Skip to main content

Best Practices & Tips

Building strategies is part science, part art. This guide shares proven principles and hard-learned lessons to help you create effective, profitable strategies.

The Golden Rules​

Rule 1: Keep It Simple​

Principle: Simple strategies often outperform complex ones.

Why:

  • Fewer conditions = more signals = better statistics
  • Easier to understand what's working (or not)
  • Less likely to be over-fitted to past data
  • Easier to explain and maintain

In Practice:

  • Start with 1-2 conditions per side
  • Add complexity only if backtests show clear improvement
  • If a 6-condition strategy performs the same as a 2-condition one, use the simpler version

Example:

  • ✅ Good: RSI < 30 (buy), RSI > 70 (sell)
  • ❌ Over-complex: RSI < 30 AND BB touches lower AND MFI < 20 AND ADX > 25 AND MACD crosses above signal...

Rule 2: Match Strategy to Market Type​

Principle: Different strategies work in different market conditions.

Market Types:

Trending Market (ADX above 25)

  • Use: Trend-following strategies
  • Examples: SMA/EMA crossovers, MACD, Supertrend
  • Avoid: Mean reversion strategies (fight the trend)

Ranging Market (ADX below 20)

  • Use: Mean reversion strategies
  • Examples: RSI, Bollinger Bands, Stochastic
  • Avoid: Trend-following (constant whipsaws)

High Volatility

  • Use: Volatility-based strategies
  • Examples: ATR bands, Bollinger Bands
  • Adjust: Wider stops, smaller positions

Low Volatility

  • Use: Breakout strategies (anticipate next move)
  • Examples: Bollinger Squeeze, range breakouts
  • Avoid: Strategies requiring large moves

In Practice:

  • Check ADX before deploying
  • Build multiple strategies for different regimes
  • Switch strategies when market changes (or use adaptive approaches)

Rule 3: Backtest Thoroughly​

Principle: Never trade a strategy you haven't tested.

Minimum Testing Requirements:

  • Test on at least 6-12 months of data
  • Include different market conditions (bull, bear, sideways)
  • Verify minimum 50-100 trades for statistical significance
  • Check performance on different assets

What to Look For:

  • Win rate above 45% (trend) or 55% (mean reversion)
  • Profit factor above 1.5
  • Maximum drawdown you can stomach emotionally
  • Consistent performance across different periods

Red Flags:

  • All wins in one month, losses spread out (luck, not edge)
  • Perfect equity curve (over-fitted)
  • Works on one asset only (not robust)
  • Less than 20 trades in 6 months (insufficient data)

In Practice:

  1. Build strategy in Strategy Builder
  2. Save it
  3. Run backtest on 12 months
  4. If good, test on different 12-month period
  5. If still good, paper trade 2-4 weeks
  6. Only then consider live deployment

Rule 4: Use ALL Logic for Entries, Consider ANY for Exits​

Principle: Be conservative entering, aggressive exiting.

Entry (Buy Conditions):

  • Use ALL logic
  • Require confirmation from multiple indicators
  • Reduces false signals
  • Higher quality trades

Exit (Sell Conditions):

  • Consider ANY logic
  • Multiple exit opportunities
  • Protects profits
  • Cuts losses quickly

Example Strategy:

Buy Conditions (ALL):

  • RSI < 35 AND
  • Price touches lower Bollinger Band AND
  • MFI < 30

(All three must confirm = high conviction entry)

Sell Conditions (ANY):

  • RSI > 70 OR
  • Price touches upper Bollinger Band OR
  • Price drops 5% from entry

(Any one triggers exit = quick profit-taking or stop-loss)

Result: Selective entries, flexible exits = better risk/reward.

Rule 5: Don't Over-Optimize​

Principle: Parameters that work perfectly on past data often fail on future data.

The Over-Optimization Trap:

You test RSI period values:

  • Period 13: 45% win rate
  • Period 14: 62% win rate ← "Perfect!"
  • Period 15: 47% win rate

You choose 14 because it's "optimal." But in live trading, it performs like a 45% strategy.

