╔══════════════════════════════════════════════════════════╗
║                                                          ║
║          ⚠️  LIVE TRADING RISKS REAL CAPITAL  ⚠️           ║
║                                                          ║
║  • You can lose ALL deployed capital                    ║
║  • Bugs in strategy code cause significant losses       ║
║  • Market conditions change - backtest ≠ live           ║
║  • NEVER deploy without thorough backtesting            ║
║  • Start with small capital to validate live behavior   ║
║  • Monitor deployments actively (daily minimum)         ║
║  • Define exit criteria BEFORE deploying                ║
║                                                          ║
║            THIS IS NOT A SIMULATION                      ║
║           REAL MONEY WILL BE TRADED                      ║
║                                                          ║
╚══════════════════════════════════════════════════════════╝

Use MCPSearch to select: mcp__workbench__deployment_create
Use MCPSearch to select: mcp__workbench__deployment_list
Use MCPSearch to select: mcp__workbench__deployment_stop

Type: Direct wallet trading
Setup: Immediate (no additional requirements)
Limit: Max 1 active deployment per wallet
Complexity: Lower
Best for: Testing, personal trading, single strategy
Cost: $0.50 to create

Advantages:
✓ Simple setup
✓ Immediate deployment
✓ No minimum balance requirement

Disadvantages:
✗ Only 1 deployment per wallet
✗ No public performance tracking
✗ Personal wallet at risk

Type: Professional vault setup
Setup: Requires 200+ USDC in wallet
Limit: Unlimited deployments
Complexity: Higher
Best for: Multiple strategies, professional trading, public showcasing
Cost: $0.50 per deployment

Advantages:
✓ Unlimited deployments
✓ Public TVL and performance tracking
✓ Professional infrastructure
✓ Separate from personal wallet

Disadvantages:
✗ Requires 200+ USDC setup
✗ More complex configuration
✗ Public performance visibility

Choose EOA if:
- First deployment (testing live behavior)
- Running single strategy
- Want simple setup
- Don't need multiple simultaneous strategies

Choose Vault if:
- Running multiple strategies
- Want professional setup
- Need public performance tracking
- Trading with significant capital
- Building track record

Leverage = Position size / Available capital

1x leverage: $1000 capital → $1000 position
2x leverage: $1000 capital → $2000 position
3x leverage: $1000 capital → $3000 position

Key points:
- Leverage multiplies BOTH gains AND losses
- Higher leverage = higher risk
- Liquidation risk increases with leverage
- Start conservative (1-2x)

Conservative (1x):
- No amplification
- Lower returns, lower risk
- Recommended for first deployments
- Drawdown ≈ backtest drawdown

Moderate (2-3x):
- 2-3× returns and risk
- Requires careful monitoring
- Only after 1x deployment validated
- Drawdown ≈ 2-3× backtest drawdown

Aggressive (4-5x):
- 4-5× returns and risk
- Very risky, high liquidation chance
- NOT recommended for most users
- Drawdown ≈ 4-5× backtest drawdown
- Can lose entire capital quickly

Backtest: 15% max drawdown
1x deployment: 15% expected drawdown
2x deployment: 30% expected drawdown (may hit margin call)
3x deployment: 45% expected drawdown (very likely liquidation)

Rule: Keep leverage low enough that backtest drawdown × leverage < 25%

Strategy controls position size via code (85-95% margin usage)
Deployment leverage multiplies available margin
Total risk = Strategy position size × Deployment leverage

Example:
- Capital: $1000
- Strategy uses 90% margin
- Deployment leverage: 2x
- Actual position: $1000 × 0.90 × 2 = $1800

Position size is LARGER than capital (risk of liquidation)

Example thresholds:
- Stop if down 10% from starting capital
- Stop if down 15% from peak
- Stop if drawdown >1.5× backtest max drawdown

Write down your threshold:
"I will stop this deployment if capital drops to $______"

DO NOT move this threshold once deployed (discipline is critical)

First 24-48 hours: Check every 2-4 hours
First week: Check daily minimum
First month: Check every 2-3 days
After 1 month: Weekly check acceptable (if performing well)

NEVER:
- Deploy and forget
- Ignore for >1 week during first month
- Assume backtest = live performance

WRONG approach:
- Backtest shows 50% annual return
- Deploy $10,000 immediately
- If strategy fails, lose significant capital

