Nardi L. vs Wu Y.

Match & Event

Executive Summary

Totals

Game Spread

Key Risks:

• CRITICAL: Wu Y. has 0 matches played in last 52 weeks tour-level data - no reliable hold/break statistics
• Data quality: MEDIUM - missing primary statistics for Wu Y.
• Market odds unavailable - cannot calculate edge
• Best of 5 format requires 5-set game distribution modeling with insufficient data

RECOMMENDATION: PASS - Insufficient data quality to generate actionable totals or handicap recommendations.

Data Quality Assessment

Critical Data Issues

Wu Y. Data Availability:

• ✗ Main profile shows 0 matches played (last 52 weeks tour-level)
• ✗ Hold %: 0% (UNUSABLE)
• ✗ Break %: 0% (UNUSABLE)
• ✓ Clutch stats available (15 matches analyzed) - but from different dataset
• ✓ Recent form available (10 matches, likely challengers/qualifiers)

Nardi L. Data Availability:

• ✓ 16 matches played (last 52 weeks tour-level)
• ✓ Hold %: 73.9%
• ✓ Break %: 19.1%
• ✓ Complete statistics available

Impact on Modeling:

• Cannot reliably model game distributions without Wu’s tour-level hold/break rates
• Clutch stats (BP saved: 47.1%, BP conversion: 40.2%) provide rough proxies but lack precision
• Using challenger-level statistics against tour-level data creates systematic bias
• Best of 5 format increases variance - requires higher data quality threshold

Data Quality Rating: MEDIUM (insufficient for actionable recommendations)

Nardi L. - Complete Profile

Rankings & Form

Hold/Break Analysis

Game Distribution Metrics

Serve Statistics

Return Statistics

Clutch Statistics

Key Games

Playing Style

Recent Form Analysis

• Last 9 Record: 5-4 (stable trend)
• Dominance Ratio: 0.96 (near break-even)
• Avg Games/Match: 21.6
• Form Trend: Stable (but at low level)

Wu Y. - Complete Profile

Rankings & Form

Hold/Break Analysis

Proxy Statistics (From Clutch Data - Lower Quality)

Note: BP saved % is NOT equivalent to hold %. A player can save break points but still lose service games. These proxies are unreliable for game distribution modeling.

Game Distribution Metrics (From Recent Form - Likely Challengers)

Key Games (15 matches analyzed)

Playing Style

Recent Form Analysis

• Last 10 Record: 9-1 (declining trend noted - likely peaked)
• Dominance Ratio: 1.26 (but against lower competition)
• Form Context: Strong challenger/qualifying form, but no tour-level L52W data

Matchup Quality Assessment

Elo Comparison

Quality Rating: LOW (average Elo: 1473 - both below 1900)

Elo Edge: Nardi L. by 240 points

• Significant gap (>200 points) - Nardi L. is substantial favorite
• Wu Y. is 253 Elo points below Nardi on hard courts
• However: Wu’s Elo based on limited data, reliability questionable

Recent Form Analysis

Form Indicators:

• Nardi: Stable but poor form at tour level (DR 0.96 = losing games)
• Wu: Excellent at challenger level (DR 1.26) but declining trend, no tour-level data

Form Advantage: UNCLEAR - Different competition levels make comparison invalid

Competition Context:

• Nardi’s 5-4 record: Against ATP tour-level opponents
• Wu’s 9-1 record: Likely against challenger/qualifier opponents
• Direct form comparison is misleading

Data Quality Issues Prevent Full Analysis

Why PASS is Required

1. Wu Y. Has No Tour-Level Statistics (Last 52 Weeks)

• Zero matches played at tour level in database
• Hold % and Break % are fundamental for game distribution modeling
• Cannot be substituted with BP saved/conversion rates
• Challenger-level statistics systematically differ from tour-level

2. Best of 5 Format Increases Data Requirements

• 3-set model requires hold/break rates for 2-3 sets
• 5-set model requires hold/break rates for 3-5 sets
• Higher variance in 5-set matches = need tighter data quality
• Missing data for one player makes 5-set modeling unreliable

3. No Market Odds Available

• Cannot calculate edge even if model was reliable
• No benchmark for fair value assessment
• Cannot compare model to market consensus

4. Proxy Statistics Are Insufficient

Breaking down why BP saved ≠ Hold %:

Example Service Game:
- Server faces 0-40 (3 BPs)
- Saves 2 BPs (wins 2 points)
- Loses 3rd BP → Game lost

Result:
- BP Saved %: 66.7% (2/3) ✓
- Hold %: 0% (lost the game) ✗

This demonstrates why BP saved cannot proxy for hold %.
Wu's 47.1% BP saved does NOT mean 47% hold rate.
Typical relationship: Hold % = 70-85% even with 60% BP saved.

