Dumbbell Strategy Construction Guide: A Systematic Framework to Achieve Sharpe Ratio >1.0

Based on the fund manager’s year-end summary background you provided, I will systematically analyze how to construct and execute the “dumbbell strategy” to achieve excellent risk-adjusted returns.

---

The dumbbell strategy is an

that distributes assets to two distinct ends:

• Left End (High Certainty End)
  : Provides stable returns, downside protection, liquidity buffer
• Right End (High Odds End)
  : Provides excess return potential, asymmetric return opportunities

[1]:

• Avoid the mediocrity trap of “medium risk, medium return”
• Balance safety and aggressiveness
• Maintain strategy flexibility in different market environments

Traditional Barbell strategy is mainly used in bond investments (short-term + long-term bonds) [2], while the fund manager’s

lies in:

• Expanding the risk dimension from “tenor” to the “certainty-odds” spectrum
• Applying to equity and equity investment fields
• Emphasizing discounted cash flow (DCF) as a unified valuation framework

---

:

Financial Quality Dimension:
├─ Free Cash Flow (FCF) positive and stable for 5 consecutive years
├─ ROIC > WACC (creates real economic value)
├─ Debt/EBITDA <3.0 (low leverage risk)
└─ Cash Conversion Cycle < industry median

Business Moat Dimension:
├─ Industry Position (Top 3 or niche leader)
├─ Pricing Power (able to pass on cost increases)
├─ Customer Stickiness (high switching cost or network effect)
└─ Predictability (business model less affected by macroeconomic shocks)

:

• Intrinsic value discount ≥30% (based on DCF model)
• Historical quantile <40% (5-year valuation level)
• Dividend yield >3% (provides cash return floor)

1. High-Quality Dividend Stocks
   : Utilities, consumer staples, healthcare leaders
2. Cash-Like Assets
   : Short-term treasury bonds, money market funds (provides liquidity)
3. Defensive Growth Stocks
   : Stable growth (10-15%) and reasonably valued enterprises

• Single Stock Position
  : ≤5% (avoid excessive exposure to individual stock risk)
• Industry Concentration
  : Single industry ≤30%
• Expected Volatility
  : Annualized <15%
• Maximum Drawdown Target
  : < -8%

---

:

Upside Potential / Downside Risk ≥5:1

Typical Scenarios:
├─ Mispricing Opportunities: Market overreaction, temporary negative news
├─ Industry Reversal: Cycle bottom, policy inflection point, technological breakthrough
├─ Transformation Targets: Management change, strategic adjustment, asset restructuring
└─ Emerging Fields: Early-stage growth stocks, disruptive technologies

:

• Option Value Thinking
  : Limited downside, huge upside
• Clear Catalysts
  : Visible catalytic events within 12-24 months
• Sufficient Liquidity
  : Avoid small-cap liquidity traps
• Reliable Management
  : Historical performance and integrity records

(Lessons from losses at Sunac (融创)):

• Avoid investments where “story outweighs fundamentals”
• Distinguish between “cycle reversal” and “value trap”
• Maintain extreme caution towards high-leverage businesses
• Establish disciplined rules for “stop-loss and take-profit”

:

f* = (bp - q) / b

Where:
b = Odds (profit/loss ratio)
p = Success probability
q = Failure probability (1-p)

Conservative Adjustments in Practical Application:
├─ Use half-Kelly or 1/4-Kelly (avoid over-concentration)
├─ Set maximum single position limit (≤3% for high-odds end)
├─ Dynamic adjustment (adjust positions based on win rate changes)
└─ Correlation control (avoid high correlation among high-odds targets)

:

• 10-15 high-odds targets
• Single position: 1-3%
• Total exposure:25-35%
• Remaining as cash reserves (waiting for new opportunities)

---

Asset Class	Allocation Ratio	Expected Return	Expected Volatility	Sharpe Ratio Contribution
High Certainty Core	50-60%	8-12%	10-15%	0.6-0.8
High Odds Opportunities	25-35%	15-30%	25-40%	0.4-0.6
Cash/Liquidity Reserve	10-15%	3-4%	1-2%	0.1-0.2
Total Portfolio	100%	10-15%	12-18%	>1.0

(Market overvaluation, rising volatility):

• VIX >25 or market PE >80th percentile of historical quantile
• Increase high-certainty end to 65-70%
• Reduce high-odds end to20-25%

(Market panic, mispriced opportunities):

• VIX >35 or systemic sell-off in the market
• Reduce high-certainty end to 45-50%
• Increase high-odds end to35-40%

---

:

Target: Correlation between high-certainty end and high-odds end <0.3

Implementation Methods:
├─ Industry diversification (cyclical vs defensive)
├─ Geographic diversification (A-shares, Hong Kong stocks, US stocks)
├─ Factor diversification (value, growth, quality, momentum)
└─ Time diversification (gradual position building, batch take-profit)

:

Total Risk = Risk from High-Certainty End + Risk from High-Odds End + Risk from Others

Ideal State:
├─ Risk contribution from high-certainty end ≈40%
├─ Risk contribution from high-odds end ≈40%
└─ Risk contribution from cash etc. ≈20%

Practical Operation:
├─ Regularly calculate marginal risk contribution (MCR) of each end
├─ Adjust positions based on MCR (instead of simple equal-weight allocation)
└─ Optimize portfolio weights using covariance matrix

