TennisEdge's AI model finds no betting value in A. Gea vs M. Zheng (Los Cabos on 2026-07-30) — the market is efficient, with M. Zheng favoured at 1.53.
A. Gea (A) vs M. Zheng (B) - Los Cabos, hard, 1/8-finals
MODELS vs MARKET
ELO(A)=0.516 ML(A)=0.493 sim/blend(A)=0.449 | book impl(A)=0.38
Betfair ML: Gea 2.50 (eff 2.425, BE 0.412) / Zheng 1.64 (eff 1.608, BE 0.622)
Gea ML edge: ELO +10.3pp, ML +8.1pp, sim/blend +3.7pp (models SPLIT).
Zheng ML edge: negative on every model (market prices Zheng ~62% > any model).
Handicap: Gea +2.5 @1.89 sim-cover 0.559 -> +1.7pp (sub-bar). Zheng -2.5 @2.0 -> negative.
Totals: Under 22.5 @1.95 -> -1.4pp; Over @1.98 -> -3.0pp. No alt edge.
ELO/RANKING SANITY (decisive)
Gea is ELO-favourite (0.516) yet ranked 241 vs Zheng 182 -> classic high-ELO/modest-rank
= rating farmed at Challenger level (Granby final run beating mostly sub-1650 fields).
Discount the ELO edge. The trustworthy sim/blend gives only 3.7pp (below 5%).
RECENT FORM / SHAPE / PATH
Gea (21, #241): today Won 1/16 vs Kwon 1785 4-6 7-6 2-6 (GRIND 3-setter, good scalp but ugly).
Prior: LOST Granby Final to Vukic (07-19); Challenger-heavy hard run; also lost to
Walton 1450 (grass) recently. Shape STABLE-to-mixed; level Challenger-calibrated.
Zheng (22, #182): today Won 1/16 vs Landaluce 1776 6-4 6-3 (DOM over Strong, straights).
Also beat Walton 1767 (07-24), reached Bloomfield Hills SF. Shape SHARPENING; beating
comparable/better fields more cleanly. Path: match-sharp, 5 matches/10d.
Both last played today (07-29).
FATIGUE (Step 2)
Gea won a 3-set match today -> extra-set haircut zone, but age 21 (<=22) softens one tier;
net slight disadvantage, and it CANNOT create a Gea edge (G1). Zheng's win today was 2 sets
(~neutral). No extreme spot flips this. Fatigue points AGAINST Gea, i.e. toward the market.
NARRATIVE
No let-down (Gea's Granby final 07-19, 10d ago). No surface transition (both hard).
No title recency. Nothing rescues a Gea overlay.
GATES
G1 edge-origin: FAIL - the >5% edge exists only via the farmed/discounted ELO (and lone ML);
the credible sim/blend is 3.7pp (<5%), and the qualitative read contradicts, not confirms.
G2 consensus: stake-tier only - models SPLIT and pro-Gea signal contaminated (farmed ELO).
G3 fatal-risk: n/a (ML, non-blocking).
VERDICT: PASS. Backing Gea means fighting both the market (sharpest estimate, Zheng ~62%) AND
the shape/path read (Zheng SHARPENING + DOM today; Gea grindy 3-set win + slight fatigue) on a
Challenger-farmed ELO the ranking contradicts. Zheng offers no value either (overpriced vs every
model). No alt-market clears its bar. No bet.