TennisEdge's AI model finds no betting value in S. Costoulas vs N. Hibino (Livesport Prague Open on 2026-07-18) — the market is efficient, with N. Hibino favoured at 1.84.
═══ S. Costoulas vs N. Hibino — Livesport Prague Open (clay, Semi-finals) ═══
MODELS
ELO: 56.2% A | ML: 37.4% A (62.6% B) | Blend: 56.2% A
Market: 1.85 / 1.84 (implied ~49.9% A) — pure pick'em
Value flag: A +6.3% (ELO-driven) — but ML flags B at roughly the same size. Models cancel.
RECENT FORM (last 5 shown; all fetched matches used)
Costoulas: 1-4 in last 5
07-14 Athens Q 1/16 L vs Valentova (1500) 1-6 2-6 — blown out
06-23 Wimbledon QF L vs Hunter (1606) 5-7 4-6
06-03 Foggia 1/8 L vs Pieri (1368, Weak) — loss to weak opp
06-01 Foggia 1/16 W vs Sakellaridi (1348) in 3 — scrappy vs Weak
05-28 Zaragoza 1/8 L vs Podoroska (1607) GRIND loss (TB)
Older DOM wins exist (FO junior run, Zaragoza) but all vs Weak/Avg; only Strong-opp win was a GRIND (Jeanjean TB, Apr).
Hibino: 3-2 in last 5, but heavy volume
07-14 Athens Q L vs Tauson 5-7 4-6
07-12 Athens Q Final W vs Tona 6-4 6-0 DOM (vs default-1500)
07-09 Corroios 1/8 L vs Gao 0-0 (ret? data oddity)
07-07 Corroios 1/16 W vs Havlickova (1622) 6-0 6-4 DOM vs Avg
07-03 Elvas QF L vs Lee 4-6 2-6
Best signal: DOM Final win vs Valdmannova (1773, Strong) Birmingham 05-31 — grass, 6 weeks old.
FATIGUE
Costoulas: 1 match / 57 min in 14d — fresh
Hibino: 4 matches / 11 sets / 279 min in 14d — heavy
Court-time diff 222 min > 60 → clear fatigue disadvantage Hibino.
NARRATIVE
Hibino: Won Athens qual Final 07-12 (~6d ago) → WEAK letdown, partially confirmed by 07-14 loss.
Costoulas: no Final letdown; lost FO Final 05-21 (stale). No venue history for either at Prague.
Data caveat: both players show 07-14 Athens losses yet appear in a Prague SF — schedule/feed inconsistency.
VERDICT — PASS
ELO edge on Costoulas (+6.3%) is killed by live form: 1-4 stretch, blown out by a 1500 four days ago, loss to a 1368. ML edge on Hibino is dented by 279 court-minutes in 14 days and a qual-final letdown spot. Models point opposite ways, market splits them at a pick'em, and the qualitative read validates neither side. No alt markets priced. Efficient coin flip — no bet.