5. Data Quality Multiplier

• Completeness: MEDIUM
• Confidence multiplier: 0.6
• Any edge calculation would need to be reduced by 40%
• Combined with missing odds → cannot reach 2.5% edge threshold

What We Cannot Calculate (And Why)

Game Distribution Modeling

Requires:

• Player 1 hold %: ✓ Available (Nardi: 73.9%)
• Player 2 hold %: ✗ MISSING (Wu: 0% = no data)
• Player 1 break %: ✓ Available (Nardi: 19.1%)
• Player 2 break %: ✗ MISSING (Wu: 0% = no data)

Impact:

• Cannot model P(6-0), P(6-1), P(6-2), P(6-3), P(6-4), P(7-5), P(7-6) for each set
• Cannot calculate P(tiebreak) without both players’ hold rates
• Cannot model straight sets vs 5-set probability
• Cannot generate expected total games distribution

Totals Analysis

Cannot Calculate:

• Expected total games (requires game distribution)
• 95% confidence interval (requires distribution)
• Fair totals line (requires expected value)
• P(Over X.5) for any threshold (requires distribution)

Would Need:

• Wu’s actual tour-level hold % (not BP saved %)
• Wu’s actual tour-level break % (not BP conversion %)
• Minimum 10-15 tour-level matches for reliability
• Best of 5 historical data for both players

Handicap Analysis

Cannot Calculate:

• Expected game margin (requires game distribution)
• Fair spread line (requires margin calculation)
• P(Nardi covers -X.5) for any line
• Coverage probabilities without distribution

Additional Issues:

• Best of 5 vs Best of 3 game counts differ significantly
• Nardi’s 5-set performance unknown (only 3-set data)
• Wu’s 5-set performance unknown (no tour data at all)

Alternative Approach Considered (And Rejected)

Could We Use Elo-Only Model?

Elo Differential: Nardi +240 hard court Elo

Typical Elo-to-Spread Conversion:

• 100 Elo points ≈ 63% win probability in 3-set match
• 240 Elo points ≈ 78% win probability
• Expected set differential ≈ +0.84 sets
• Expected game margin ≈ 4-5 games (rough)

Why This Is Insufficient:

1. Elo predicts match winner, not game totals
2. Same Elo gap can produce 6-2, 6-3 (20 games) OR 7-6, 7-6 (26 games)
3. Hold/break rates determine game count, not Elo
4. Best of 5 format requires set-level modeling
5. Wu’s Elo reliability questionable (limited tour data)

Edge Requirement: 2.5% minimum

• Elo-only model CI would be ±5-6 games (extremely wide)
• Cannot achieve 2.5% edge with such wide uncertainty
• Professional modeling requires hold/break foundation

Player Comparison Matrix

Statistical Comparison (Where Available)

Key Matchup Insights

Cannot Generate Actionable Insights Without:

• Wu’s tour-level hold/break rates
• Comparable competition data
• Best of 5 historical performance
• Market odds for validation

What We Can Say:

• Nardi is substantial favorite by Elo (+240 points)
• Both players are error-prone (W/UFE < 1.0)
• Nardi struggles on serve (73.9% hold, 56.9% 1st serve in)
• Wu’s form is strong but against lower competition
• Match will likely be high-variance given both players’ styles

What We Cannot Say:

• Expected game total
• Fair spread line
• Whether totals lean over or under
• Whether Nardi covers any specific game handicap
• Any edge vs market odds (no odds available)

Why Professional Standards Require PASS

Tennis Totals Modeling Best Practices

Minimum Data Requirements:

1. ✓ Player 1 hold % (Nardi: available)
2. ✗ Player 2 hold % (Wu: MISSING)
3. ✓ Player 1 break % (Nardi: available)
4. ✗ Player 2 break % (Wu: MISSING)
5. ✗ Market odds (MISSING)
6. ✗ Best of 5 historical data for both players (MISSING)

Failed Criteria: 4 out of 6 minimum requirements

Edge Calculation Framework

For any recommendation:
1. Calculate model fair line (❌ Cannot - missing data)
2. Calculate model probability distribution (❌ Cannot - missing data)
3. Extract market odds (❌ Not available)
4. Convert market odds to no-vig probability (❌ Cannot)
5. Calculate edge: Model P - Market P (❌ Cannot)
6. If edge ≥ 2.5% → recommend (❌ Cannot reach this)
7. If edge < 2.5% → PASS (✓ PASS required)

Outcome: Cannot complete steps 1-5, therefore step 7 (PASS) is automatic.