Why This Happens:

  • You fitted to noise, not signal
  • Past performance was luck
  • Market conditions change

How to Avoid:

  • Use standard parameters (14 for RSI, 20 for BB, etc.)
  • If adjusting, use round numbers (10, 20, 50, not 17 or 23)
  • Test optimized parameters on out-of-sample data
  • Prefer robust strategies that work with a range of parameters

In Practice:

  • ✅ Good: RSI period 14 (standard, battle-tested)
  • ❌ Over-optimized: RSI period 17.3 (suspiciously specific)

Indicator Selection Best Practices​

Combine Different Types​

Effective Combinations:

Momentum + Volatility:

  • RSI (momentum) + Bollinger Bands (volatility)
  • Why: Momentum confirms oversold, BB confirms price extreme
  • Example: Buy when RSI < 30 AND price touches lower BB

Trend + Momentum:

  • SMA (trend) + MACD (momentum)
  • Why: SMA shows direction, MACD confirms momentum
  • Example: Buy when price above SMA 50 AND MACD crosses above 0

Price + Volume:

  • Price levels + MFI/OBV (volume)
  • Why: Price shows level, volume confirms conviction
  • Example: Buy when price below $50k AND MFI > 30 (accumulation)

Ineffective Combinations:

Two Momentum Indicators:

  • RSI + Stochastic
  • Why: Both measure the same thing (redundant)
  • Result: Adds complexity without new information

Two Moving Averages (same type):

  • SMA 20 + SMA 50
  • Why: Better to use MA crossover strategy instead
  • Result: Redundant unless doing crossover

Use Complementary Timeframes (Advanced)​

If you want multi-indicator confirmation, consider using same indicator on different timeframes:

Example:

  • Daily RSI < 40 (longer-term oversold)
  • 4-hour RSI < 30 (short-term extreme)
  • Buy when both align

Note: Strategy Builder currently works on single timeframe, but you can manually check multiple timeframes before trading.

Parameter Selection Guide​

Moving Averages (SMA, EMA, WMA)​

Common Periods:

  • 9-10: Very fast, day trading
  • 20-21: Short-term, swing trading (most popular)
  • 50: Medium-term, swing to position trading
  • 100-200: Long-term, position trading

Recommendations:

  • Start with 20 (proven sweet spot)
  • Use 50 for trend filter
  • Use 200 for major trend identification

RSI​

Period:

  • 7: Fast, day trading (noisy)
  • 14: Standard (use this)
  • 21: Slow, smoother

Thresholds:

  • 20/80: Conservative (fewer, more extreme signals)
  • 30/70: Standard (balanced)
  • 40/60: Aggressive (more signals, less extreme)

Recommendations:

  • Crypto: Use 20/80 or 25/75 (more volatile)
  • Stocks: Use 30/70 (standard)
  • Start with 14 period, 30/70 thresholds

Bollinger Bands​

Period:

  • 20: Standard (use this)
  • 10: Fast, day trading
  • 50: Slow, position trading

Standard Deviation:

  • 1.5: Tight bands (more signals, more false)
  • 2.0: Standard (balanced)
  • 2.5-3.0: Wide bands (fewer signals, more reliable)

Recommendations:

  • Start with 20 period, 2.0 std dev
  • Increase std dev in high volatility markets

MACD​

Periods:

  • Fast/Slow/Signal:
    • 12/26/9: Standard (use this)
    • 5/13/5: Faster (day trading)
    • 19/39/9: Slower (swing trading)

Recommendations:

  • Stick with 12/26/9 unless you have specific reason
  • Faster settings = more signals but more whipsaws
  • Slower settings = fewer signals but more reliable

Stochastic​

K Period:

  • 14: Standard
  • 5: Fast (day trading)

D Period:

  • 3: Standard

Thresholds:

  • 20/80: Standard

Recommendations:

  • 14/3 with 20/80 is proven default
  • Very sensitive, best for short-term trading

Other Indicators​

  • ATR: Period 14 (standard)
  • CCI: Period 20, thresholds ±100
  • ADX: Period 14, threshold 25
  • MFI: Period 14, thresholds 20/80
  • Parabolic SAR: AF step 0.02, AF max 0.2

General Rule: Start with defaults. They're defaults for a reason — they work.

