Skip to main content

BB_Strategy04 Strategy: Double-Track Bollinger Bands "Extreme Bottom-Fishing"

Nickname: Extreme Challenger
Profession: Specializes in picking up chips outside 2 standard deviations
Timeframe: 1 hour

1. What Is This Strategy?​

Simply put, BB_Strategy04 is a strategy that:

  • Uses 72-period (3 days) Bollinger Bands, draws two channels (1σ and 2σ)
  • Specializes in bottom-fishing near the 2 standard deviation lower band
  • Runs when price rebounds to the 2 standard deviation upper band

Like an extreme sports enthusiast — it goes where others dare not: when price drops outside 2σ, it says "opportunity is here" 🎢

2. Core Configuration: Wide Stoploss + High Return Targets​

Take-Profit Rules (ROI Table)​

Right after buy    → Run at 22.6% profit
3 hours later      → Run at 6.27% profit
10 hours later     → Run at 3.77% profit
33.4 hours later   → Run at break-even

Translation: This strategy wants to make 22.6% right after buying — pretty ambitious! But the longer time goes, the more it loses its temper — after 33 hours, it'll accept break-even.

Comment: These numbers are precise to 14 decimal places... can you believe it? 22.597784040439192%? This is obviously "memorized answers" from optimization 😅

Stoploss Rules​

Fixed Stoploss: -32.5% (only admit defeat after losing one-third)
Trailing Stoploss: Enabled but not finely tuned

Translation: The stoploss is super loose — only cut losses at 32.5%. This... is either super confident, or gambler mentality 🎰

3. One Buy Condition: Picking Up Bargains at the Cliff Edge​

This strategy's buy logic is just one thing: price breaks below the 2 standard deviation lower band.

🎯 The Only Buy Condition​

Buy Conditions:
1. Close < 2σ Lower Band (price breaks below 2 std dev channel)
2. Close > 2σ Lower Band × 67.5% (but don't drop too much)

In Plain English:

"Price broke below the 2 standard deviation lower band! This is extreme oversold, high probability of rebound. But wait... don't drop too far from the lower band (more than 32.5%), if it drops too hard it might be a crash, I'm out."

Buy Zone Diagram:

        ═══════════════ 2σ Upper Band ← Sell Zone
              │
        ─────────────── 1σ Upper Band
              │
        ················ Middle Band
              │
        ─────────────── 1σ Lower Band
              │
        ═══════════════ 2σ Lower Band
              │
              ↓ ★ Buy Zone!
              │ (Price < 2σ lower band, but > 2σ lower band × 67.5%)
              │
              ✗ Forbidden Zone! Too dangerous!

Why 72 Periods?​

window = 24 * 3  # 72 hours = 3 days

In Plain English: Look at the past 3 days of data to draw Bollinger Bands, filter out those intraday small fluctuations, only focus on truly large deviations.

4. Protection Mechanisms: Super Tolerant "Life Insurance"​

Protection TypeParameterPlain English
Fixed Stoploss-32.5%"Only admit defeat after losing one-third, tough enough!"
Trailing StoplossEnabled but not tuned"It exists, but wasn't carefully configured"

Comments:

  • 32.5% stoploss... this means if you buy $100, you don't cut until it drops to $67.5 💀
  • This is either super long-term mentality, or poor risk control
  • Suggestion: Change to 15-20% is more reasonable

5. Sell Logic: Wait for Price to Reach the Clouds​

5.1 The Only Sell Signal​

Sell Condition:
Close > 2σ Upper Band

In Plain English:

"Price broke through the 2 standard deviation upper band! This means it rose too aggressively, high probability of pullback. Sell now and lock in profits!"

5.2 Four Defensive Lines​

Exit MethodTrigger ConditionPlain English
ROI ExitReached profit target"Made enough, run!"
Trailing StoplossPullback triggered"Protect profits!"
Fixed Stoploss32.5% loss"Really can't hold on anymore..."
Signal SellBreaks above 2σ upper band"Rose too much, run!"

