Recommendation Engine¶

This document explains how FinWiz generates BUY, HOLD, and SELL recommendations based on composite scores and decision thresholds.

Overview¶

The recommendation engine is the final decision layer that translates quantitative scores into actionable investment recommendations. It uses clear, deterministic thresholds aligned with industry-standard grading systems.

Key Principle: Recommendations are based on composite scores that combine fundamental (40%), technical (30%), and risk (30%) analysis, with adaptive weights for quality companies (50%/25%/25%).

Recommendation Framework¶

Three-Tier System¶

FinWiz uses a simplified three-tier recommendation system:

Recommendation	Composite Score	Grade Range	Meaning
BUY	≥ 0.85	A+, A, B+, B	Strong opportunity - acquire or add to position
HOLD	0.65 - 0.85	C+, C	Mixed signals - maintain position, monitor closely
SELL	< 0.65	D, F	Weak fundamentals or elevated risk - reduce or exit position

Decision Logic¶

Python

def generate_recommendation(composite_score: float) -> str:
    """Generate investment recommendation based on composite score."""
    if composite_score >= 0.85:
        return "BUY"
    elif composite_score >= 0.65:
        return "HOLD"
    else:
        return "SELL"

Portfolio Decision Thresholds¶

When evaluating existing portfolio holdings, the threshold alignment is critical:

Python

# Portfolio Holdings Processor
if composite_score >= 0.65:
    decision = "KEEP"  # C grade or better
    rationale = "Acceptable performance - maintain position"
else:
    decision = "SELL"  # D or F grade
    rationale = "Underperforming - recommend exit"

Key Alignment: The KEEP threshold (0.65) matches the HOLD threshold to ensure consistency:

C grade (0.65-0.70): Minimum acceptable quality - HOLD new positions, KEEP existing positions
D grade (0.50-0.65): Below acceptable - Don't buy, SELL existing
F grade (<0.50): Critical issues - Strong sell signal

Grade-to-Recommendation Mapping¶

Detailed Grade Breakdown¶

Grade	Score Range	Recommendation	Action	Rationale
A+	0.95-1.00	BUY	Strong buy - maximum conviction	Exceptional across all dimensions
A	0.85-0.95	BUY	Buy - high conviction	Excellent fundamentals and technicals
B+	0.80-0.85	BUY	Buy - moderate conviction	Solid opportunity, minor weaknesses
B	0.75-0.80	BUY	Buy - acceptable	Good overall, some concerns
C+	0.70-0.75	HOLD	Hold - monitor closely	Mixed signals, watch for changes
C	0.65-0.70	HOLD	Hold - significant concerns	Minimum acceptable quality
D	0.50-0.65	SELL	Sell - multiple red flags	Below acceptable standards
F	0.00-0.50	SELL	Strong sell - critical issues	Fundamental problems or high risk

Conviction Levels¶

BUY Recommendations have varying conviction levels:

Score 0.95+: Maximum conviction - Core portfolio position
Score 0.85-0.95: High conviction - Significant allocation
Score 0.80-0.85: Moderate conviction - Standard position size
Score 0.75-0.80: Acceptable - Smaller position, more monitoring

Recommendation Rationale Generation¶

Each recommendation includes a detailed rationale explaining the decision:

Template Structure¶

Python

rationale = (
    f"{ticker} receives a {grade} grade with a composite score of {score:.2f}. "
    f"Fundamental analysis (score: {fundamental_score:.2f}) shows "
    f"ROE of {roe:.1%}, debt-to-equity of {debt_to_equity:.2f}, and revenue growth of {revenue_growth:.1%}. "
    f"Technical analysis (score: {technical_score:.2f}) indicates {trend_direction} trend with RSI at {rsi:.1f}. "
    f"Risk assessment (score: {risk_score:.2f}) shows {volatility:.1%} volatility and maximum drawdown of {max_drawdown:.1%}. "
    f"{conclusion}"
)

Conclusion Logic¶

Python

def generate_conclusion(composite_score: float, grade: str, recommendation: str) -> str:
    """Generate conclusion text based on recommendation."""
    if recommendation == "BUY":
        if composite_score >= 0.95:
            return "Exceptional investment opportunity with outstanding metrics across all dimensions."
        elif composite_score >= 0.85:
            return "Strong investment opportunity with excellent fundamentals and favorable risk profile."
        else:
            return "Solid investment opportunity despite minor weaknesses in some areas."

    elif recommendation == "HOLD":
        return "Mixed signals across fundamental, technical, and risk factors suggest a HOLD recommendation pending further developments."

    else:  # SELL
        if composite_score < 0.50:
            return "Critical issues in fundamentals, technicals, or risk profile warrant immediate exit."
        else:
            return "Weak fundamentals, unfavorable technical setup, or elevated risk profile warrant a SELL recommendation."

