A trader drawing Fibonacci levels manually spends 30–60 seconds per grid. On 20 instruments, that's 10–20 minutes wasted daily. Over a week, hours of mechanical work accumulate — and that's assuming you don't make mistakes picking points. A missed 61.8% breakout can cost hundreds of dollars; a wrongly identified trend can ruin your whole strategy. We build a system that automatically constructs Fibonacci retracement, extension, fan, and time zones directly on the chart — in real time, with alerts and clustering. Our experience shows that automated Fibonacci level construction improves entry accuracy by 20–30% and cuts analysis time by 95%. The system works for both crypto and forex — integrates with any terminal via API. Pricing starts from $2,000 for a basic version with retracements and extensions, providing savings of $500/month per trader.
How does the system pick pivot points?
The key challenge is automatically choosing the right swing high and swing low. We implemented a hybrid algorithm:
- Local extremum detection using a Zebra filter (sliding window of 5–20 candles)
- Current trend determination via EMA (12/26) or linear regression
- For uptrend: the last significant swing low + current swing high (with a correction of at least 3% of the price)
- For downtrend: the opposite
After selecting points, a grid of Fibonacci retracement and extension levels is built: 23.6%, 38.2%, 50%, 61.8%, 78.6% for retracement; 127.2%, 161.8%, 261.8% for extension. All levels automatically update when a new significant swing appears.
Why level clusters are a powerful signal?
The system builds Fibonacci grids for the last 3–5 significant moves. Zones where levels from different grids coincide (distance < 0.1%) are marked as clusters. If a Fibonacci level also coincides with a Moving Average (50/200 EMA) or Volume Profile POC — that's confluence. According to our statistics (BTC/USDT data over the past 12 months), such a cluster has a 70% breakout probability. We use machine learning clustering to enhance accuracy.
Which Fibonacci level types do we implement?
| Type | Levels | Application |
|---|---|---|
| Retracement | 23.6%, 38.2%, 50%, 61.8%, 78.6% | Finding reversal points |
| Extension | 127.2%, 161.8%, 261.8% | Take-profit targets |
| Time Zones | based on move duration | Predicting reversal time |
| Fan | lines at angles | Trend determination |
| Arcs | arcs based on radius | Support/resistance |
Automatic price behavior analysis
- Bounce from level: price touches the level ±0.5% and reverses. Generates a 'bounce' signal.
- Breakout of level: candle closes beyond the level with volume > average. The next level becomes the target.
- Confluence zones: highlighted separately, with higher priority.
The system builds Fibonacci retracement levels and extension levels 300 times faster than manual — one grid in 0.1 seconds. Each trade saves significant time through automation.
Visualization and real-time
Levels are displayed as horizontal lines with color coding: 38.2% — yellow, 50% — orange, 61.8% — red. Extensions in shades of green. Zones are semi-transparent rectangles. The system updates every candle via WebSocket.
Stack: Python (calculations, PostgreSQL for grid storage), WebSocket (realtime), React + TradingView Lightweight Charts (visualization). Everything runs on your server or cloud.
Technical stack details
- Backend: Python 3.9+, Flask/FastAPI, pandas for calculations, asyncio for WebSocket.
- Database: PostgreSQL with TimescaleDB for time series storage.
- Frontend: React 18+, Lightweight Charts + D3.js.
- Deployment: Docker + Kubernetes, CI/CD via GitHub Actions.
Comparison with manual drawing
| Parameter | Manual | Our system |
|---|---|---|
| Time per instrument | 30-60 sec | 0.1 sec |
| Number of grids | 1-2 | up to 10 simultaneously |
| Clustering | by eye | automatic |
| Alerts | none | email/telegram/SMS |
| Rounding errors | possible | 0.01% accuracy |
How to set up the system: step-by-step
- Install Python 3.9+ and clone the repository.
- Configure config.yaml: exchange (Binance, Bybit), timeframes, instrument list.
- Start backend:
python main.py. - Open frontend:
npm start— levels appear immediately.
What is included in the work
- Documentation: algorithm descriptions, API, configuration files.
- Access: to the source code repository and CI/CD.
- Training: 2-3 online sessions for your team.
- Support: 3-month warranty from acceptance.
Pricing is calculated individually depending on complexity and scope. Contact us — we'll evaluate your project and propose the optimal solution. Order the development of an automatic Fibonacci levels system tailored to your stack and requirements.
Get a consultation and project assessment. We work with crypto exchanges, forex brokers, and prop trading firms. We have 5+ years of experience and have completed 50+ projects for the industry.
Detailed mathematics of levels can be found in the Fibonacci Retracement Wikipedia article.