TX Investment Report vs Quantitative Analysis - Side-by-Side Comparison

Date: December 3, 2025

Quick Verdict Comparison

Expected Return Analysis

Key Disagreement: Report assumes 7.17% dividend yield is maintained. Quantitative analysis assigns 80% probability of dividend cut, reducing expected yield to 4.39%.

Impact: Expected total return is 2.78% lower than report implies from dividend component alone.

Risk Metrics Comparison

Key Disagreement: Report's "20-25% maximum drawdown" is far too optimistic for a stock with 35% annual volatility. Monte Carlo simulation shows 9-15% probability of 50-70% drawdowns.

Impact: Investors may be unprepared for the psychological toll of -40-50% declines.

Factor Exposure Analysis

Report's Narrative vs Quantitative Decomposition

Factor Model Expected Return:

E(R) = 4.50% (Rf) + 7.74% (Market) + 6.00% (Value) - 2.50% (Quality) + 3.00% (Momentum) + 4.39% (Dividend)
     = 23.13% annualized

Key Disagreement: Report characterizes quality as "acceptable" based on strong balance sheet. Quantitative analysis penalizes quality due to poor ROE (4.9% vs 15% normalized) and extreme earnings volatility (Q3 -88% miss is still within 1.1σ of normal variance—meaning earnings are THAT volatile).

Risk-Adjusted Return Metrics

Sharpe Ratio Comparison

*Report did not explicitly calculate Sharpe ratio or state volatility assumptions.

Reference: Elements of Quantitative Investing, Chapter 3.3 - Sharpe Ratio

SR = (E[R] - Rf) / σ
SR_TX = (23.13% - 4.50%) / 35.0% = 0.532

Quantitative Finding: TX's Sharpe of 0.532 is 60% better than market (0.333), confirming superior risk-adjusted returns. This validates the investment thesis mathematically.

But: The Information Ratio of 0.416 is below institutional quality threshold (0.50), reflecting poor earnings quality and high idiosyncratic risk.

Dividend Sustainability Analysis

Report vs Quantitative Assessment

Dividend Scenario Analysis (Quantitative)

Key Finding: Report acknowledges dividend is "unsustainable" but doesn't quantify the impact. Quantitative analysis shows 80% probability of cut, reducing expected yield to 4.39%—a material 2.78% shortfall from the assumed 7.17%.

Impact on Total Return: Subtract 2.78% from report's 15-20% expected return → 12.22-17.22% (still acceptable but less attractive).

Position Sizing Analysis

Report Recommendation vs Kelly Optimization

Kelly Criterion (Reference: Chapter 9 - Portfolio Management):

f* = (E[R] - Rf) / σ²
f* = (23.13% - 4.50%) / (0.35)² = 174.78%
Half-Kelly = 87.39%

Why Report's 5% is Correct Despite Kelly Suggesting 87%:

1. Model Uncertainty: Cyclical timing risk (22-37% probability of no recovery)
2. Earnings Volatility: Q3 -88% miss shows extreme variance
3. Dividend Cut Risk: 80% probability creates income volatility
4. Idiosyncratic Risk: 26% company-specific volatility limits concentration
5. Behavioral Risk: Can investor handle -50% drawdown on 87% position?

Quantitative Verdict: Report's 5% recommendation is prudent and reflects good risk management. This is a case where qualitative judgment correctly overrides quantitative optimization.

Earnings Quality Assessment

Report's "Acceptable" vs Quantitative "Poor"

Statistical Test (Reference: Chapter 5 - Evaluating Risk):

• Mean surprise (4 quarters): -4.25%
• Standard deviation: 78.91%
• Z-score Q3 miss: -1.06σ

Finding: While -88% miss feels catastrophic, it's only 1.1 standard deviations from mean given TX's earnings history. This means extreme variance is normal for TX, not an aberration.

Implication: Report's "acceptable" quality assessment is too generous. Quantitative analysis correctly penalizes quality with -2.50% factor drag.

Cyclical Recovery Probability

Report's "Expected 2026-27" vs Quantitative Probability Model

Mean Reversion Model:

• Current ROE: 4.86%
• Normalized ROE: 15.0%
• Reversion gap: 10.14%
• Half-life: 2 years
• Probability formula: P(recovery) = 1 - e^(-t/half-life)

Key Finding: Report's "expected recovery 2026-27" is directionally correct (63-78% probability) but doesn't quantify the 22-37% risk that recovery DOESN'T happen in that timeframe.

Impact: There's a meaningful probability the cyclical thesis is wrong or takes longer than expected.

