Best Practices

Learn proven techniques and strategies for developing effective trading strategies with Lona.

Strategy Development Philosophy

Start Simple, Then Iterate

The most successful strategy development follows this pattern:

1. Begin with Core Concept: Test the basic idea
2. Validate Logic: Ensure it works as expected
3. Refine Parameters: Optimize key values
4. Add Filters: Improve signal quality
5. Test Robustness: Verify across markets and timeframes

Golden Rule: A simple strategy that works beats a complex strategy that doesn’t.

The 80/20 Principle

80% of your results come from:

• ✅ Solid core logic
• ✅ Appropriate indicators
• ✅ Clear entry/exit rules
• ✅ Risk management (when needed)

20% of your results come from:

• ⚠️ Complex optimizations
• ⚠️ Micro-adjustments
• ⚠️ Minor refinements

Focus on the 80% first!

Interview Phase Best Practices

Communicate Clearly

Do:

✅ "Buy when the 10-period SMA crosses above the 30-period SMA"
✅ "Exit when RSI reaches 70"
✅ "Use 14-period RSI with 30/70 thresholds"

Don’t:

❌ "Use moving averages"  (which ones? what periods?)
❌ "Trade RSI"  (how? what thresholds?)
❌ "Make it profitable"  (not specific enough)

Provide Context

Help Lona understand your reasoning:

Good:

"I want to use RSI for mean reversion. When RSI falls below 30,
the market is oversold. I want to buy when RSI crosses back above
30, signaling recovery. Exit when RSI reaches 70 (overbought)."

Why it’s good:

• Explains the strategy type (mean reversion)
• Defines what indicators mean (oversold/overbought)
• Specifies exact entry/exit triggers
• Shows the logic flow

Answer Questions Completely

When Lona asks clarifying questions:

Effective Response:

Lona: "What moving average periods would you like?"
You: "Use 20 for the fast MA and 50 for the slow MA. 
      Both should be simple moving averages."

Ineffective Response:

Lona: "What moving average periods would you like?"
You: "The usual ones"  (ambiguous)

Strategy Design Best Practices

Choose Appropriate Indicators

Match indicators to your strategy type:

Trend-Following:

• ✅ Moving Averages (SMA, EMA)
• ✅ MACD
• ✅ ADX
• ❌ RSI (momentum indicator)
• ❌ Stochastic (oscillator)

Mean Reversion:

• ✅ RSI
• ✅ Stochastic
• ✅ Bollinger Bands
• ❌ Moving Average Crossovers
• ❌ Trend indicators

Momentum:

• ✅ RSI
• ✅ MACD
• ✅ Rate of Change
• ❌ Mean reversion indicators

Use Logical Combinations

Good Combinations:

Trend Filter + Timing:

Trend: 50-period MA (identify direction)
Timing: MACD crossover (precise entry)
Logic: Only long when price > MA AND MACD crosses up

Momentum + Confirmation:

Momentum: RSI for overbought/oversold
Confirmation: Volume spike
Logic: Buy on RSI < 30 AND volume > average

Bad Combinations:

Conflicting Signals:

❌ Trend-following MA + Mean reversion RSI
   (One says trend, other says reverse)

Redundant Indicators:

❌ Three different moving averages (10, 20, 30 MA)
   (All measure the same thing)

Define Clear Exit Rules

Every strategy needs exits:

Good Exit Rules:

• ✅ “Exit when indicator gives opposite signal”
• ✅ “Close position when MA crosses below price”
• ✅ “Exit after holding for N bars”
• ✅ “Take profit at X% gain”

Poor Exit Rules:

• ❌ “Exit when profitable” (when exactly?)
• ❌ “Close if it drops” (how much?)
• ❌ “Exit soon” (not specific)

Parameter Selection

Start with Standard Values:

Then Adjust Based on:

• Market volatility
• Your trading timeframe
• Backtest results
• Strategy responsiveness needs

Parameter Optimization

Systematic Testing Approach

Step 1: Baseline Test

Test with standard parameters
Document results
Establish baseline performance

Step 2: One Variable at a Time

Change only fast MA period: 5, 10, 15, 20
Keep slow MA constant at 50
Test each variation
Compare to baseline

Step 3: Combine Best Values

Best fast MA: 15
Now test slow MA: 30, 50, 70, 100
Keep fast MA at 15
Find optimal combination

Step 4: Final Validation

Test best combination on different:
- Time periods
- Market conditions
- Symbols

Avoiding Over-Optimization

Warning Signs of Over-Optimization:

