DeltaForge

Stop watching charts. Start trading systems.

Quick Start · Why DeltaForge · Features · Docs · 📊 Live Dashboard · Disclaimer

---

🏁 Project Sunset — October 2026

DeltaForge has been archived. After 228 trades across 140 days of development and paper trading (Jun–Oct 2026), the project is being retired.

Final Numbers

Metric	Value
Deployed Capital	₹45,000
Final Capital	₹145,025
Peak Capital	₹177,670
Total Return	+222%
Total Trades	228
Win Rate	47%
Period	Jun 2, 2026 – Oct 19, 2026

What worked

• Signal generation — PULLBACK, TREND_RIDE, and STOCH_CROSS strategies consistently identified good entries
• Trailing stops — STEP_TRAIL and STRATEGY_TRAIL captured 2–3× more on winners than losses
• Risk management — direction caps, lot caps, and HARD_CAP prevented catastrophic single-day blowups
• Real-premium integration — system evolved to use actual option LTPs from the broker API instead of a formula model

What didn't

• HARD_CAP exits destroyed ₹189,450 of profit (63% of gross gains) — the #1 P&L killer across the entire run
• Engine reliability — silent crashes and launchd not restarting cost 15+ trading days with zero activity
• Over-optimization — constant filter tuning (RSI gates, direction whitelists, ADX thresholds) often blocked more profitable trades than bad ones
• Paper ≠ Real — the premium model diverged significantly from actual option market behavior; switching to real premiums late in the project revealed the gap

Lessons learned

1. Ship to live (even 1 lot) early — paper trading validates signals, not P&L
2. Fewer filters, wider stops — let the strategy breathe
3. Reliability > strategy — a system that doesn't run can't make money
4. Real option pricing is non-negotiable from day one

The live dashboard remains available as a record of the full trading history.

— Moving on to the next thing.

---

Your strategy deserves better than a spreadsheet and a prayer.

DeltaForge is an autonomous intraday options trading system for Nifty 50 on the NSE. You define the rules — it watches the market, finds entries, manages risk, sizes positions, and gets out. Paper or live, same engine, same signals, no surprises.

Works with: Angel One SmartAPI · Slack · iMessage · Telegram

Disclaimer — This is educational software, not financial advice. Options trading involves substantial risk of loss. See DISCLAIMER.md.

---

Quick Start

Prerequisites: Python 3.9+, macOS or Linux, Angel One trading account

git clone https://github.com/mimran-khan/deltaforge.git
cd deltaforge
make install              # venv + deps + registers `df` CLI
cp .env.example .env      # add your Angel One creds
df selftest               # verify everything works

Start paper trading:

$ df trade

[09:15:01] Engine started in PAPER mode
[09:15:02] WebSocket connected — streaming Nifty ticks
[09:22:15] SIGNAL  Stochastic Cross  BUY  conf=72
[09:22:15] RISK    9/9 gates passed
[09:22:16] ENTRY   NIFTY 24500 CE @ ₹185.40  qty=50
[10:14:33] EXIT    ₹213.20  P&L +₹1,390  (+15.0%)
[15:30:00] EOD     3 trades | Net P&L +₹3,820 | Capital ₹13,820

Go live when you're ready:

df trade --live           # requires confirmation

---

Why DeltaForge

Most retail options tools are either overpriced terminals you can't customize, or fragile scripts taped together with cron jobs. DeltaForge is neither.

	Manual / Spreadsheet	Paid Terminals	DeltaForge
Execution	You click buttons all day	Alerts, you still click	Fully autonomous — entry to exit
Risk management	"I'll stop at 5% loss" (you won't)	Basic SL/target	9-gate filter + independent kill switch
Position sizing	Fixed lots, guesswork	Basic calculators	Auto-compounds with equity, scales down on drawdown
Backtesting	Copy-paste into Excel	Separate tool, different logic	Same engine — backtest = paper = live
Alerts	Check the app	Email (if lucky)	Instant Slack, iMessage, or Telegram
Cost	Free (your time isn't)	₹2,000–₹15,000/mo	Free and open source

---

Features

Feature	Description
🧠 Multi-Strategy Engine	Trend Ride, Pullback-in-Trend, Supertrend Flip, Stochastic Cross — all run in parallel with regime-aware confidence scoring
🌐 Multi-Asset Trading	Nifty 50 options (NSE) + MCX Crude Oil Mini + CDS USDINR futures — each with independent risk pools
🛡️ Adaptive Risk System	Daily loss limits (10%), consecutive-loss brakes, drawdown tiers, adaptive mode that adjusts confidence/lots in real-time
🔴 Kill Switch	Independent watchdog process. Halts trading on breach, manual override via CLI or dashboard. Works even if the main engine hangs.
📈 Compound Sizing	Scales lots with capital growth. Position sizing adapts to equity curve with configurable per-lot capital allocation.
📊 Live Dashboard	FastAPI + WebSocket — capital, trades, risk status, halt toggle. All at localhost:8900
💬 Instant Alerts	Slack (recommended), iMessage (macOS), or Telegram — entries, exits, errors, end-of-day P&L reports
🔄 Paper ↔ Live Parity	Identical code path across backtest, paper, and live. If it works in paper, it works in production. Zero divergence.
📉 Walk-Forward Backtest	100+ day replays with production engine. Realistic costs (brokerage + STT + slippage). Compound lot sizing.
⏰ Automated Scheduling	macOS launchd agent for daily auto-start, code sync, health monitoring, and session management

