Bitcoin Scalping Bot
Rule-based BTC paper trading experiment with live Coinbase price feeds and a separate Python backend.
Problem
I wanted to see how automated trading actually works: exchange APIs, reacting to price moves, position sizing, logging results. Without putting real money on logic I had not tested yet.
Solution
Two related implementations: a browser-based paper trading dashboard on this site, and a local Python/Flask backend that runs the same style of rule-based strategy against Binance market data.
How It Works
- Entry: when BTC drops $50 below a rolling reference high, enter a long paper position.
- Exit: sell when price rises $50 above entry, or hit a 0.15% stop loss.
- Position sizing: up to 20% of balance, capped at $100 per trade.
- Live data: Coinbase REST spot price + WebSocket feed; Chart.js candle chart from Coinbase Exchange API.
- Logging: trade history and strategy state displayed on the dashboard.
Tech Stack
Challenges
- Keeping live price updates smooth when WebSocket drops. REST polling as fallback.
- Separating the public paper-trading demo from any real API keys or live trading logic.
- Strategy tuning: fixed dollar thresholds behave differently at varying BTC price levels.
What I Learned
Most of the work was plumbing: price feeds, state management, position sizing, logging. Not machine learning. A simple rule set is easy to write and hard to make money with.
Status
The hosted dashboard is a paper trading demo for educational purposes. No real money or API keys are used in the browser version. A separate local Python/Flask interface connects to Binance for backend experimentation and is not accessible through this public site.
Try the browser paper trading dashboard with live Coinbase price data.