Shadow Mode¶

Shadow mode lets you test strategies without executing real trades. Signals are generated and logged, but no orders are placed.

Why Use Shadow Mode?¶

Test new strategies - Verify logic before risking capital
Validate configuration - Ensure settings work as expected
Learn the system - Understand how PolyBot operates
Debug issues - Troubleshoot without financial risk

Enabling Shadow Mode¶

Per Strategy¶

# Enable shadow mode for a specific strategy
polybot strategy shadow arbitrage --enable

# Disable shadow mode (go live)
polybot strategy shadow arbitrage --disable

Check Status¶

# List strategies with shadow status
polybot strategy list

Output:

┌────────────┬─────────────────────────┬─────────┬────────┐
│ Name       │ Description             │ Enabled │ Shadow │
├────────────┼─────────────────────────┼─────────┼────────┤
│ arbitrage  │ YES+NO arbitrage        │ ✓       │ ✓      │
│ stat_arb   │ Statistical arbitrage   │ ✓       │ ✗      │
│ ai_model   │ AI model predictions    │ ✗       │ ✗      │
└────────────┴─────────────────────────┴─────────┴────────┘

How Shadow Mode Works¶

Normal Mode¶

Price Update → Strategy Scan → Signal → Executor → Order Placed

Shadow Mode¶

Price Update → Strategy Scan → Signal → Logged (no execution)

In shadow mode: - Strategies receive real price data - Signals are generated normally - Signals are logged to the strategy logs database - No orders are sent to the exchange

Viewing Shadow Signals¶

Dashboard¶

View shadow signals in the web UI:

http://localhost:8000/ui → Strategy Logs

CLI¶

# View recent strategy logs
polybot logs strategy arbitrage --tail 50

# Filter by signal type
polybot logs strategy arbitrage --type signal

API¶

curl http://localhost:8000/api/strategy-logs?strategy=arbitrage

Analyzing Shadow Performance¶

Track hypothetical performance:

Signal count - How many opportunities found?
Signal quality - What prices were signaled?
Timing - When do signals occur?
Strategy comparison - Which strategies find more opportunities?

Transitioning to Live¶

When you're ready to go live:

1. Review Shadow Results¶

# Check shadow performance
polybot logs strategy arbitrage --days 7

Look for: - Consistent signal generation - Reasonable prices - No errors or anomalies

2. Start Small¶

# Set conservative limits
MAX_POSITION_SIZE_USD=50
MAX_TOTAL_EXPOSURE_USD=500

3. Disable Shadow Mode¶

polybot strategy shadow arbitrage --disable

4. Monitor Closely¶

Watch the first few trades carefully: - Check execution prices - Verify position tracking - Monitor P&L

Best Practices¶

Run shadow for at least 24-48 hours - See different market conditions
Compare to expectations - Are signals what you expected?
Check error logs - Any issues during shadow period?
Test all strategies - Shadow each before going live
Keep some strategies in shadow - Compare live vs shadow performance

Troubleshooting¶

No signals in shadow mode¶

Check strategy is enabled: polybot strategy list
Verify market conditions match strategy requirements
Check for errors: polybot logs --level ERROR

Too many signals¶

Review strategy thresholds
Check if parameters are too aggressive
Verify data quality

Shadow and live disagreement¶

Execution slippage is normal
Live has real market impact
Timing differences matter