Comparison to Standard Requirements

What a HIGH Confidence Recommendation Needs:

• ✓ Hold % both players (surface-adjusted)
• ✓ Break % both players (opponent-adjusted)
• ✓ Sample size ≥15 matches each
• ✓ Tiebreak data with ≥10 TBs each
• ✓ Market odds available
• ✓ Model-empirical alignment check
• ✓ Edge ≥ 5%
• ✓ Data quality: HIGH

What This Match Has:

• ✓ Nardi data available
• ✗ Wu tour-level data: ZERO
• ✗ Sample size: Wu has 0 tour matches
• ✗ Tiebreak data: Wu has 0 tour TBs
• ✗ Market odds: Not available
• ✗ Alignment check: Cannot perform
• ✗ Edge: Cannot calculate
• ✗ Data quality: MEDIUM (insufficient)

Failed: 7 out of 8 requirements for HIGH confidence Failed: 5 out of 6 requirements for even LOW confidence

Recommendations

Totals Recommendation

Rationale: Wu Y. has zero tour-level matches in the last 52 weeks, making it impossible to reliably estimate hold % and break % - the fundamental inputs for game distribution modeling. BP saved % (47.1%) cannot substitute for hold %, as they measure different outcomes. Additionally, market odds are unavailable, preventing any edge calculation even if the model could be constructed. Best of 5 format requires higher data quality than Best of 3, making this data gap even more critical.

Game Spread Recommendation

Rationale: Expected game margin calculation requires game distribution modeling from hold/break rates for both players. Wu’s missing tour-level statistics make this impossible. While Nardi is a substantial Elo favorite (+240 points), converting Elo to game spreads without hold/break context produces confidence intervals of ±5-6 games - far too wide to identify any 2.5%+ edge, especially without market odds for comparison.

Pass Conditions

This match fails multiple pass criteria:

1. ✗ Missing critical statistics - Wu Y. has no tour-level hold/break data
2. ✗ Data quality below threshold - MEDIUM quality insufficient for Best of 5 modeling
3. ✗ No market odds available - Cannot calculate edge or compare to consensus
4. ✗ Insufficient sample size - 0 tour matches for Wu Y. (minimum 10-15 required)
5. ✗ Format mismatch - Best of 5 requires higher data quality than available
6. ✗ Proxy statistics inadequate - BP saved/conversion cannot substitute for hold/break %

When This Match Would Become Bettable:

• Wu Y. plays 10-15 tour-level matches → establish reliable hold/break rates
• Market odds become available → enable edge calculation
• Both players have Best of 5 match history → reduce format uncertainty
• Model expected total has 95% CI within ±3 games → manageable variance

Professional Recommendation: Wait for better data quality or market opportunities. A PASS is a valid and valuable outcome when data quality cannot support the required 2.5% edge threshold.

Risk & Unknowns

Variance Drivers

Identified (But Cannot Quantify):

• Tiebreak Volatility: Cannot model without Wu’s hold rate
• Both Players Error-Prone: W/UFE ratios (0.76, 0.97) suggest high volatility
• Best of 5 Format: Increases variance vs Best of 3
• Nardi’s Weak Serve: 56.9% 1st serve in % creates break opportunities
• Wu’s Unknown Tour-Level Performance: Complete uncertainty

Data Limitations

Critical Gaps:

1. Wu Y. has ZERO tour-level matches in last 52 weeks
2. Wu’s hold % and break % are unavailable (fundamental statistics)
3. Wu’s tiebreak performance at tour level unknown
4. Best of 5 historical performance missing for both players
5. Market odds unavailable for edge calculation
6. No tour-level H2H history

Impact on Analysis:

• Cannot construct game distribution model
• Cannot calculate expected total games
• Cannot calculate expected game margin
• Cannot perform model-empirical validation
• Cannot calculate edge vs market
• Cannot meet 2.5% edge threshold