🚩 Too Perfect: 90%+ win rate, minimal drawdown 🚩 Too Specific: Works only with very precise parameters 🚩 Not Robust: Small parameter changes = dramatic performance drop 🚩 Curve Fitting: Optimized to every historical wiggle 🚩 Too Complex: 10+ parameters to tune

How to Avoid:

✅ Test Out-of-Sample: Reserve some data for validation ✅ Use Parameter Ranges: Strategy should work with similar values ✅ Cross-Validate: Test on different symbols/timeframes ✅ Simplicity First: Fewer parameters = more robust ✅ Economic Logic: Strategy should make intuitive sense

Parameter Ranges to Explore

Moving Averages:

Fast: 5-30 (try: 5, 10, 15, 20)
Slow: 30-200 (try: 30, 50, 100, 200)

RSI:

Period: 2-30 (try: 7, 14, 21)
Oversold: 20-40 (try: 20, 25, 30)
Overbought: 60-80 (try: 70, 75, 80)

Bollinger Bands:

Period: 10-30 (try: 10, 20, 30)
Std Dev: 1-3 (try: 1.5, 2.0, 2.5)

Testing Best Practices

Data Selection

Use Representative Data:

✅ Include Different Market Conditions:

• Bull markets
• Bear markets
• Sideways/ranging markets
• High volatility periods
• Low volatility periods

✅ Adequate Time Period:

• Minimum: 2-3 years
• Recommended: 5+ years
• Include at least one full market cycle

✅ Multiple Symbols:

• Test on 3-5 different symbols
• Include correlated and uncorrelated markets
• Mix asset classes if possible

Minimum Sample Sizes

For reliable statistics:

If you have < 50 trades:

• Test on longer time periods
• Test on more symbols
• Question if strategy trades enough
• Consider adjusting sensitivity

Metric Interpretation

Focus on Risk-Adjusted Returns:

❌ Wrong Focus:

"This strategy returned 150%!"
(But with 80% drawdown...)

✅ Right Focus:

"This strategy returned 35% with 12% max drawdown
and Sharpe ratio of 1.8"
(Sustainable risk/reward)

Key Metrics Priority:

1. Sharpe Ratio (risk-adjusted return)
2. Maximum Drawdown (worst-case loss)
3. Total Return (absolute performance)
4. Win Rate (consistency)
5. Profit Factor (wins vs losses)

Forward Testing Mindset

Remember:

• ✅ Past performance ≠ future results
• ✅ Backtest shows what COULD have happened
• ✅ Real trading has costs (fees, slippage)
• ✅ Market conditions change
• ✅ Strategy may need adaptation

Before Going Live:

1. Paper trade for 30+ days
2. Start with small position sizes
3. Monitor for slippage/costs
4. Be ready to adjust or stop

Iterative Development

The Development Cycle

1. Create → 2. Test → 3. Analyze → 4. Refine → repeat

Iteration Example:

Version 1:

Strategy: MA Crossover (10/30)
Result: 15% return, 25% drawdown
Analysis: Too many whipsaws in ranging markets

Version 2:

Strategy: MA Crossover (10/30) + volume filter
Result: 22% return, 18% drawdown
Analysis: Better, but still some false signals

Version 3:

Strategy: MA Crossover (20/50) + volume filter
Result: 28% return, 12% drawdown
Analysis: More stable, fewer trades but higher quality

When to Move Forward

A strategy is ready for next phase when:

✅ Logic makes intuitive sense ✅ Sharpe ratio > 1.0 (preferably > 1.5) ✅ Maximum drawdown tolerable (<20%) ✅ Sufficient trade count (>50 trades) ✅ Works across multiple symbols ✅ Performance stable across time periods

When to Pivot

Consider a new approach when:

❌ Strategy loses money consistently ❌ Win rate < 30% (unless huge winners) ❌ Maximum drawdown > 40% ❌ Performance extremely inconsistent ❌ No logical explanation for why it would work ❌ Can’t improve after 5+ iterations

Common Mistakes to Avoid

1. Over-Complicating

Mistake:

"Use MA crossover AND RSI AND MACD AND volume AND
price action AND support/resistance AND..."

Better:

"Use MA crossover confirmed by volume"

Why: More indicators ≠ better strategy. More complexity = more ways to fail.

2. Ignoring Drawdowns

Mistake:

"This returned 100%!" 
(ignoring the 60% drawdown)

Better:

"This returned 30% with only 10% drawdown"

Why: You need to survive the drawdowns to see the returns.