---

What It Does

graph LR
    A["📡 Angel One WebSocket"] --> B["🕯 CandleBuilder"]
    B --> C["🧠 MultiStrategyEngine"]
    C --> D{"🛡️ RiskEngine + AdaptiveMode"}
    D -->|"✅ pass"| E["💰 PositionManager"]
    D -->|"❌ reject"| F(("Skip"))
    E --> G["📊 Dashboard"]
    E --> H["💬 Alerts"]
    I["🔴 Kill Switch"] -.->|"halt"| D
    J["📡 Futures Feed"] --> K["🌐 MultiAssetEngine"]
    K --> D

Loading

---

Usage

# Trading
df trade                    # paper mode (default)
df trade --live             # live (confirmation required)
df trade -v                 # verbose / debug output

# Monitoring
df status                   # capital, drawdown, risk state
df logs -f                  # follow today's log
df db stats                 # per-strategy performance
df db summary               # today's trade summary

# Backtesting
df backtest --days 180      # default capital ₹10,000
df backtest --days 365 --capital 50000

# Testing & Verification
df selftest                 # pre-flight engine check
df broker                   # test broker connection
df alert "test"             # verify alert channel

# Dashboard
make dashboard              # web UI at localhost:8900
make run                    # trading + dashboard together

---

Configuration

Copy .env.example and fill in your credentials:

# Angel One SmartAPI (required)
ANGEL_API_KEY=your_api_key
ANGEL_CLIENT_ID=your_client_id
ANGEL_PASSWORD=your_password
ANGEL_TOTP_SECRET=your_totp_secret

# Trading
TRADING_MODE=paper
STARTING_CAPITAL=10000

# Risk Management
DAILY_LOSS_LIMIT_PCT=10          # halt after 10% daily loss
CAPITAL_PER_LOT=3000             # lot sizing aggressiveness

# Alerts — choose: "slack", "imessage", or "telegram"
ALERT_METHOD=slack
SLACK_BOT_TOKEN=xoxb-your-token
SLACK_CHANNEL_ID=C0XXXXXXXXX

All parameters live in config/settings.py and are overridable via .env. See .env.example for every configurable option.

---

Paper vs Live

	Paper	Live
Signal generation	Identical engine	Identical engine
Order execution	Simulated (premium model)	Angel One SmartAPI
Costs	Realistic (brokerage + STT + slippage)	Actual
Risk gates	All 9 enforced	All 9 enforced

Same code. Same signals. Same risk. If it works in paper, it works in live.

---

Documentation

Topic	Description
Strategies & Indicators	How each strategy generates signals, indicator reference, options model
Risk Model & Kill Switch	9-gate system, drawdown tiers, real-time monitoring, kill switch
Dashboard & API	REST endpoints, WebSocket, security configuration, alerts
Architecture & Design	Data flow, system components, project structure, design decisions

---

Live Performance Dashboard

📊 View Live Dashboard →

Auto-generated trading report with interactive equity curves, calendar heatmaps, trade analytics, and strategy breakdowns. Deployed via GitHub Pages — updated with each push.

---

Roadmap

Timeline	Milestone
Shipped	Multi-strategy engine, adaptive risk system, multi-asset (Nifty + MCX + CDS), compound sizing, automated daily scheduling, walk-forward backtesting, Slack/iMessage/Telegram alerts, FastAPI dashboard
Next	BankNifty support, strategy marketplace, ML-based regime detection
Future	Mobile companion alerts, cloud sync, community strategies, broker-agnostic execution

---

Built With

• Angel One SmartAPI — Broker connectivity & WebSocket market data
• FastAPI — Dashboard HTTP layer
• Loguru — Structured logging
• Pandas — Indicator computation & data analysis
• SQLite — Trade persistence
• WebSocket — Real-time market feed

---

Contributing

We're building the trading system we wish existed — help us make it better.

Contribute: New strategies, risk improvements, broker integrations, indicators, bug fixes, docs

• Contributing Guide
• Report Bug
• Request Feature
• Good First Issues

---

Testing

make test                   # full test suite
make test-e2e               # end-to-end only
make test-component         # component tests

---

License

MIT — see LICENSE. Full trading disclaimer in DISCLAIMER.md.

---

Your strategy deserves better than a spreadsheet and a prayer.

⭐ Star · 🔄 Share · 🤝 Contribute