Data Quality Assessment

Completeness: MEDIUM

• Player 1 (Nardi): ✓ Complete
• Player 2 (Wu): ✗ Critical gaps
• Market odds: ✗ Missing
• Best of 5 context: ✗ Missing

Reliability: LOW

• Wu’s clutch stats from mixed-level competition
• Wu’s recent form likely from challengers
• Elo rating based on limited tour exposure
• No validation possible without tour-level statistics

Confidence Multiplier: 0.6 (40% reduction for MEDIUM quality)

• Even if edge could be calculated, would need >4% raw edge to reach 2.5% adjusted edge
• Without market odds, cannot calculate any edge
• PASS is required regardless of confidence adjustment

Verification Checklist

Core Statistics

• [✓] Hold % collected for Player 1 (Nardi: 73.9%)
• [✗] Hold % collected for Player 2 (Wu: MISSING - 0 tour matches)
• [✓] Break % collected for Player 1 (Nardi: 19.1%)
• [✗] Break % collected for Player 2 (Wu: MISSING - 0 tour matches)
• [✗] Tiebreak statistics adequate (Wu: 0 tour TBs)
• [✗] Game distribution modeled (CANNOT - missing Wu data)
• [✗] Expected total games calculated (CANNOT - missing Wu data)
• [✗] Expected game margin calculated (CANNOT - missing Wu data)
• [✗] Totals line compared to market (NO ODDS AVAILABLE)
• [✗] Spread line compared to market (NO ODDS AVAILABLE)
• [✗] Edge ≥ 2.5% for any recommendations (CANNOT CALCULATE EDGE)
• [N/A] Confidence intervals appropriately wide (CANNOT CONSTRUCT MODEL)
• [✓] NO moneyline analysis included

Enhanced Analysis

• [✓] Elo ratings extracted (Nardi: 1593 hard, Wu: 1353 hard)
• [✓] Recent form data included (Nardi: 5-4, Wu: 9-1)
• [✓] Clutch stats analyzed (available for both, but Wu from mixed competition)
• [✓] Key games metrics reviewed (available for both)
• [✓] Playing style assessed (both error-prone, W/UFE < 1.0)
• [✓] Matchup Quality Assessment completed (noted data limitations)
• [✓] Data Quality Issues section completed
• [✓] Pass Rationale clearly documented

Decision Criteria

• [✗] Minimum data requirements met - FAILED (Wu has 0 tour matches)
• [✗] Market odds available - FAILED (not found for this match)
• [✗] Edge calculation possible - FAILED (cannot model + no odds)
• [✓] Pass recommended when criteria not met - COMPLETED

Final Verification: Report correctly identifies data quality issues and recommends PASS. No false precision or unsupported recommendations made.

Sources

1. TennisAbstract.com - Player statistics (Last 52 Weeks Tour-Level Splits)
   • Nardi L.: Complete tour-level data (16 matches)
   • Wu Y.: NO TOUR-LEVEL DATA (0 matches last 52 weeks)
   • Note: Wu’s clutch/form data from different time period or competition level
2. Briefing File Data - Provided via nardi_l_vs_wu_y_briefing.json
   • Collection timestamp: 2026-01-19T14:20:19.868633Z
   • Data quality: MEDIUM
   • Critical gap: Wu Y. tour-level statistics missing
3. Sportsbet.io - Match odds
   • Status: NOT FOUND
   • Error: “Match not found for Nardi L. vs Wu Y. in date range [‘2026-01-19’, ‘2026-01-20’, ‘2026-01-18’]”
   • Impact: Cannot calculate edge

Conclusion

This match analysis demonstrates the importance of data quality standards in professional tennis betting. While Nardi L. is a clear Elo favorite (+240 points), the absence of tour-level hold % and break % statistics for Wu Y. makes it impossible to construct a reliable game distribution model.

Key Takeaway: Totals and handicap modeling requires hold/break rates as fundamental inputs. These cannot be substituted with break point statistics or challenger-level data. Without these inputs, any expected total or spread would have a confidence interval of ±5-6 games, making it impossible to identify edges ≥2.5%.

Professional Recommendation: PASS on both totals and spread markets until:

1. Wu Y. establishes a tour-level statistical profile (10-15 matches minimum)
2. Market odds become available for edge calculation
3. Both players have Best of 5 historical data for format-specific modeling

A PASS recommendation is a successful outcome when data quality cannot support actionable betting recommendations. Protecting capital by avoiding low-quality setups is as important as identifying +EV opportunities.