TennisEdge's AI model finds no betting value in F. Diaz Acosta vs F. Cina (Kitzbuhel on 2026-07-20) — the market is efficient, with F. Diaz Acosta favoured at 1.61.
═══ F. Diaz Acosta vs F. Cina — Kitzbuhel 1/16 (clay) ═══
MODELS vs MARKET
Book: A 1.61 (59.0%) / B 2.32 (43.1%) Betfair ML: 1.65 / 2.50
ELO: A 65.5% (+6.4 vs market)
ML: A 55.1% (-4.0 vs market)
Models split — ELO sees value on A, ML sees A slightly overpriced. No consensus.
RECENT FORM (last 5 each; classification uses full 90d)
Diaz Acosta (A):
07-14 Bastad R16 L vs Midon (1646) 6-3 4-6 2-6 — loss to Avg
07-11 Braunschweig SF L vs Gaston (1641) 7-5 6(4)-7(7) 5-7 — GRIND loss
07-10 Braunschweig QF W vs McDonald (1637) 6-3 6-2 — DOM vs Avg
07-09 Braunschweig R16 W vs Gaubas (1650) straight sets — vs Strong
07-07 Braunschweig R32 W vs Molleker (1505) 3 sets — GRIND vs Avg
Two straight losses to ~1645-level opps; Milan title 07-05 (3 TBs vs Cecchinato 1693) = GRIND champion. 6 days full rest since 07-14.
Cina (B):
07-19 Kitzbuhel Q-Final W vs Giustino (1558) 6-4 6-2 — DOM vs Weak
07-18 Kitzbuhel Q-SF W vs Michalski (1528) 3 sets — vs Weak
07-14 Gstaad R32 L vs Halys (1710) 3rd-set TB — GRIND loss vs Strong
07-12 Gstaad Q-Final W vs Hemery (1607) 6-4 6-3 — DOM vs Avg
07-11 Gstaad Q-SF W vs Zahraj (1354) 7-5 6-3 — vs Weak
Pattern: beats Weak/Avg, loses every Strong test (Halys 1710, Kasnikowski 1651, De Jong 1727). Form shine is opponent-quality inflated.
FATIGUE
A: 6 matches/14d, 933 court-min (avg 155!), but 6 DAYS REST since 07-14.
B: 6 matches/14d, 4 in last 7d, 805 min, played back-to-back incl. YESTERDAY.
Court-time diff 128min against A historically, but freshness clearly favors A. Slight edge A.
NARRATIVE
B "Won Final" 1d ago = qualifying final → mechanical LETDOWN flag, but quali wins usually mean match-sharp, venue-acclimated (2-0 on these courts). Signal discounted both ways.
A: Milan TITLE 15d ago (clay) — but followed by 2 losses. Momentum stalled.
VERDICT: PASS
ELO's +6.4% edge on A is not validated: ML model disagrees (55.1% < market), and A's live form (consecutive losses to 1646/1641) undercuts the 65% ELO read. Betting B instead fails too: his run is built on Weak/Avg scalps, he loses every Strong test, and he played yesterday while A had 6 days rest. Market ~59-60% A looks about right.
Alt markets: A -2.5 @1.96 model edge (61.6% vs 51%) inherits the inflated 65.5% ELO prob — not trusted. B +2.5 @1.96 has a tight-match case (both grinders) but sim says 38.4% cover; conflicting anchors. Total 22.5: model under 53.9% vs 51.5% implied — thin, and the grind-heavy profiles argue over. No stakeable edge anywhere.