Real-World Examples¶

Example 1: BUY Recommendation (ASML)¶

Score Breakdown:

Composite: 0.765 (B grade)
Fundamental: 0.84 (A)
Technical: 0.72 (C+)
Risk: 0.58 (D)

Quality Company: Yes (3/3 criteria) → Adaptive weights (50/25/25)

Recommendation: BUY

Rationale:

"ASML receives a B grade with a composite score of 0.77. Fundamental analysis (score: 0.84) shows ROE of 53.9%, debt-to-equity of 0.14, and revenue growth of 2.6%. Technical analysis (score: 0.72) indicates sideways trend with RSI at 50.0. Risk assessment (score: 0.58) shows 38.9% volatility and maximum drawdown of -23.4%. Despite high volatility typical of semiconductor sector, excellent fundamentals and quality company status justify a BUY recommendation."

Key Decision Factors:

✓ Exceptional fundamentals (ROE 53.9%, Margins 29.4%, Low debt 0.14)
✓ Quality company detection triggered adaptive weighting
✓ High volatility acceptable for cyclical semiconductor sector
✓ Strong competitive moat (ASML is monopoly in EUV lithography)

Example 2: SELL Recommendation (GES)¶

Score Breakdown:

Composite: 0.564 (D grade)
Fundamental: 0.26 (F)
Technical: 0.72 (C+)
Risk: 0.58 (D)

Quality Company: No (0/3 criteria) → Standard weights (40/30/30)

Recommendation: SELL

Rationale:

"GES receives a D grade with a composite score of 0.56. Fundamental analysis (score: 0.26) shows ROE of 7.6%, debt-to-equity of 2.97, and revenue growth of 2.3%. Technical analysis (score: 0.72) indicates sideways trend with RSI at 50.0. Risk assessment (score: 0.58) shows 27.0% volatility and maximum drawdown of -23.4%. Weak fundamentals, unfavorable technical setup, or elevated risk profile warrant a SELL recommendation."

Key Decision Factors:

✗ Poor fundamentals (ROE 7.6%, Debt 297%, Margins 1.01%)
✗ Altman Z-Score 2.14 indicates bankruptcy risk
✗ Struggling retail apparel sector with intense competition
✓ Neutral technicals insufficient to override fundamental concerns

Example 3: HOLD Recommendation (Hypothetical)¶

Score Breakdown:

Composite: 0.72 (C+ grade)
Fundamental: 0.68 (C)
Technical: 0.75 (B)
Risk: 0.74 (C+)

Recommendation: HOLD

Rationale:

"Ticker receives a C+ grade with a composite score of 0.72. Fundamental analysis (score: 0.68) shows moderate financial health with room for improvement. Technical analysis (score: 0.75) indicates upward trend with positive momentum. Risk assessment (score: 0.74) shows acceptable volatility profile. Mixed signals across fundamental, technical, and risk factors suggest a HOLD recommendation pending further developments."

Key Decision Factors:

~ Moderate fundamentals - not strong enough for BUY
✓ Positive technical trends suggest potential improvement
~ Risk profile acceptable but not exceptional
⚠ Decision: Monitor closely, re-evaluate in 3-6 months

Special Considerations¶

Quality Company Premium¶

Quality companies with 2+ of the following criteria receive adaptive 50/25/25 weighting:

High ROE (≥20%)
Low Debt (Debt/Equity ≤0.5)
Strong Margins (≥15%)

This adjustment recognizes that short-term volatility is less relevant for companies with durable competitive advantages and strong fundamentals.

Impact: A company scoring 0.77 with standard weights might score 0.82 with adaptive weights, potentially pushing it from HOLD to BUY territory.

Sector-Specific Adjustments¶

Certain sectors have different risk/volatility profiles:

Sector	Typical Volatility	Typical Debt	Notes
Technology	30-40%	Low-Moderate	Higher volatility acceptable
Utilities	15-25%	High	Regulated, stable cash flows
Financials	25-35%	High (leverage)	Debt metrics different
Consumer Staples	15-25%	Moderate	Defensive characteristics
Energy	35-50%	Variable	Commodity price sensitivity

Future Enhancement: Sector-specific scoring adjustments are planned (not yet implemented).

