TennisEdge's AI model finds no betting value in E. Shibahara vs E. Dong (Vancouver on 2026-07-25) — the market is efficient, with E. Shibahara favoured at 1.40.
E. Shibahara vs E. Dong - Vancouver (hard)
MODELS vs MARKET
ELO: 83.2% A | ML: 82.7% A | Combined model: 77.9% A
Book: 1.40 / 2.75 (implied 66.3% A) | Betfair ML: 1.02 / 1.02 - placeholder, no real exchange market
Raw model-market gap: ~11.6% on A, but book-priced only (no executable exchange price).
RECENT FORM
Shibahara (A, age 28, rank 190): NO matches in 90d, no fallback data. Zero live form - ELO is stale.
Dong (B, unranked): 4 matches last 14d, all W35/ITF hard.
07-24 L vs L. Cabrera (1552) QF - lopsided scoreline
07-23 W vs L. Poling (1168)
07-22 W vs Mushika Mio (1331) - GRIND (tiebreak, 3 sets)
07-14 L vs A. Arseneault (1420)
Dong ceiling is low: wins only vs Weak (<1500) opponents, losses to 1420/1552.
FATIGUE
A: 0 matches 14d, fully rested (but rest here = inactivity, not freshness).
B: 4 matches 14d, 12 sets, 287 court-min, back-to-back x2, played yesterday. Heavily loaded.
Age: Dong age unknown - age-neutral. Fatigue edge to Shibahara, but this may only confirm, never create, an edge.
NARRATIVE
No finals/letdown signals. No venue history for either. Dong arrives from Saskatoon (same surface - no transition).
ELO/ranking sanity: Shibahara ELO-favourite at 83% yet ranked 190 with ZERO singles matches in the window - the rating is stale/uncorroborated. Framework: discount the ELO edge unless live form independently backs it; there is no live form at all.
GATES
G1 edge-origin: PASS - model-vs-market gap (~78% vs 66%) exists numerically before fatigue/narrative.
G2 consensus: PASS - market, ELO and ML all favour A; no consensus conflict.
G3 fatal-risk: FAIL - the entire edge rests on a stale ELO for a player with zero matches in 90d; if the rating is inflated/rusty, the thesis is gone. That risk directly negates the bet.
PRICING
No executable exchange price (betfair_ml 1.02/1.02 is a dead market). Book-priced ML needs edge >= 8%. Discounting the stale 78-83% model prob for a 90-day layoff, honest win prob for A ~0.72-0.75 -> edge vs 1.40 book ~1-4%. Below the 8% book-priced bar. Small event without exchange liquidity is exactly the losing segment.
VERDICT: PASS. Shibahara should win on class (predicted winner), but the staked edge depends entirely on an unverifiable stale rating, there is no exchange market, and the discounted book edge is thin. No alt markets priced.