Skip to main content

SuperTrend Strategy Deep Dive

Strategy Number: #400 (400th of 465 strategies)
Strategy Type: Triple Supertrend indicator trend following
Timeframe: 1 hour (1h)

1. Strategy Overview​

SuperTrend is a multi-layer trend following strategy based on the Supertrend indicator. The strategy uses three sets of Supertrend indicators (buy group) and three sets of Supertrend indicators (sell group). Entry occurs when all three buy indicators simultaneously show "up" signal, and exit occurs when all three sell indicators simultaneously show "down" signal. This multiple confirmation mechanism ensures trade reliability.

Core Features​

FeatureDescription
Buy Conditions3 buy Supertrend indicators simultaneously show "up"
Sell Conditions3 sell Supertrend indicators simultaneously show "down"
Protection MechanismsTrailing stop + Stepped ROI
Timeframe1 hour (1h)
Dependenciestalib, numpy

2. Strategy Configuration Analysis​

2.1 Basic Risk Parameters​

# ROI exit table
minimal_roi = {
    "0": 0.087,      # Immediate requirement: 8.7% profit
    "372": 0.058,    # After ~6 hours: 5.8%
    "861": 0.029,    # After ~14 hours: 2.9%
    "2221": 0        # After ~37 hours: allow 0% profit exit
}

# Stop loss setting
stoploss = -0.265   # 26.5% hard stop loss

# Trailing stop
trailing_stop = True
trailing_stop_positive = 0.05
trailing_stop_positive_offset = 0.144
trailing_only_offset_is_reached = False

Design Rationale:

  • ROI table uses stepped decrease, from 8.7% down to 0%
  • Stop loss value -0.265 is moderate, working with trailing stop to protect profits
  • Trailing stop activates after 5% profit, with offset value of 14.4%
  • Does not require reaching offset value before enabling trailing stop

2.2 Hyperparameter Configuration​

# Buy hyperparameters
buy_params = {
    "buy_m1": 4, "buy_m2": 7, "buy_m3": 1,    # Supertrend multipliers
    "buy_p1": 8, "buy_p2": 9, "buy_p3": 8,    # Supertrend periods
}

# Sell hyperparameters
sell_params = {
    "sell_m1": 1, "sell_m2": 3, "sell_m3": 6,  # Supertrend multipliers
    "sell_p1": 16, "sell_p2": 18, "sell_p3": 18, # Supertrend periods
}

2.3 Hyperparameter Range Definition​

# Buy multiplier range: 1-7
buy_m1 = IntParameter(1, 7, default=4)
buy_m2 = IntParameter(1, 7, default=4)
buy_m3 = IntParameter(1, 7, default=4)

# Buy period range: 7-21
buy_p1 = IntParameter(7, 21, default=14)
buy_p2 = IntParameter(7, 21, default=14)
buy_p3 = IntParameter(7, 21, default=14)

# Sell multiplier range: 1-7
sell_m1 = IntParameter(1, 7, default=4)
sell_m2 = IntParameter(1, 7, default=4)
sell_m3 = IntParameter(1, 7, default=4)

# Sell period range: 7-21
sell_p1 = IntParameter(7, 21, default=14)
sell_p2 = IntParameter(7, 21, default=14)
sell_p3 = IntParameter(7, 21, default=14)

3. Buy Conditions Detailed Analysis​

3.1 Supertrend Indicator Principle​

The Supertrend indicator calculates dynamic support/resistance lines based on ATR (Average True Range):

Basic Formula:
Upper Band = (High + Low) / 2 + Multiplier × ATR
Lower Band = (High + Low) / 2 - Multiplier × ATR

Trend Judgment:
Price below upper band → "down" (downtrend)
Price above lower band → "up" (uptrend)

3.2 Triple Buy Confirmation Mechanism​

The strategy uses three sets of Supertrend indicators with different parameters for buy judgment:

def populate_entry_trend(self, dataframe, metadata):
    dataframe.loc[
        (
            (dataframe[f'supertrend_1_buy_{m1}_{p1}'] == 'up') &
            (dataframe[f'supertrend_2_buy_{m2}_{p2}'] == 'up') &
            (dataframe[f'supertrend_3_buy_{m3}_{p3}'] == 'up') &
            (dataframe['volume'] > 0)
        ),
        'buy'] = 1

Three Indicator Groups Comparison:

