TennisEdge's AI model finds no betting value in R. A. Burruchaga vs A. Blockx (Estoril on 2026-07-24) — the market is efficient, with A. Blockx favoured at 1.62.
═══ R. A. Burruchaga vs A. Blockx — Estoril QF (clay) ═══
MODELS A (Burruchaga) B (Blockx)
ELO 45.6% 54.4%
ML 52.5% 47.5%
Sim 40.8% 59.2%
Book 2.30 1.62
Betfair 2.28 1.76 (fair B ~56.4%)
RECENT FORM
Burruchaga (rank 56, age 24): 4-2 last 14d
07-22 Estoril R16 W vs Borges (1779, Strong) 3 sets GRIND
07-21 Estoril R32 W vs Wawrinka (1552, Avg) 3 sets GRIND
07-17 Umag SF L vs Merida Aguilar (1772) straight sets
07-16 Umag QF W vs Ugo Carabelli (1822) 3 sets GRIND
07-15 Umag R16 W vs Cobolli (1915, Elite) straight sets DOM
Blockx (rank 36, age 21): 1-1 last 14d
07-22 Estoril R16 W vs Jacquet (1619, Avg) 3 sets w/ TB GRIND
07-14 Umag R32 L vs Droguet (1780) 3 sets
06-30 Wimbledon L vs Zverev (1859) 4 sets, two TBs — pushed elite
May: Madrid SF run — beat Ruud (1987), F. Cerundolo (1914), FAA (1850)
FATIGUE
Burruchaga: 6 matches/14d, 4 in 7d, 18 sets, 746 court-min, 4 back-to-backs; QF = 7th match in 10 days, 3rd straight day
Blockx: 2 matches/14d, 6 sets, 316 court-min
Court-time gap 430 min -> Burruchaga heavily disadvantaged (age 24, baseline — no softening)
NARRATIVE
Let-down: none either side (no recent finals)
Venue: Burruchaga 2-0 Estoril this week; Blockx 1-0
Transition: none — both on clay
ELO/rank: Blockx rank 36 with only a modest ELO edge — fast-rising 21yo (Madrid SF, wins over Ruud/FAA/Cerundolo); ELO likely understates him. Do not lean on the small ELO gap for A.
VERDICT: PASS
Qualitative read favors Blockx (fresh vs a player on fumes), but the exchange already prices it: my B win prob ~0.59 vs effective 1.72 (1.76 less 5% comm) = 58.1% implied -> ~1% edge. Min odds 1.78 > 1.76 executable -> automatic pass on ML. Handicap B -1.5 @1.90 needs 57.6% cover; both players are grinding 3-setters this week, cover ~52% -> no. Under 22.5: sim says 59.5% but the sim's handicap block is internally inconsistent (A +2.5 cover 34.5% < A win prob 40.8%), so its margin outputs are not trusted; own under read ~50% vs 51.3% implied -> no. ML model's lean to A is killed by the 746-min workload. Market efficient — no stake.