Valuation Metrics Comparison

Graham Number Analysis

Key Disagreement: Report presents reaching Graham Number as "base case." Quantitative analysis shows it's a 31.6% probability event—closer to a favorable scenario (70th percentile) than a median outcome.

True Base Case (Median): +27% over 3 years (~8.5% annualized), not +67% to Graham Number.

Forward P/E Implied Discount Rate:

Using Gordon Growth Model:
P = E(1+g) / (r-g)
$37.68 = $7.34(1.03) / (r - 0.03)
r = 23.06%

Finding: Market is pricing in a 23% required return—far above CAPM's 12.24%. This reflects extreme skepticism about earnings recovery.

Monte Carlo Simulation Results

Report Did Not Model, Quantitative Analysis Provides Full Distribution

3-Year Holding Period (10,000 simulations):

Probability of Key Outcomes:

• Any profit: 65.5%
• Reaching Graham Number ($63.16): 31.6%
• 50%+ gain: 39.0%
• 50%+ loss: 9.0%
• Dividend maintained: 20.0%

Value Added by Quantitative Analysis: Report gives point estimates (15-20%) without distribution. Monte Carlo reveals the full risk/reward landscape:

• Median outcome (+27%) is reasonable
• Right tail (50%+ gains) is 39%—attractive asymmetry
• Left tail (50%+ losses) is 9%—meaningful but manageable at 5% position size

Risk Management Comparison

Stop Loss Strategy

VaR-Based Stop Loss Validation:

• 95% VaR (1-year): -34.44%
• Stop loss at -15% catches deterioration early
• At 5% position, max loss = $750 (0.75% of $100k portfolio)

Quantitative Endorsement: Stop loss at $32 is well-calibrated to catch cyclical thesis failure before catastrophic losses.

Monitoring Requirements

Value Added: Quantitative analysis adds specific, measurable exit criteria beyond the $32 stop loss.

Where Report and Quantitative Analysis AGREE ✓

5 Critical Agreements

1. Investment Merit: Both conclude TX is a BUY/ACCEPTABLE for enterprising value investors
2. Position Size: Both recommend 5% for neutral investors (report was prudent despite Kelly suggesting 87%)
3. Balance Sheet Strength: Both recognize exceptional financial condition (D/E 13.9%, CR 2.46)
4. Value Opportunity: Both identify 40% margin of safety to Graham Number as compelling
5. Cyclical Recovery Probability: Both see 2026-27 recovery as plausible (quant: 63-78% probability)
6. Entry Caution: Both recommend waiting for pullback to $34-36 or at least tactical patience
7. Stop Loss Level: Both agree on $32 as appropriate exit trigger

Where They DISAGREE ❌

5 Critical Disagreements

1. Dividend Sustainability: Report assumes 7.17% yield; Quant expects 4.39% (80% cut probability)

   • Impact: -2.78% on expected return

2. Tail Risk Magnitude: Report claims 20-25% max drawdown; Quant models 50-70% possible (9% probability)

   • Impact: Investor psychology preparation

3. Base Case Definition: Report presents Graham Number as base case; Quant shows it's 31.6% probability (favorable scenario)

   • Impact: Expectation management

4. Quality Assessment: Report says "acceptable"; Quant says "POOR" (IR 0.416 < 0.50)

   • Impact: -2.50% quality factor penalty

5. Earnings Volatility Framing: Report treats Q3 -88% miss as outlier; Quant shows it's 1.1σ (normal for TX)

   • Impact: Understanding that extreme variance is inherent, not aberrant

Net Assessment: Report Quality

What the Report Got RIGHT ✓

1. Directional Conclusion: BUY at 5% is correct
2. Position Sizing: 5% recommendation is optimal (better than Kelly's 87%)
3. Stop Loss: $32 is well-calibrated
4. Balance Sheet Analysis: Correctly identifies fortress balance sheet
5. Value Framework: 40% margin of safety calculation is accurate
6. Cyclical Timing: 2026-27 recovery is plausible (63-78% probability)
7. Entry Tactics: Patience for pullback is justified

What the Report MISSED or UNDERSTATED ⚠️

1. Dividend Cut Quantification: Didn't model the 80% probability
2. Tail Risk Distribution: Understated downside scenarios (50-70% drawdowns)
3. Expected Return Precision: Gave 15-20% range without probability distribution
4. Quality Factor Penalty: Too generous on earnings quality assessment
5. Base Case Semantics: Presented favorable scenario (Graham Number) as base case
6. Risk-Adjusted Metrics: Didn't calculate Sharpe ratio, IR, VaR
7. Implied Discount Rate: Missed the 23% market-implied rate (skepticism signal)

