WIMBLEDON 03/07/26 - matches R32 - 9.5/10 score !

Today there were 8 matches at WIMBLEDON 3rd round (R32)

🎯 Correct Elo Predictions (Top 4 Matches) All platforms agreed

The player with the higher pre-match Surface Elo won every match in the top half:

Jannik Sinner (2754) defeated Jenson Brooksby (2601)

Félix Auger-Aliassime (2622) defeated Michael Zheng (2550)

Novak Djokovic (2883) defeated Arthur Rinderknech (2555)

Alejandro Davidovich Fokina (2634) defeated Marton Fucsovics (2620)

❌ Elo Upsets (Bottom 4 Matches)

TenELOs vs. Other Platforms

TenELOs (Grass Elo) correctly predicted 7 / 8...Only missed the Jan-Lennard Struff upset over Daniil Medvedev.

Major Sportsbooks (Odds) and Official ATP Rankings / Seedings predicted 4 / 8...Suffered massive losses on Paul, Fonseca, Medvedev, and Jódar. Failed on underdogs with superior grass skill sets.

Why TenELOs Outperformed the Competition

The reason TenELOs wiped the floor with sportsbooks and official seedings comes down to its mathematical isolation of surface-specific data. Traditional platforms suffered heavily because they grouped overall momentum or clay/hardcourt achievements into their projections.

The Jódar vs. Mochizuki Matchup:

Sportsbooks heavily favored Jódar based on his superior overall rank and 23rd tournament seeding. Both players entered match in top peak grass form (2550 ELO). TenELOs ignored the noise, focusing purely on Jodar raw unexperienced pre tournament 2400 Grass Elo compared to Mochizuki's 2498, easily identifying the correct favorite.

The Safiullin vs. Fonseca Matchup: Fonseca was the heavily backed favorite in the betting market due to a recent hot streak. TenELOs correctly recognized Safiullin’s proven grass-court efficacy (2594 Elo vs. Fonseca's 2554), resulting in another predictive victory.

Handling True Even Matches: In the Hurkacz vs. Paul matchup, sportsbooks priced Paul as a heavy favorite. TenELOs accurately modeled them as a dead-even 2669 tie, completely shifting expectations to a toss-up where Hurkacz eventually prevailed.

By evaluating matchups purely on a player's historical grass metrics, TenELOs proved that surface specialization is the ultimate predictor in grand slam tennis—making standard betting lines look highly flawed.

📈 Revised Overall Elo Accuracy Factoring in those exact baseline numbers, the Surface Elo rating was an absolute powerhouse for the day 1 of Round of 32:

6 out of 7 matches with a definitive raw rating gap were predicted correctly (Sinner, Auger-Aliassime, Djokovic, Davidovich Fokina, Safiullin, and Mochizuki).

Jan-Lennard Struff over Daniil Medvedev remains the single, true Elo upset across the entire board.

On a scale of 0 to 10, TenELOs' performance for this specific set of Round of 32 matches rates as a 9.5 out of 10. Here is why it scores near-perfection, alongside the minor detail that keeps it from a flawless 10:

Why it gets a 9.5 Unmasked Hidden Value (10/10 execution): Its biggest triumph was the Mochizuki vs. Jódar match. Standard seedings, general ATP rankings, and major sportsbooks heavily favored Jódar. TenELOs' surface data uncovered that Mochizuki possessed a massive 98-point grass advantage (2498 vs. 2400) prior to the tournament. The model identified the correct favorite when everyone else saw a major upset.

Beating the Market on Undervalued Players: By isolating grass metrics, it correctly leaned into Roman Safiullin (2594 Elo) over João Fonseca (2554 Elo), completely going against betting sites that heavily backed Fonseca based on his overall tour hype.

Accurate Match Contextualization: It correctly flagged Hurkacz and Paul as a dead-even toss-up (2669 Elo) rather than agreeing with bookmakers who labeled Paul a definitive favorite.

Why it missed a perfect 10

The only deduction comes from the Jan-Lennard Struff vs. Daniil Medvedev match. Medvedev held a commanding 76-point Elo advantage over Struff (2678 vs. 2602). TenELOs mathematically favored Medvedev to win comfortably, meaning it was completely blindsided by Struff's straight-sets victory.

Because predicting sports with 100% accuracy is mathematically impossible due to human variance, an 87.5% success rate (7 out of 8) on a grand slam grass court is about as elite as a predictive model can get.