6. This Strategy's "Personality Traits"​

✅ Strengths (Praise Section)​

  1. Double Channel Design: Draws both 1σ and 2σ simultaneously, more intuitive for charting
  2. Captures Extremes: Specializes in bottom-fishing outside 2σ, opportunities are rare but high quality
  3. Long-period Filtering: 72 periods filters noise, signals more reliable
  4. Simple Logic: Buy below lower band, sell above upper band, clear and straightforward
  5. Buy Protection: Has decline limit, avoids catching falling knives

⚠️ Weaknesses (Complaint Section)​

  1. ROI Parameters Too Precise: 14 decimal places, obviously "memorized answers" from optimization 😂
  2. Stoploss Too Wide: 32.5% stoploss, how much money do you want to lose before stopping?
  3. RSI Calculated for Nothing: Code calculates RSI but doesn't use it at all, waste of computational resources
  4. Trailing Stoploss Not Tuned: Enabled but default parameters, same as not configured
  5. No Trend Judgment: Doesn't care about major direction, might get buried抄底 against trend

7. When to Use It: Applicable Scenarios​

Market EnvironmentRecommended ActionReason
Range-bound (high volatility)✅ Strongly RecommendedBest scenario, captures extreme deviations
Range-bound (low volatility)⚠️ Okay to UseFew signals, scarce opportunities
Slow Bull (steady uptrend)❌ Don't UseMight sell too early, miss后续 gains
Sharp Decline (waterfall)❌ Absolutely Don't UseStoploss too wide, might lose badly
Single-sided Trend❌ Don't UseNo trend judgment, counter-trend operations

8. Summary: How Good Is This Strategy Really?​

One-Sentence Evaluation​

"Extreme challenger, specializes in picking up bargains outside 2σ. But stoploss too wide, ROI parameters overfitted, use with caution."

Who Should Use It?​

  • ✅ Traders who like extreme operations
  • ✅ People doing high-volatility range-bound markets
  • ✅ People willing to adjust stoploss themselves
  • ✅ People wanting to learn Bollinger Bands 2σ usage

Who Should NOT Use It?​

  • ❌ Risk-averse traders
  • ❌ People who don't want to tune parameters themselves
  • ❌ Trend traders
  • ❌ People追求 automation (parameters overfitted)

My Suggestions​

  1. Tighten Stoploss: Change -32.5% to -15% or -20%
  2. Simplify ROI: Those numbers precise to 14 decimal places are too fake, make them simpler
  3. Add Trend Filter: At least add a moving average direction judgment
  4. Small Position Test: Don't go all-in immediately, test the waters first

9. What Markets Can This Strategy Make Money In?​

9.1 Core Logic: Mean Reversion After Extreme Deviation​

BB_Strategy04 is a typical extreme mean reversion strategy. Its profit philosophy:

"Price stays within 2σ 95% of the time, once it runs outside it's an extreme deviation. After extreme deviation, high probability of regression — I'll eat this regression profit."

Statistical Support:

  • 1σ covers approximately 68% of prices
  • 2σ covers approximately 95% of prices
  • Breaking 2σ is a low-probability event (approximately 5%)

Therefore: Signals are few, but theoretically "high quality".

9.2 Performance in Different Markets (Plain English Version)​

Market TypePerformance RatingPlain English Explanation
📈 Slow Bull Market⭐⭐☆☆☆Sells too early, misses后续 gains
🔄 Range-bound Market (High Volatility)⭐⭐⭐⭐⭐Best scenario! Regression profits after extreme deviations
📉 Sharp Decline Market⭐☆☆☆☆Stoploss too wide, might get buried
⚡️ Single-sided Trend⭐☆☆☆☆No direction judgment, counter-trend operations bleed money

One-Sentence Summary: High-volatility range-bound markets are home turf, other markets are disasters.

10. Want to Run This Strategy? Check These Configurations First​

10.1 Trading Pair Configuration​

Configuration ItemRecommended ValueComment
Timeframe1hKeep default
VolatilityHigh volatility instrumentsLow volatility instruments have too few signals
Number of Pairs5-20 pairsSignals are already few, watch more instruments

10.2 Hardware Requirements (Almost None)​

This strategy's computational requirements are very small:

Number of PairsMinimum RAMRecommended RAMExperience
1-10 pairs1GB2GBSilky smooth
10-50 pairs2GB4GBMore than enough

10.3 Backtest vs Live Trading​

Where's the Trap?