RIGHT approach:
- Deploy $500-1000 initially (5-10% of intended size)
- Monitor for 1-2 weeks
- Validate live behavior matches backtest
- If successful, scale up gradually
- Reduce risk during validation phase

Scaling schedule example:
Week 1-2: $1,000 (test)
Week 3-4: $2,000 (if performing well)
Week 5-6: $4,000 (if still performing well)
Month 2+: Scale to full size gradually

1. P&L vs backtest expectation:
   - Is live performance similar to backtest?
   - Track daily, weekly, monthly returns
   - Compare to backtest metrics

2. Drawdown:
   - Current drawdown from peak
   - Compare to backtest max drawdown
   - If exceeds backtest max × 1.5, be concerned

3. Trade execution:
   - Are trades executing as expected?
   - Check fill prices (slippage)
   - Verify trade frequency matches backtest

4. Win rate and profit factor:
   - Track live win rate
   - Should be close to backtest win rate
   - If diverges >20%, investigate

5. Market regime:
   - Has market character changed?
   - Trending → ranging or vice versa
   - Strategy may stop working if regime shifts

Live drawdown >30% OR >1.5× backtest max drawdown
Example:
- Backtest max drawdown: 15%
- Threshold to stop: 22.5% (1.5× backtest)
- Current live drawdown: 25%
→ STOP IMMEDIATELY

Why: Strategy may be broken or market changed

Live win rate <50% of backtest win rate
Example:
- Backtest win rate: 55%
- Threshold to stop: 27.5% (50% of backtest)
- Live win rate after 20 trades: 25%
→ STOP IMMEDIATELY

Why: Strategy logic not working in live market

Much higher or lower trade frequency than backtest
Example:
- Backtest: 2-3 trades per day
- Live: 15 trades per day
→ STOP IMMEDIATELY

Why: Strategy may be malfunctioning

10+ consecutive losing trades (when backtest shows max 5-6)
→ STOP IMMEDIATELY

Why: Strategy edge may have disappeared

- Orders not executing
- Repeated API errors
- Position sizing errors
- Strategy crashes/restarts frequently
→ STOP IMMEDIATELY

Why: Technical problems = unpredictable risk

Market conditions fundamentally different from backtest period
Examples:
- Extreme volatility event (>3× normal)
- Major regulatory news
- Exchange issues
→ STOP, REASSESS, decide if/when to restart

Why: Strategy designed for different conditions

1. Compare metrics:
   | Metric         | Backtest | Live  | Variance |
   |----------------|----------|-------|----------|
   | Sharpe         | 1.5      | 1.3   | -13%     |
   | Drawdown       | 12%      | 15%   | +25%     |
   | Win rate       | 52%      | 49%   | -6%      |
   | Profit factor  | 1.8      | 1.6   | -11%     |

2. Evaluate variance:
   - Small variance (<20%) → Expected, continue ✓
   - Moderate variance (20-40%) → Monitor closely, may be temporary
   - Large variance (>40%) → Significant concern, consider stopping

3. Decision:
   - If metrics acceptable: Continue monitoring
   - If metrics concerning: Investigate cause
   - If red flags present: Stop deployment

Review:
- Total return vs expectation
- Max drawdown experienced
- Trade execution quality
- Any technical issues

Decide:
- Scale up capital (if performing well)
- Continue same size (if acceptable)
- Scale down or stop (if underperforming)

1. Final pre-deployment check:
   ☑ Backtested 6+ months (Sharpe 1.4, drawdown 14%)
   ☑ Code reviewed (no bugs found)
   ☑ Risk management validated
   ☑ Starting capital: $500 (can afford to lose)
   ☑ Monitoring plan: Check daily for first week
   ☑ Exit criteria: Stop if down >20% or drawdown >25%

2. Check credit balance:
   get_credit_balance()
   → Balance: 100 USDC ✓ (sufficient for deployment $0.50)

3. Deploy:
   deployment_create(
       strategy_name="RSIMeanReversion_M",
       symbol="BTC-USDT",
       timeframe="1h",
       leverage=1,  # Conservative for first deployment
       deployment_type="eoa"
   )
   → Deployment ID: abc123
   → Status: Active
   → Cost: $0.50

4. Monitor closely:
   Day 1: Check every 4 hours
   Day 2-7: Check daily
   Track: P&L, drawdown, trade execution

5. After 1 week:
   Review performance vs backtest
   If good: Continue and consider scaling up
   If poor: Stop and analyze what went wrong