Indicator GroupMultiplier (m)Period (p)Sensitivity
Supertrend_1_buy48Medium
Supertrend_2_buy79Slower (larger multiplier)
Supertrend_3_buy18Fast (smaller multiplier)

Design Logic:

  • Three indicators cover different sensitivity levels
  • Fast indicator captures early signals
  • Medium indicator confirms trend
  • Slow indicator filters false breakouts
  • All three must confirm before entry, improving reliability

3.3 Indicator Pre-calculation Mechanism​

def populate_indicators(self, dataframe, metadata):
    # Pre-calculate all possible Supertrend combinations
    for multiplier in self.buy_m1.range:
        for period in self.buy_p1.range:
            dataframe[f'supertrend_1_buy_{multiplier}_{period}'] = ...

Optimization Features:

  • Pre-calculate all parameter combinations
  • Use cached results at runtime
  • Facilitates quick switching during hyperparameter optimization

4. Sell Logic Detailed Analysis​

4.1 Triple Sell Confirmation Mechanism​

def populate_exit_trend(self, dataframe, metadata):
    dataframe.loc[
        (
            (dataframe[f'supertrend_1_sell_{m1}_{p1}'] == 'down') &
            (dataframe[f'supertrend_2_sell_{m2}_{p2}'] == 'down') &
            (dataframe[f'supertrend_3_sell_{m3}_{p3}'] == 'down') &
            (dataframe['volume'] > 0)
        ),
        'sell'] = 1

Three Sell Indicator Groups Comparison:

Indicator GroupMultiplier (m)Period (p)Sensitivity
Supertrend_1_sell116Fast (small multiplier)
Supertrend_2_sell318Medium
Supertrend_3_sell618Slow (large multiplier)

4.2 Trailing Stop Protection​

trailing_stop = True
trailing_stop_positive = 0.05          # Activate after 5% profit
trailing_stop_positive_offset = 0.144  # Offset value 14.4%
trailing_only_offset_is_reached = False # Don't require reaching offset

Trailing Stop Logic:

  • Activates after profit reaches 5%
  • Stop loss line follows price increases
  • Triggers sell when price drops, protecting profits

5. Technical Indicator System​

5.1 Supertrend Indicator Implementation​

def supertrend(self, dataframe, multiplier, period):
    # Calculate basic upper/lower bands
    df['TR'] = ta.TRANGE(df)              # True Range
    df['ATR'] = ta.SMA(df['TR'], period)  # ATR average
    
    # Calculate basic upper/lower bands
    df['basic_ub'] = (high + low) / 2 + multiplier * ATR
    df['basic_lb'] = (high + low) / 2 - multiplier * ATR
    
    # Calculate final upper/lower bands (considering previous period state)
    df['final_ub'] = ...
    df['final_lb'] = ...
    
    # Set Supertrend value
    df[st] = ...
    
    # Mark trend direction
    df[stx] = 'up' if close > st else 'down'

5.2 Core Calculation Logic​

Supertrend's final band calculation considers the previous period's state:

# Final upper band
if basic_ub < final_ub[i-1] or close[i-1] > final_ub[i-1]:
    final_ub = basic_ub
else:
    final_ub = final_ub[i-1]  # Keep unchanged

# Final lower band
if basic_lb > final_lb[i-1] or close[i-1] < final_lb[i-1]:
    final_lb = basic_lb
else:
    final_lb = final_lb[i-1]  # Keep unchanged

6. Risk Management Features​

6.1 Triple Confirmation Filter​

Risk TypeProtection Measure
False BreakoutThree indicators must confirm simultaneously, filtering single-indicator misjudgments
Trend ReversalThree sell indicators confirm reversal signal
Profit GivebackTrailing stop locks in profits

6.2 Stepped ROI​

Time Interval       Profit Requirement
───────────────────────────────────────
0 minutes           8.7%
6 hours             5.8%
14 hours            2.9%
37 hours            0%

Design Logic:

  • High profit requirement for short-term, locking in quick profit opportunities
  • Lower requirements for long-term, avoiding forced holding for too long

6.3 Trailing Stop Parameter Analysis​

ParameterValueDescription
trailing_stop_positive0.05Activate trailing after 5% profit
trailing_stop_positive_offset0.144Offset when price rises 14.4% from entry
trailing_only_offset_is_reachedFalseDon't require reaching offset to start trailing