Market Conditions¶

Recommendations should be contextualized with market conditions:

Bull Market: HOLD recommendations may warrant closer review for upgrade to BUY
Bear Market: BUY recommendations should be even higher conviction
High Volatility: Risk scores may temporarily elevate, consider fundamental strength
Sector Rotation: Technical scores may lag during rotation periods

Confidence Levels¶

Each recommendation includes a confidence level based on data quality and completeness:

Python

confidence_level = (
    0.3 * fundamental_confidence +
    0.3 * technical_confidence +
    0.2 * risk_confidence +
    0.2 * data_quality_score
)

Confidence Thresholds:

High (≥0.80): All data fresh, complete, validated
Medium (0.60-0.80): Some data gaps or age concerns
Low (<0.60): Significant data issues, treat with caution

Recommendation Override: If confidence <0.50, recommendation should be downgraded:

BUY → HOLD (insufficient confidence for entry)
HOLD → HOLD (maintain position but monitor)
SELL → SELL (exit regardless of confidence)

Implementation Guidelines¶

For Portfolio Managers¶

New Positions: Only initiate on BUY recommendations (score ≥0.85)
Existing Positions: KEEP on HOLD (≥0.65), exit on SELL (<0.65)
Position Sizing: Scale based on score within BUY range:
0.95+: Full position
0.85-0.95: 75-100% position
0.75-0.85: 50-75% position
Rebalancing Triggers: Re-evaluate when score crosses threshold ±0.05

For Individual Investors¶

Risk Tolerance Adjustment:
Conservative: Only BUY on score ≥0.90, SELL on <0.70
Moderate: Standard thresholds (BUY ≥0.85, SELL <0.65)
Aggressive: BUY on ≥0.80, SELL on <0.60
Time Horizon:
Long-term (5+ years): Emphasize fundamentals, tolerate higher volatility
Medium-term (1-5 years): Balanced approach, standard weights
Short-term (<1 year): Emphasize technicals and risk, consider tighter thresholds

Limitations and Disclaimers¶

Important Limitations:

Backward-Looking: Scores based on historical data, not predictive guarantees
Quantitative Only: Does not include qualitative factors (management quality, brand value, ESG)
Market Agnostic: Does not account for overall market conditions or sector rotation
Data Dependent: Recommendations only as good as underlying data quality
Not Financial Advice: Educational tool, not personalized investment advice

Best Practices:

Use as one input among many in investment decision process
Validate with independent research and analysis
Consider personal financial situation, goals, and risk tolerance
Monitor positions regularly, don't set-and-forget
Consult qualified financial advisor for personalized guidance

Deep Analysis Scoring System - How scores are calculated
Investment Methodology - Overall philosophy
Risk Framework - Risk assessment details
Data Quality - Data validation standards

References¶

Academic Research¶

Modern Portfolio Theory: Markowitz (1952) - Risk-return optimization
Efficient Market Hypothesis: Fama (1970) - Information and pricing
Behavioral Finance: Kahneman & Tversky - Decision biases

Industry Standards¶

CFA Institute: Standards of practice for investment recommendations
FINRA Rule 2111: Suitability requirements for recommendations
SEC Regulation Best Interest: Fiduciary standards

Last Updated: November 2025 Version: 2.0 (Post-Threshold Alignment) Maintainer: FinWiz Development Team

Recommendation Engine¶

Overview¶

Recommendation Framework¶

Three-Tier System¶

Decision Logic¶

Portfolio Decision Thresholds¶

Grade-to-Recommendation Mapping¶

Detailed Grade Breakdown¶

Conviction Levels¶

Recommendation Rationale Generation¶

Template Structure¶

Conclusion Logic¶

Real-World Examples¶

Example 1: BUY Recommendation (ASML)¶

Example 2: SELL Recommendation (GES)¶

Example 3: HOLD Recommendation (Hypothetical)¶

Special Considerations¶

Quality Company Premium¶

Sector-Specific Adjustments¶

Market Conditions¶

Confidence Levels¶

Implementation Guidelines¶

For Portfolio Managers¶

For Individual Investors¶

Limitations and Disclaimers¶

Related Documentation¶

References¶

Academic Research¶

Industry Standards¶