Common Mistakes to Avoid​

Mistake 1: Building Strategies in Isolation​

Problem: Creating strategy without understanding market context.

Reality: Same strategy performs differently in different markets.

Solution:

  • Check current market regime (trending vs ranging)
  • Build strategies appropriate for that regime
  • Or build adaptive strategy that works in multiple regimes

Mistake 2: Chasing Perfect Win Rate​

Problem: Trying to build 90%+ win rate strategy.

Reality: High win rate often means small wins, large losses.

Solution:

  • Focus on profit factor, not win rate
  • 40% win rate with 1:3 risk/reward beats 70% with 1:1
  • Accept that losses are part of trading

Mistake 3: No Risk Management​

Problem: Great signals, no stop-loss or position sizing.

Reality: One bad trade wipes out weeks of gains.

Solution:

  • Strategy Builder creates signals, not complete system
  • Always implement stop-losses in bot settings
  • Use proper position sizing (1-2% risk per trade)
  • Remember: signals + risk management = complete strategy

Mistake 4: Trading Every Signal​

Problem: Bot trades every signal regardless of quality.

Reality: Some signals are better than others.

Solution:

  • Use ALL logic for higher quality
  • Add trend filter (only buy in uptrend)
  • Implement signal quality scoring (advanced)

Mistake 5: Ignoring Transaction Costs​

Problem: Backtest shows profit, live trading loses money.

Reality: Fees and slippage eat high-frequency strategies.

Solution:

  • Include realistic fees in backtest (0.1-0.2%)
  • Avoid strategies with 50+ trades per month unless fees are very low
  • Swing trading usually better than day trading for this reason

Mistake 6: Abandoning Strategy Too Early​

Problem: Strategy has 5 losses in a row, you switch.

Reality: Even 60% win rate strategies can have long losing streaks.

Solution:

  • Commit to minimum 100 trades before judging
  • Understand that variance exists
  • Track performance but don't react to short-term results
  • Only abandon if drawdown exceeds backtest max drawdown significantly

Mistake 7: Not Adapting to Market Changes​

Problem: Strategy worked great in 2024, fails in 2025.

Reality: Markets evolve, volatility changes, correlations shift.

Solution:

  • Monitor strategy performance monthly
  • Be willing to switch strategies when market regime changes
  • Have multiple strategies for different conditions
  • Backtest periodically on recent data

Strategy Design Checklist​

Before saving and deploying any strategy, run through this checklist:

✅ Design Quality​

  • Strategy name is clear and descriptive
  • I understand why each condition should work
  • Conditions use different indicator types (not redundant)
  • I'm using 1-4 conditions per side (not over-complex)
  • Parameters are standard or slightly adjusted, not over-optimized
  • Buy logic is ALL (confirmation) or I have good reason for ANY
  • Strategy makes logical sense for the market type I'm trading

✅ Testing Quality​

  • Backtested on at least 6-12 months of data
  • Minimum 50 trades in backtest (100+ preferred)
  • Win rate is realistic (40-70%)
  • Profit factor above 1.5
  • Maximum drawdown is acceptable to me emotionally
  • Tested on out-of-sample period and still performs
  • Tested on different assets (if planning to use on multiple)

✅ Risk Management​

  • I have defined stop-loss rules in bot settings
  • I have defined position sizing rules
  • I understand maximum loss per trade
  • I understand maximum drawdown I could experience
  • I'm prepared emotionally for losing streaks

✅ Deployment Plan​

  • Paper trading planned for 2-4 weeks before live
  • Starting with small position sizes
  • Have plan for when to increase size (if ever)
  • Have plan for when to pause/stop strategy
  • Monitoring plan in place (daily/weekly checks)

If you can't check every box, don't deploy yet. Fix the gaps first.