3. Cherry-Picking Results

Mistake:

Testing only on bull market data
Testing only on one favorable symbol
Testing only on best time period

Better:

Test across multiple:
- Market conditions
- Symbols
- Time periods

Why: Real trading won’t be cherry-picked.

4. Parameter Over-Fitting

Mistake:

"It works perfectly with MA periods of 17.3 and 43.7!"

Better:

"It works well with MA periods around 15-20 and 40-50"

Why: Overly specific parameters won’t work in future.

5. Ignoring Transaction Costs

Mistake:

Strategy trades 50 times per day
Backtest shows 2% average profit per trade

Reality:

Trading fees: 0.1% per trade
Slippage: 0.1% per trade
Net profit: 1.8% → becomes 1.2%
After 50 trades: costs eat significant portion

Better: Factor in costs from the start.

6. Abandoning Too Quickly

Mistake:

"First test didn't work, strategy is bad"

Better:

"First test showed issues. Let me adjust
parameters and try different timeframes"

Why: Good ideas often need refinement.

7. Chasing Perfect

Mistake:

Tweaking forever to get that extra 0.5% return

Better:

Strategy is profitable and robust. Moving to
paper trading to validate in real market

Why: Perfection is the enemy of progress.

Documentation and Organization

Keep a Trading Journal

Document Each Strategy:

Strategy Name: MA Crossover 20/50 v3
Date Created: 2024-01-15
Core Concept: Trend following with MA crossover
Indicators: 20 SMA, 50 SMA
Entry: 20 crosses above 50
Exit: 20 crosses below 50
Parameters Tested: 10/30, 15/40, 20/50, 25/75
Best Results: 20/50 on SPY (28%, Sharpe 1.8, DD 12%)
Notes: Works well in trending markets, avoid ranging periods
Next Steps: Add volume filter to reduce whipsaws

Track Your Learning

What Worked:

• Document successful approaches
• Note which indicators work well together
• Record optimal parameter ranges

What Didn’t Work:

• Save failed strategies too (learn from mistakes)
• Note why they failed
• Avoid repeating same mistakes

Advanced Tips

Combining Strategies

Once you have multiple successful strategies:

Portfolio Approach:

Strategy A: Trend-following (35% allocation)
Strategy B: Mean-reversion (35% allocation)
Strategy C: Momentum (30% allocation)

Benefit: Diversification reduces overall risk

Adaptive Strategies

Concept: Adjust strategy based on market conditions

Example:

If market volatility low:
  → Use mean reversion
If market volatility high:
  → Use trend following

Implementation: Create separate strategies, deploy based on conditions

Walk-Forward Analysis

Advanced Testing Method:

1. Optimize on Period 1 (e.g., 2020)
2. Test on Period 2 (e.g., 2021)
3. Re-optimize on Periods 1-2
4. Test on Period 3 (e.g., 2022)
5. Continue rolling forward

Benefit: Simulates realistic optimization and testing flow

Mindset and Discipline

Patience is Key

✅ Strategy development takes time ✅ Good ideas need refinement ✅ Testing reveals issues ✅ Iteration improves results

Realistic Expectations

Achievable Goals:

• 15-30% annual return
• Sharpe ratio > 1.0
• Maximum drawdown < 20%
• Win rate 45-55%

Unrealistic Goals:

• 100%+ annual return
• 90%+ win rate
• Never losing trades
• Zero drawdown

Continuous Learning

✅ Study why strategies work/fail ✅ Learn from each backtest ✅ Read about trading concepts ✅ Understand market dynamics ✅ Keep improving your approach

Quick Reference Checklist

Before Creating Strategy

• Clear idea of core concept
• Appropriate indicators identified
• Entry/exit rules defined
• Parameter ranges considered

During Interview

• Communicated clearly
• Answered all questions
• Specified exact values
• Explained logic

After Code Generation

• Reviewed code explanation
• Verified parameters make sense
• Understood strategy logic

Before Execution

• Appropriate data selected
• Parameters configured
• Initial capital set
• Ready to analyze results

After Results

• Analyzed all metrics
• Studied equity curve
• Checked trade count
• Evaluated drawdowns
• Documented findings

Optimization Phase

• Changed one variable at a time
• Tested systematically
• Validated on multiple markets
• Avoided over-fitting
• Kept strategy simple

What’s Next?

Now that you know best practices:

• Troubleshooting: Solve common issues
• Glossary: Reference for terms
• Creating Strategies: Apply these practices

Next: Review Troubleshooting for solutions to common problems.