forge data

Fetch, update, and manage historical market data. Data is stored as Parquet files under alpha-strategies/data/historical/.

Examples

# Initial fetch
forge data fetch USDJPY
forge data fetch CL=F
forge data fetch SPY QQQ NAS100

# Incremental update (all symbols)
forge data update

# Update a single symbol
forge data update --symbol USDJPY

# Check what's saved
forge data list

forge strategy

Create, register, and inspect strategy JSON definitions. Strategies must be registered before use in backtests or optimization.

Examples

# Create from template
forge strategy create --template ema_crossover --id usdjpy_ema_v1 \
  --out data/strategies/usdjpy_ema_v1.json

# Register after editing the JSON
forge strategy save data/strategies/usdjpy_ema_v1.json

# List registered strategies
forge strategy list

# Inspect a strategy definition
forge strategy show usdjpy_ema_v1

forge backtest

Run backtests, view history, and analyse results. All results are automatically saved to SQLite.

Examples

# Basic backtest
forge backtest run USDJPY --strategy usdjpy_ema_v1

# With date range
forge backtest run USDJPY --strategy usdjpy_ema_v1 \
  --start 2021-01-01 --end 2024-12-31

# JSON output
forge backtest run USDJPY --strategy usdjpy_ema_v1 --json

# Walk-forward test (5 folds)
forge backtest wft USDJPY --strategy usdjpy_ema_v1 --folds 5

# List results sorted by Sharpe
forge backtest list --sort sharpe --limit 20

# Open interactive chart in browser
forge backtest chart USDJPY --strategy usdjpy_ema_v1

forge optimize

Bayesian parameter search with Optuna. Includes overfitting risk evaluation, walk-forward optimization, and one-click apply.

Examples

# 300 trials, 4 workers, maximize Sharpe
forge optimize run USDJPY --strategy usdjpy_ema_v1 \
  --metric sharpe_ratio --trials 300 -j 4 --save

# View optimization history
forge optimize history --strategy usdjpy_ema_v1 --sort score

# Apply best result as a new strategy
forge optimize apply data/results/usdjpy_ema_v1_opt.json \
  --to-strategy usdjpy_ema_v1_optimized

# Check parameter robustness
forge optimize sensitivity USDJPY --strategy usdjpy_ema_v1_optimized

forge pine

Auto-generate TradingView Pine Script v6 from a strategy definition. Supports machine-learning indicators such as HMM by embedding trained parameters.

Examples

# Generate Pine Script (saved to output/pinescript/)
forge pine generate --strategy usdjpy_ema_v1_optimized

# Embed trained HMM parameters
forge pine generate --strategy cl_hmm_bb_rsi_v1 --with-training-data

# Preview without writing to disk
forge pine preview --strategy usdjpy_ema_v1_optimized

# Import existing .pine file as a strategy
forge pine import output/pinescript/usdjpy_ema_v1.pine --id usdjpy_ema_imported

forge dashboard

Launch the Streamlit-based Web UI for backtest visualisation, strategy rankings, and idea tracking.

forge dashboard
forge dashboard --port 8501 --no-open

forge license

Activate and manage your LemonSqueezy license. Activation data is cached in ~/.forge/license.json.

End-to-End Strategy Development Workflow

A typical flow from raw data to live execution. This pairs naturally with a coding agent (e.g. Claude Code) for automated parameter exploration and strategy generation.

• Fetch historical data

  forge data fetch USDJPY

• Create a strategy from a template Generate a JSON scaffold, edit parameters, then register it.

  forge strategy create --template ema_crossover \
    --id usdjpy_ema_v1 \
    --out data/strategies/usdjpy_ema_v1.json
  
  # Edit the JSON, then register
  forge strategy save data/strategies/usdjpy_ema_v1.json

• Run a backtest

  forge backtest run USDJPY --strategy usdjpy_ema_v1
  
  # Visual equity curve
  forge backtest chart USDJPY --strategy usdjpy_ema_v1

• Optimize parameters Bayesian search with Optuna, then apply the best result.

  forge optimize run USDJPY --strategy usdjpy_ema_v1 \
    --metric sharpe_ratio --trials 300 -j 4 --save
  
  forge optimize apply data/results/usdjpy_ema_v1_opt.json \
    --to-strategy usdjpy_ema_v1_optimized

• Walk-forward validation Detect overfitting with out-of-sample testing.

  forge backtest wft USDJPY --strategy usdjpy_ema_v1_optimized \
    --folds 5
  
  forge optimize sensitivity USDJPY \
    --strategy usdjpy_ema_v1_optimized

• Generate Pine Script Export a TradingView alert script from the optimized strategy.

  forge pine generate --strategy usdjpy_ema_v1_optimized

  Output: output/pinescript/usdjpy_ema_v1_optimized.pine

Deploying Pine Script to TradingView

Paste the generated .pine file into TradingView and set up an alert to trigger alpha-strike.

• Open the Pine Editor Open a chart in TradingView and click the "Pine Editor" tab at the bottom.
• Paste the script Copy the contents of the generated .pine file and paste it into the editor. Click "Add to chart" (▶).
• Create an alert Click the bell icon (Alerts) → "Create alert". Set the condition to your script, enable "Webhook URL", and paste the alpha-strike endpoint.
• Tip: in-script alert conditions Define alertcondition in Pine Script for cleaner alert setup.

  longSignal = ta.crossover(ema_fast, ema_slow)
  alertcondition(longSignal, title="Long Entry", message="long")

Connecting TradingView to alpha-strike

alpha-strike receives TradingView webhook alerts and forwards orders to OANDA or moomoo.

1. Configure environment variables

Create alpha-strike/.env from the provided example:

# Required
WEBHOOK_PASSPHRASE=your-secret-passphrase

# OANDA
OANDA_API_KEY=your-personal-access-token
OANDA_ACCOUNT_ID=your-account-id
OANDA_ENV=PRACTICE    # or LIVE

# moomoo
MOOMOO_HOST=127.0.0.1
MOOMOO_PORT=11111
MOOMOO_TRD_ENV=SIMULATE   # or REAL

2. Webhook URL

3. Payload format

{
  "passphrase": "your-secret-passphrase",
  "broker": "oanda",
  "asset_class": "FX",
  "action": "buy",
  "ticker": "USDJPY",
  "quantity": 1000
}

4. Ticker to OANDA instrument mapping

5. Verify connectivity

# Health check
curl http://localhost:8080/health
# → {"status":"ok"}

# Test order (use PRACTICE / SIMULATE environment)
curl -X POST http://localhost:8080/webhook \
  -H "Content-Type: application/json" \
  -d '{"passphrase":"your-secret","broker":"oanda","asset_class":"FX","action":"buy","ticker":"USDJPY","quantity":1000}'