7. Strategy Advantages and Limitations​

✅ Advantages​

  1. Triple Confirmation Reliability: Three indicators confirm simultaneously, reducing false signals
  2. Buy/Sell Separation Design: Buy and sell use different parameter groups, flexible adjustment
  3. Hyperparameter Optimizable: 12 hyperparameter ranges, easy for Hyperopt optimization
  4. Trailing Stop Protection: Automatically trails after profit, locking in gains
  5. Calculation Pre-caching: All combinations pre-calculated, efficient runtime

⚠️ Limitations​

  1. Strict Entry Conditions: Triple confirmation may miss early opportunities
  2. Non-official Indicator Implementation: Supertrend implementation not verified against original paper
  3. Moderate Stop Loss: 26.5% stop loss may incur significant losses in sharp drops
  4. ATR Stability Dependency: ATR calculation requires sufficient historical data (startup_candle_count=18)

8. Applicable Scenario Recommendations​

Market EnvironmentRecommended ConfigurationNotes
Trending MarketDefault configurationTriple confirmation filters false breakouts
Oscillating MarketLower multipliersIncrease sensitivity
Sharp DropTighten stop lossPrevent large losses
Low Volatility MarketRaise multipliersExpand volatility range threshold

9. Applicable Market Environment Details​

SuperTrend is a multi-layer trend following strategy. Based on its triple Supertrend indicator design, it is most suitable for stable trend markets, while signals may be sparse in oscillating markets.

9.1 Strategy Core Logic​

  • Triple Confirmation Mechanism: Three Supertrend indicators with different parameters confirm simultaneously
  • Buy/Sell Separation: Buy and sell groups use different parameters, adapting to different exit strategies
  • Dynamic Support/Resistance: Dynamic bands based on ATR follow price fluctuations

9.2 Performance in Different Market Environments​

Market TypePerformance RatingReason Analysis
📈 Stable Trend⭐⭐⭐⭐⭐Triple confirmation reliably captures trend
🔄 Oscillating Market⭐⭐⭐☆☆Few signals but high reliability
📉 Sharp Drop⭐⭐☆☆☆26.5% stop loss may incur large losses
⚡ Low Volatility⭐☆☆☆☆ATR too small, indicator insensitive
⚡ High Volatility⭐⭐⭐⭐☆Large ATR, bands follow more flexibly

9.3 Key Configuration Recommendations​

Configuration ItemRecommended ValueNotes
stoploss-0.15~-0.20Tighten stop loss to prevent large losses
trailing_stop_positive0.03Increase trailing sensitivity
buy multipliers2-4Medium sensitivity
sell multipliers1-3Fast exit response

10. Important Note: The Cost of Complexity​

10.1 Learning Cost​

SuperTrend involves understanding the Supertrend indicator:

  • ATR (Average True Range) calculation and meaning
  • Supertrend upper/lower band calculation logic
  • Final band state continuation mechanism
  • Triple confirmation design philosophy

Recommended reading time: 1-2 hours for deep understanding

10.2 Hardware Requirements​

Number of PairsMinimum MemoryRecommended Memory
1-10 pairs2GB4GB
10-30 pairs4GB8GB
30+ pairs8GB16GB

Computational Burden:

  • Pre-calculate all hyperparameter combinations (~7×21×3×2 = 882 combinations)
  • Each combination requires iterating historical data to calculate bands

10.3 Differences Between Backtesting and Live Trading​

  • Backtesting Environment: Uses pre-calculation cache, fast execution
  • Live Trading Environment: Requires real-time Supertrend indicator calculation
  • Recommendation: Ensure startup_candle_count is set correctly (18)

10.4 Suggestions for Manual Traders​

Manual traders can learn from:

  • Triple confirmation concept: Multiple indicators must confirm simultaneously before entry
  • Buy/sell separation design: Use different parameters for entry and exit
  • Trailing stop usage: Enable trailing protection after profit

11. Summary​

SuperTrend is a clearly designed multi-layer trend following strategy. Its core value lies in:

  1. Triple Confirmation Reliability: Three indicators confirm simultaneously reducing false signals
  2. Buy/Sell Separation Flexibility: Entry and exit parameters independently optimized
  3. Adjustable Hyperparameters: 12 parameter ranges easy to optimize
  4. Trailing Stop Protection: Locks in profits after gaining

For quantitative traders, this is an intermediate strategy suitable for stable trend markets, with clear logic and adjustable parameters. It is recommended to find parameter combinations suitable for your trading pairs through Hyperopt optimization.

Final Reminder: No matter how good the strategy is, the market will humble you without warning. Test with small positions, survival comes first! 🙏