Skip to main content

LinRegBisect_M Strategy: In-Depth Analysis

Strategy Number: #221 (the 221st of 465 strategies)
Strategy Type: Linear Regression / Trendline Analysis / Multi-Timeframe Verification
Timeframe: 15 Minutes (15m)

I. Strategy Overview​

LinRegBisect_M is an improved version of the linear regression-based trend strategy LinRegBisect. This strategy introduces a multi-timeframe verification mechanism on top of the original, enhancing signal reliability.

The name "LinReg" stands for Linear Regression, "Bisect" refers to using the regression line to dichotomize the price space, and "_M" indicates the Multi-timeframe enhanced version. This statistically grounded approach, combined with multi-timeframe verification, enables more accurate identification of trend reversal points.

Core Characteristics​

AttributeDescription
Entry Conditions2 sets of linear regression-based buy signals + multi-timeframe verification
Exit Conditions1 set of base exit signals + take-profit and stop-loss
Protections1 set of entry protection parameters
Timeframe15-minute primary timeframe + 1-hour informational timeframe
Dependenciesscipy, pandas, numpy

II. Strategy Configuration Analysis​

2.1 Base Risk Parameters​

# ROI Exit Table
minimal_roi = {
    "0": 0.12,
    "30": 0.08,
    "60": 0.04,
    "120": 0
}

# Stop Loss Settings
stoploss = -0.10

# Trailing Stop
trailing_stop = True
trailing_stop_positive = 0.03
trailing_stop_positive_offset = 0.035

Design Philosophy:

  • Higher Take-Profit Target: 12% initial target, paired with multi-timeframe verification to improve accuracy
  • Standard Stop Loss: -10% stop loss magnitude
  • Trailing Stop: Enabled to lock in profits

2.2 Linear Regression Parameters​

# Linear Regression Parameters
lr_window = 30         # Regression window size
lr_threshold = 0.012   # Slope threshold (stricter than original)
bisect_deviation = 0.025  # Deviation threshold (more relaxed than original)

III. Entry Conditions Details​

3.1 Multi-Timeframe Verification Mechanism​

The core innovation of this strategy is the introduction of the 1-hour timeframe as a confirmation layer:

# Multi-Timeframe Verification
# 1. 15-minute chart: Regression line slope turns positive
# 2. 1-hour chart: EMA60 direction is upward
# Both conditions must be met simultaneously to trigger a buy signal

3.2 Linear Regression Conditions​

Condition #1: Regression Line Slope Turns Positive + Large Cycle Confirmation​

# Logic
- 15-minute: Calculate linear regression over the most recent 30 candles
- Regression line slope transitions from negative to positive
- 1-hour: EMA60 direction confirms upward
- Entry Timing: Both timeframe signals are aligned

Condition #2: Price Deviation Below Regression Line + Large Cycle Oversold​

# Logic
- Price deviates significantly below the regression line
- Deviation exceeds the 2.5% threshold
- 1-hour: RSI < 40 oversold confirmation
- Entry Timing: Deviation reversal with large cycle support

IV. Exit Logic Details​

4.1 Multi-Layer Take-Profit System​

Take-Profit Point    12%   Hold 0-30 minutes
Take-Profit Point    8%    Hold 30-60 minutes
Take-Profit Point    4%    Hold 60-120 minutes
Stop-Loss Point      -10%  Any time

4.2 Trailing Stop​

When profit exceeds 3%, a trailing stop automatically activates, locking in at least 3.5% profit.

V. Technical Indicator System​

5.1 Core Indicators​

Indicator CategorySpecific IndicatorPurpose
Trend IndicatorLinear Regression SlopeTrend direction and strength
Deviation IndicatorPrice-Regression Line DeviationOverbought/oversold assessment
Confirmation IndicatorEMA60 (1h)Large cycle trend confirmation
Oversold IndicatorRSI (1h)Large cycle extreme positions

5.2 Multi-Timeframe Verification Principle​

# Small timeframe is used for entry precision
# Large timeframe is used for trend direction confirmation
# The combination reduces false signals

VI. Risk Management Highlights​

6.1 Dual Timeframe Filtering​

The multi-timeframe verification mechanism significantly reduces false signal probability, but may decrease trade frequency.

6.2 Slope Threshold Adjustment​

Compared to the original LinRegBisect, the _M version uses a stricter slope threshold (0.012 vs 0.01), requiring more definitive trend signals.

VII. Strategy Strengths and Limitations​

✅ Strengths​

  1. Multi-Timeframe Verification: Large cycle confirmation improves signal reliability
  2. Objective and Quantitative: Statistical method-based, high objectivity
  3. Clear Trends: Slope clearly reflects trend direction
  4. Adaptability: Adjustable parameters to suit different markets

⚠️ Limitations​

  1. Reduced Trade Frequency: Multi-timeframe filtering reduces trading opportunities
  2. Linear Assumption: Assumes trend is linear, which may not be accurate
  3. Inherent Lag: Regression line has natural lag
  4. Parameter Sensitivity: Parameters need to be adjusted according to market conditions

VIII. Applicable Scenarios Recommendations​

Market EnvironmentRecommended ConfigurationNotes
Stable TrendWindow 30, Threshold 0.012Most effective when trend is clear
High VolatilityIncrease deviation threshold to 3%Avoid false signals
Ranging MarketDisable multi-timeframe verificationIncrease trading opportunities

IX. Applicable Market Environment Details​

LinRegBisect_M is the multi-timeframe enhanced version of LinRegBisect. Code volume is approximately 200 lines, based on statistical methods combined with multi-timeframe analysis.

Its Money-Making Philosophy: Use large cycle trend confirmation to filter out noise from the small cycle

  • Multi-Timeframe Verification: Small cycle entry, large cycle confirmation
  • Strict Filtering: Fewer trades, higher win rate
  • Trend Following: Only trade in clear trends

9.1 Performance in Different Market Environments​

Market TypePerformance RatingAnalysis
📈 Slow Bull⭐⭐⭐⭐⭐Clear trends, multi-timeframe confirmation effective
🔄 Ranging⭐⭐⭐☆☆High-frequency false signals filtered
📉 Crash⭐⭐☆☆☆Trend reversals difficult to capture
⚡ Fast Bull⭐⭐⭐⭐☆Lag during sharp rallies

9.2 Key Configuration Recommendations​

Configuration ItemRecommended ValueNotes
lr_window30Regression window size
lr_threshold0.012Slope threshold
bisect_deviation0.025Deviation threshold

X. Important Notes: The Cost of Complexity​

10.1 Learning Curve​

Multi-timeframe strategies require understanding the relationships between different timeframes. Testing on a demo account is recommended first.

10.2 Hardware Requirements​

Number of Trading PairsMinimum RAMRecommended RAM
5-101GB2GB
20-302GB4GB

10.3 Backtesting vs Live Trading Differences​

Multi-timeframe strategies may perform better in backtesting than live trading because multi-timeframe relationships in historical data are easier to realize.

10.4 Manual Trading Recommendations​

It is recommended to read large cycle charts to confirm trend direction before referring to small cycle signals for entry.

XI. Summary​

LinRegBisect_M is a multi-timeframe enhanced linear regression trend strategy. Its core value lies in:

  1. Improved Signal Reliability: Large cycle confirmation reduces false signals
  2. Statistical Foundation: Objective quantitative methods
  3. Clear Trends: Slope clearly reflects market direction
  4. Adjustable Parameters: Adaptable to market conditions

For quantitative traders, this strategy is suitable for investors who pursue high win rates and are willing to sacrifice trade frequency. It is recommended for use in markets with clear trends, with sufficient backtesting validation.