> Do not run the cell that initializes the Interactive Brokers Live Trading (Live Trading Integration).
### License
This project is licensed under the [Modified MIT License](./LICENSE).
### Copyright
(c) 2025 [Finbarrs Oketunji](https://finbarrs.eu). All Rights Reserved.
---
My non-negotiable fee for a custom enterprise trading system is £150,000. [Let's talk](mailto:f+trading@finbarrs.eu) about your requirements.