Quantitative Grade: B+ for Report Quality

Why B+:

• Directional conclusion is correct (BUY at 5%)
• Key fundamentals accurately identified
• Risk management framework is sound
• Position sizing is prudent

Why not A:

• Underweights dividend cut risk
• Understates tail risk magnitude
• Overstates quality factor
• Lacks quantitative rigor on probability distributions

Value Added by Quantitative Analysis:

1. Probabilistic framing (31.6% Graham Number vs "base case")
2. Dividend cut scenario analysis (4.39% expected yield vs 7.17% assumed)
3. Full return distribution (Monte Carlo with 10,000 trials)
4. Risk-adjusted metrics (Sharpe 0.532, IR 0.416, VaR -34%)
5. Factor model decomposition (market +7.74%, value +6%, quality -2.5%)

Final Synthesis: Quantitative Verdict

The Numbers Support the Thesis (with Caveats)

VERDICT: ACCEPTABLE HYPOTHESIS (B+ Grade)

Confidence Score: 80% (4/5 quantitative factors passed)

Quantitative Evidence FOR Investment:

1. Sharpe Ratio 0.532 (60% better than market) ✓
2. Expected return 23.13% is attractive ✓
3. Value factor loading is extreme (P/B 0.61) ✓
4. Balance sheet prevents bankruptcy ✓
5. Cyclical recovery probability 63-78% ✓

Quantitative Evidence AGAINST:

1. Dividend cut probability 80% ❌
2. Earnings quality poor (IR 0.416) ❌
3. Tail risk 50-70% drawdowns (9% probability) ❌
4. High idiosyncratic volatility (26%) ❌
5. Quality factor penalty -2.50% ❌

Net Assessment: The bull case quantitatively outweighs the bear case at a 5% position size. But this is high risk, high return—not "margin of safety + steady income."

Bottom Line: Report vs Quantitative

Where Report Shines ✓

• Practical wisdom: 5% position sizing despite Kelly suggesting 87%
• Risk management: $32 stop loss is well-calibrated
• Narrative clarity: Easy to understand for non-quants
• Comprehensive analysis: Multi-agent approach covered all angles

Where Quantitative Adds Value ✓

• Probabilistic precision: 31.6% Graham Number, 80% dividend cut, 63-78% recovery
• Risk metrics: Sharpe 0.532, VaR -34%, IR 0.416
• Factor decomposition: Market +7.74%, Value +6%, Quality -2.5%
• Distribution modeling: Monte Carlo shows full risk/reward landscape

Recommendation: USE BOTH ✓

For Investment Decision: Use quantitative analysis

• Sharpe 0.532 > Market 0.333 = proceed
• Expected return 23.13% with 5% position = acceptable risk/reward

For Risk Management: Use report's framework + quantitative triggers

• $32 stop loss (report)
• Dividend cut >50% exit (quantitative)
• Payout ratio >120% for 2Q exit (quantitative)

For Expectation Setting: Combine both

• Expected return: 20-25% (report's high end, quant's base)
• Dividend yield: 4-5% (assume cut happens)
• Max drawdown: 40-50% (not report's 20-25%, not quant's tail 70%)
• Probability of profit: 65% (quantitative)

The Final Word

Original Report: Directionally correct, practically wise, but understates dividend risk and tail scenarios.

Quantitative Analysis: Mathematically rigorous, adds precision, validates thesis with Sharpe 0.532 and 80% confidence.

Synthesis: Proceed with 5% position, use $32 stop loss, prepare for dividend cut, expect 20-25% returns with 40-50% volatility.

Quantitative Verdict: B+ (Good Hypothesis) - compelling risk-adjusted returns but requires discipline and risk tolerance.

Files Generated:

1. This comparison: /home/pengacau/pasar-malam/output/tx-report-vs-quantitative-comparison.md
2. Full quantitative report: /home/pengacau/pasar-malam/output/tx-quantitative-verdict-2025-12-03.md
3. Executive summary: /home/pengacau/pasar-malam/output/tx-quantitative-summary-2025-12-03.md
4. Metrics JSON: /home/pengacau/pasar-malam/output/tx_quantitative_metrics.json
5. Analysis script: /home/pengacau/pasar-malam/scripts/tx_quantitative_analysis.py

Analysis Date: December 3, 2025 Framework: Elements of Quantitative Investing (Chapters 2-5, 9)