Advanced Tips​

Tip 1: Build Portfolios, Not Single Strategies​

Instead of: One perfect strategy

Do this: Multiple complementary strategies

Why:

  • Diversification smooths equity curve
  • Some strategies work in trends, others in ranges
  • Reduces dependence on any single approach

Example Portfolio:

  • 40% allocation: Trend-following (SMA crossover)
  • 30% allocation: Mean reversion (RSI + BB)
  • 30% allocation: Breakout (Bollinger Squeeze)

Tip 2: Use Asymmetric Logic​

Concept: Different logic for buy vs sell

Example:

  • Buy: RSI < 30 AND BB touches lower (ALL)
  • Sell: RSI > 70 OR 5% gain reached (ANY)

Result: Selective entries, multiple exit opportunities

Tip 3: Layer Your Conditions​

Instead of: All conditions equal weight

Think of: Primary signal + filters

Example:

  • Primary: MACD crosses above 0 (main signal)
  • Filter 1: Price above 50 SMA (trend filter)
  • Filter 2: RSI above 40 (not oversold in downtrend)

Setup: All three as "ALL" logic, but mentally they have different roles

Tip 4: Consider Time-Based Filters (Manual)​

Strategy Builder doesn't have time filters, but you can manually implement:

Avoid:

  • Trading in low-volume hours (3-6 AM UTC)
  • Trading during major news (check economic calendar)
  • Trading during extreme volatility events

Monitor:

  • Is your strategy performing better certain days/times?
  • Adjust your bot schedule accordingly

Tip 5: Document Your Strategies​

Keep a strategy journal:

For each strategy, record:

  • Date created
  • Rationale (why should this work?)
  • Backtest results (win rate, profit factor, max DD)
  • Parameter settings
  • Market conditions it's designed for
  • Performance notes (monthly review)

Benefits:

  • Learn what works over time
  • Avoid repeating mistakes
  • Track evolution of your trading

Progressive Learning Path​

Month 1: Foundations​

  • Build simple 1-condition strategies
  • Learn what each indicator does
  • Backtest everything
  • Focus on understanding, not profit

Month 2: Confirmation​

  • Add second condition for confirmation
  • Experiment with ALL vs ANY logic
  • Compare single vs multi-indicator performance
  • Start paper trading best strategies

Month 3: Refinement​

  • Build multi-indicator strategies
  • Test parameter variations
  • Develop strategies for different market types
  • Track paper trading results

Month 4: Deployment​

  • Go live with small positions
  • Monitor performance religiously
  • Make minor adjustments if needed
  • Continue building new strategies for testing

Month 5+: Mastery​

  • Build strategy portfolios
  • Implement adaptive approaches
  • Develop your own trading theories
  • Share insights with community

Final Thoughts​

The Best Strategy is One You:

  1. Understand — You know why it should work
  2. Trust — You've tested it thoroughly
  3. Can Execute — You can follow the signals without second-guessing
  4. Can Stomach — You can handle the drawdowns emotionally

No strategy wins every trade. Focus on:

  • Process over results
  • Consistency over perfection
  • Learning over profit (initially)
  • Risk management always

The Strategy Builder gives you the tools. Your discipline, patience, and continuous learning determine success.

What's Next?​

You now have everything you need to build, test, and deploy profitable trading strategies using Kelor's Strategy Builder.

Suggested Next Steps:

  1. Build your first simple strategy (RSI or SMA crossover)
  2. Backtest it thoroughly
  3. Refine based on results
  4. Paper trade for 2-4 weeks
  5. Deploy with small size
  6. Monitor, learn, improve

Remember: The journey of a thousand profitable trades begins with a single well-tested strategy.

Happy trading!