trading-best-practices（交易最佳实践）

Skill: Trading Best Practices

When to use this skill

• Before implementing a new trading strategy
• When modifying risk management logic
• Quarterly review of existing strategies
• Before going live with real capital
• When performance degrades unexpectedly

Purpose

This skill ensures that trading implementations follow current best practices and avoid common pitfalls in algorithmic trading. It includes mechanisms to stay updated with the latest financial research and market structure changes.

Critical Trading Principles

1. Risk Management (Non-negotiable)

Position Sizing:

• Never risk more than 1-2% of capital per trade
• Use Kelly Criterion or fixed fractional sizing
• Account for correlation between positions

Stop Losses:

• ALWAYS use stop losses (no exceptions)
• Place stops based on volatility (ATR) not arbitrary percentages
• Never move stops against your position

Drawdown Protection:

• Maximum drawdown threshold: 20% (conservative) to 30% (aggressive)
• Implement circuit breakers for daily loss limits
• Use high-water mark tracking

2. Strategy Development

Avoid Overfitting:

• ❌ Don't optimize on the same data you test on
• ✅ Use walk-forward analysis
• ✅ Test on out-of-sample data
• ✅ Prefer simple strategies with fewer parameters

Backtesting Integrity:

• Account for transaction costs (commissions + slippage)
• Use realistic fill assumptions (no perfect fills at close)
• Avoid look-ahead bias (only use data available at decision time)
• Include survivorship bias (test on delisted stocks too)

Statistical Validation:

• Minimum 100+ trades for statistical significance
• Sharpe Ratio > 1.0 (preferably > 1.5)
• Profit Factor > 1.5
• Win rate should match strategy type (trend: 40-50%, mean reversion: 55-65%)

3. Market Microstructure Awareness

Execution Quality:

• Use limit orders to control slippage
• Avoid market orders on illiquid assets
• Be aware of bid-ask spread costs
• Consider market impact for larger positions

Regime Awareness:

• Strategies perform differently in bull/bear/sideways markets
• Adapt position sizing to market volatility (VIX)
• Reduce exposure during high uncertainty events

4. Common Pitfalls to Avoid

Research Workflow

To stay current with financial innovation, perform quarterly reviews:

Step 1: Research Latest Practices

# Use web search to find recent research
# Topics to research:
# - "algorithmic trading best practices 2026"
# - "quantitative finance risk management"
# - "market microstructure changes"
# - "regulatory changes algorithmic trading"

Step 2: Review Current Implementation

Compare findings against:

• src/domain/risk/ - Risk management logic
• src/application/strategies/ - Strategy implementations
• docs/STRATEGIES.md - Strategy documentation

Step 3: Identify Gaps

Document any practices we're missing or doing incorrectly.

Step 4: Update Implementation

If gaps found:

1. Create issue/task for improvement
2. Follow /implement workflow
3. Backtest changes thoroughly
4. Update this skill with new learnings

Checklist: Strategy Implementation

Before implementing ANY new strategy:

• Strategy has clear entry/exit rules
• Risk per trade is defined (max 2%)
• Stop loss logic is implemented
• Position sizing accounts for volatility
• Backtested on 2+ years of data
• Tested on out-of-sample data
• Transaction costs included in backtest
• Sharpe Ratio > 1.0
• Max Drawdown < 20%
• No look-ahead bias
• Strategy logic is simple (fewer parameters = better)
• Correlation with existing strategies checked

Red Flags in Strategy Design

// ❌ RED FLAG: No stop loss
if signal == Signal::Buy {
    execute_order(symbol, quantity); // Where's the stop?
}

// ❌ RED FLAG: Fixed position size (ignores risk)
let quantity = 100; // Always 100 shares?

// ❌ RED FLAG: No transaction costs
let profit = exit_price - entry_price; // Ignores commissions/slippage

// ❌ RED FLAG: Too many parameters
struct Strategy {
    sma_period_1: usize,
    sma_period_2: usize,
    rsi_period: usize,
    rsi_oversold: f64,
    rsi_overbought: f64,
    macd_fast: usize,
    macd_slow: usize,
    // ... 20 more parameters = overfitting
}

// ✅ GOOD: Risk-based position sizing with stop
let risk_amount = capital * risk_per_trade;
let stop_distance = entry_price * atr_multiplier;
let quantity = risk_amount / stop_distance;
let stop_loss = entry_price - stop_distance;

Resources to Monitor

Academic Research:

• SSRN (Social Science Research Network)
• arXiv quantitative finance section
• Journal of Portfolio Management

Industry Standards:

• CFA Institute guidelines
• FIX Protocol updates (market structure)
• SEC/FINRA regulatory changes

Market Data:

• VIX (volatility regime)
• Sector rotation trends
• Correlation matrices

Update Frequency

• Monthly: Check VIX and market regime
• Quarterly: Research latest academic papers
• Annually: Full strategy review and revalidation
• Ad-hoc: When performance degrades or market structure changes

Integration with Other Skills

• Use benchmarking skill to validate strategies
• Use critical-review skill for code quality
• Use rust-trading skill for implementation rules
• Update documentation skill when best practices change