1. Setup vault (one-time):
   - Verify 200+ USDC in wallet
   - Decide vault name (unique, descriptive)

2. Deploy first strategy:
   deployment_create(
       strategy_name="TrendFollower_M",
       symbol="BTC-USDT",
       timeframe="4h",
       leverage=2,
       deployment_type="vault",
       vault_name="AlgoTrading_Vault_2025",
       vault_description="Multi-strategy algorithmic trading vault"
   )
   → Vault created successfully

3. Deploy second strategy (same vault):
   deployment_create(
       strategy_name="MeanReversion_L",
       symbol="ETH-USDT",
       timeframe="1h",
       leverage=1,
       deployment_type="vault",
       vault_name="AlgoTrading_Vault_2025"  # Same vault name
   )

4. Monitor all deployments:
   deployment_list()
   → Shows both strategies with individual performance

5. Manage independently:
   - Can stop one strategy without affecting other
   - Each strategy tracks separate P&L
   - Vault shows combined performance

1. Monitor deployment:
   deployment_list()
   → Strategy: MomentumBreakout_H
   → P&L: -18% (started $1000, now $820)
   → Drawdown: 28%
   → Red flag: Drawdown > 1.5× backtest max (15% × 1.5 = 22.5%)

2. Decision: STOP (red flag triggered)

3. Stop deployment:
   deployment_stop(deployment_id="abc123")
   → Status: Stopped
   → Final P&L: -$180 (-18%)

4. Analyze what went wrong:
   - Review trade history
   - Check market conditions during deployment
   - Compare to backtest assumptions
   - Identify issue (market regime change? bug? bad luck?)

5. Next steps:
   - Fix issues if identified (use improve-trading-strategies)
   - Re-backtest with improvements
   - Deploy again with smaller capital if confident
   - Or abandon strategy if fundamentally broken

1. Previously stopped deployment due to high volatility event
   Stopped during extreme market conditions

2. Market stabilizes:
   - Check current market conditions
   - Compare to backtest environment
   - Decide conditions are favorable again

3. Review strategy:
   - Re-backtest on recent data
   - Verify strategy still works
   - Check no code changes needed

4. Restart deployment:
   deployment_start(deployment_id="abc123")
   → Status: Active (resumed)

5. Monitor closely:
   - First day: Check multiple times
   - Verify execution matches expectations
   - Be ready to stop again if issues recur

1. Check balance:
   get_credit_balance() → Balance: 0.20 USDC

2. Purchase credits:
   - Visit Robonet dashboard
   - Add credits to account
   - Deployment costs $0.50

3. Retry deployment after purchase

1. Stop existing EOA deployment:
   deployment_list() → Find existing deployment
   deployment_stop(deployment_id="existing_id")

2. Or switch to Hyperliquid Vault:
   - Requires 200+ USDC in wallet
   - Allows unlimited deployments
   - Use deployment_type="vault"

3. Or use different wallet (new EOA)

1. Verify 200+ USDC in wallet:
   - Check wallet balance on Hyperliquid
   - Vault requires minimum balance

2. Check vault name unique:
   - Try different vault name
   - Vault names must be unique across Hyperliquid

3. Verify wallet permissions:
   - Ensure wallet connected properly
   - Check Hyperliquid account status

1. Slippage higher than expected:
   - Market less liquid than backtest assumed
   - Solution: Use wider stops, lower frequency trades, or stop deployment

2. Fees not properly accounted:
   - Forgot to include fees in backtest
   - Solution: Re-backtest with realistic fees (0.05-0.1%)

3. Market regime changed:
   - Trending market → ranging market
   - Solution: Strategy may not work in current conditions, stop deployment

4. Execution delays:
   - Live trades execute slower than backtest assumed
   - Solution: Use longer timeframes (1h instead of 5m)

5. Overfitted strategy:
   - Strategy memorized past data
   - Solution: Simplify strategy, re-backtest, test on out-of-sample data

Decision: If performance -30% worse than backtest, STOP and fix issues

⚠️ Trading crypto perpetuals is HIGH RISK
⚠️ Regulations vary by jurisdiction
⚠️ You are responsible for compliance with local laws
⚠️ This is NOT financial advice
⚠️ Trade at your own risk
⚠️ Only risk capital you can afford to lose 100%