:

1. Confirmation Bias
   :
   • Establish mandatory refutation mechanism (list 3 bearish reasons for each investment)
   • Regularly review investment assumptions vs actual changes

2. Loss Aversion
   :
   • Pre-set stop-loss rules (mandatory stop-loss for high-odds targets at -20%)
   • Focus on overall portfolio performance instead of individual targets

3. Anchoring Effect
   :
   • Use DCF intrinsic value instead of cost price as decision anchor
   • Regularly update valuation models (quarterly revaluation)

4. Overconfidence
   :
   • Record all investment decisions and logic
   • Regularly review and calculate real win rate
   • Use historical win rate as input parameter for Kelly Formula

---

Phase 1: Stock Pool Construction (10% Weight)
├─ Fundamental Initial Screening (financial quality, business moat)
├─ DCF Valuation (intrinsic value calculation)
└─ Classify into pools (certainty pool vs odds pool)

Phase 2: Position Decision (40% Weight)
├─ Calculate theoretical position using Kelly Formula
├─ Risk Adjustment (conservative treatment)
├─ Correlation Check (avoid over-concentration)
└─ Execute trades

Phase3: Continuous Monitoring (50% Weight)
├─ Quarterly Revaluation (update DCF model)
├─ Catalyst Tracking (high-odds targets)
├─ Risk Indicator Monitoring (VaR, maximum drawdown, Sharpe ratio)
└─ Psychological Discipline Execution (stop-loss and take-profit)

###6.2 Key Performance Indicator (KPI) Monitoring

:

• Sharpe Ratio (target >1.0)
• Maximum Drawdown (target <-15%, historical -13.3%) [0]
• Calmar Ratio (return/maximum drawdown, target >1.5)
• Information Ratio (excess return vs benchmark / tracking error)

:

• High-Certainty End: Win rate >60%, profit-loss ratio >2:1
• High-Odds End: Win rate 35-45%, profit-loss ratio >5:1
• Cash Flow: Annual dividend income >2% of portfolio value

---

###7.1 Limitations of AI & Human Advantages

:

• Information processing speed and breadth
• Pattern recognition and data analysis
• Emotional neutrality (no psychological bias)

(Business Essence Understanding):

• Qualitative judgment of business models
• Evaluation of management integrity
• Comprehensive reasoning of unstructured information
• Response to extreme black swan events

###7.2 Adaptability of Investment Framework in AI Era

:

Human Judgment (Business Essence)
    ↓
Set Investment Framework & Boundary Conditions
    ↓
AI Assistance (Data Verification & Monitoring)
    ↓
Quantitative Indicator Tracking & Early Warning
    ↓
Human Decision (Final Judgment & Execution)

:

• Use AI for financial data anomaly detection
• Use AI to monitor news and public opinion (catalyst tracking for high-odds targets)
• Use AI for backtesting and stress testing
• Humans are responsible for assumption setting and qualitative judgment in DCF modeling

---

###8.1 Characteristics of Successful Cases

1. High-Certainty End
   :
   • Long-term holding (3-5+ years)
   • Obvious compound interest effect
   • Dividend reinvestment enhances returns

2. High-Odds End
   :
   • Catalysts realized
   • Valuation repair (from undervalued → reasonable)
   • Timely take-profit (avoid mean reversion)

###8.2 Failure Lessons (Sunac (融创) Case)

:

• Misjudged “high odds” as “high certainty”
• Ignored cash flow deterioration signals
• High leverage amplified downside risk
• Emotional加仓 (attempted to average down)

:

• Establish mandatory DCF review mechanism
• Set maximum loss limit for single targets
• Increase safety margin requirements for high-debt industries
• Implement “stop-loss line” discipline (close position once触及, no discussion)

---

Evolving from “individual stock fundamentalists” to “system guardians”, the key transformations are:

1. From Stock Selection to Allocation Selection
   : Acknowledge limitations of predicting the future, respond to uncertainty through portfolio structure

2. From Offense to Balance
   : No longer pursue success in every investment, but pursue excellent risk-adjusted returns at portfolio level

3. From Intuition to System
   : Use the framework of “10% stock picking +40% betting +50% psychology” to transform investment into a repeatable, optimizable process

4. From History to Future
   : In the AI era, the ability to understand business essence (rather than information processing speed) will become core competitiveness

5. From Greed to Discipline
   : Recognize that discounted cash flow is the law of gravity for investment, all valuations will eventually return to fundamentals

: Build a system that remains robust in different market environments, achieving:

• Annualized Return
  :12-18%
• Sharpe Ratio
  :>1.0
• Maximum Drawdown
  :< -15%
• Sustainability
  : Repeatable for 10+ years

---

[1] Investopedia - “Barbell Investment Strategy: Definition, How It Works, and Examples” (https://www.investopedia.com/terms/b/barbell.asp)

[2] Investopedia - “Understanding the Barbell Investment Strategy: High-Risk and Safe Assets” (https://www.investopedia.com/articles/investing/013114/barbell-investment-strategy.asp)

[3] Investopedia - “Dumbbell: What It is, How It Works, Example” (https://www.investopedia.com/terms/d/dumbbell.asp)

[4] Jinling AI Financial Data - Fund Manager Year-End Summary Background Information