  1. ROI Parameters Overfitted: Precise to 14 decimal places, live trading probably won't work
  2. Stoploss Too Wide: Backtest might hold through, can you hold through in live trading?
  3. Extreme Events: Breaking 2σ might be a prelude to crash, not an opportunity

Recommended Process:

  1. Simplify ROI: Change those ridiculous numbers to simpler ones
  2. Tighten Stoploss: Change to -15% to -20%
  3. Demo Test: Run for 1-2 months first
  4. Small Position Live: Confirm viability before increasing position

Don't be deceived by backtests! Those precise parameters are evidence of "memorizing answers".

11. Easter Egg: The Strategy Author's "Little Tricks"​

Look closely at the code and you'll find some interesting things:

  1. ROI Parameters Ridiculously Precise:

    "0": 0.22597784040439192  # 14 decimal places?

    "I ran hyperparameter optimization countless times, finally found these magical numbers. Historical backtest perfect, future... we'll see."

  2. RSI Calculated for Nothing:

    dataframe['rsi'] = ta.RSI(dataframe)  # Calculated
    # But completely unused in buy/sell conditions...

    "Already calculated, might as well keep it, who knows if I'll use it later?"

  3. Stoploss Setting:

    stoploss = -0.32530922906811843  # Again precise to 14 decimal places

    "32.5% stoploss is the optimal solution from optimization... or maybe I just want losses to look smaller?"

  4. Buy Protection:

    close > bb_lowerband2 * (1 + self.stoploss)
    # Meaning: Price not below 67.5% of lower band

    "Okay to buy below 2σ, but don't drop too far — far enough that even stoploss can't save you."

12. Last But Not Least​

One-Sentence Evaluation​

"Double standard deviation Bollinger Bands strategy, logic makes sense but parameters overfitted. Must make major changes before live trading."

Who Should Use It?​

  • ✅ People wanting to learn Bollinger Bands 2σ usage
  • ✅ People doing high-volatility range-bound markets
  • ✅ People willing to tune parameters themselves
  • ✅ Patient players who can accept few signals

Who Should NOT Use It?​

  • ❌ People wanting to use backtest parameters directly
  • ❌ Risk-averse traders
  • ❌ People追求 high-frequency signals
  • ❌ Trend traders

Manual Trader Suggestions​

If you're trading manually, you can use this strategy's approach like this:

  • Use 72-period, 2 std dev Bollinger Bands
  • Buy: Price breaks below 2σ lower band
  • Sell: Price breaks above 2σ upper band
  • Stoploss: Set 15-20%, don't use 32.5%!
  • Filter: Add a trend judgment, don't抄底 against trend

13. ⚠️ Risk Reminder Again (Must Read This Section)​

Backtests Look Great, Be Cautious in Live Trading​

BB_Strategy04's historical backtest might show high returns — but here's the big problem:

ROI parameters precise to 14 decimal places, this is typical evidence of overfitting! The strategy "memorized" historical data's "answers", won't work with a different test.

Simply put: Memorized answers for the exam, exposed when given a different test paper.

Hidden Risks of Wide Stoploss​

32.5% stoploss means:

  • Buy $100, don't admit defeat until it drops to $67.5
  • Single-trade maximum loss can reach one-third of principal
  • A few consecutive losses can cripple the account

This stoploss setting is either super long-term mentality, or poor risk control.

Risks of Extreme Bottom-Fishing​

Breaking 2σ might be an opportunity, or it might be the beginning of a crash:

  • Normal Deviation: Price temporarily deviates, quickly回归 ✅
  • Crash Prelude: Price deviation is the start of a crash ❌

Problem: You never know which one it is.

My Suggestions (Honest Truth)​

1. Absolutely do not use backtest ROI parameters directly
2. Change stoploss to -15% or -20%
3. Add a trend filter, don't抄底 against trend
4. Small position test, observe for at least 1-2 months
5. Signals are already few, don't expect high-frequency trading

Remember: The strategy's stoploss setting is 32.5%, do you dare to use it? Light position testing, staying alive is most important! 🙏

Final Reminder: Double standard deviation Bollinger Bands makes sense, but this strategy's parameter settings are obviously overfitted. If you want to use it, change parameters first before live trading!