TennisEdge's AI model finds no betting value in A. Eala vs Q. Zheng (Washington on 2026-07-27) — the market is efficient, with Q. Zheng favoured at 1.88.
═══ A. Eala (A) vs Q. Zheng (B) — WTA Washington, hard, R32 ═══
MODELS vs MARKET
ELO: Eala 36.1% / Zheng 63.9%
ML: Eala 46.4% / Zheng 53.6%
Blended: Eala 44.2% / Zheng 55.8%
Exchange: Eala 43.1% (2.32) / Zheng 58.5% (1.71) [book 1.92/1.88 = stale/soft]
Fresh sim (baseline): Eala 30.5% / Zheng 69.5% <-- HIGH OUTLIER
Consensus (ELO/ML/exchange) clusters Zheng 54-64%, exchange 58.5%.
RECENT FORM (last 5)
Eala: L Paolini(1817) 6-4 4-6 6-3 (narrow, grass) | W Swiatek(1965 Elite!) 7-6 6-2
(GRIND, huge over-performance) | W Joint(1616) 3-6 6-2 6-0 | W Zarazua(1564)
6-1 6-2 DOM | pre: won Birmingham title, Berlin SF. SHARPENING — but ALL grass.
Hard/clay record poor (blown out by Jovic 6-2 6-2 twice).
Zheng: L Krejcikova(1832) 6-4 6-4 | W Micic(1500) | W Bouzas(1500) 7-5 6-1 |
L Siniakova 6-4 6-4 | earlier: ugly losses to Cristian(1409), Gibson(1537),
Tauson. FADING/mediocre — recent level well below her rank-32 / ELO status.
PATH & SHAPE
Eala: last match 2026-07-06 (21 days ago) — RUST risk. grass -> hard TRANSITION
(depresses level, WIDENS margins). No hard-court reps in window.
Zheng: Athens hard 07-13/15/17 (10 days ago) — modest hard reps but on a weak
qualifying field; lost her last match. First match here for both.
FATIGUE (Step 2)
Both effectively rested. Eala matches_14d=0. Zheng last played 10d ago (212 min
over 3 matches, all >7d stale) — CONTEXT only, no extreme spot. Fatigue irrelevant.
ELO/RANK SANITY
Eala rank 42, Zheng rank 32 — close. Zheng's ELO edge (64%) looks generous vs her
actual recent results (losing to 1400-1500 level). Suspect ELO/serve stats are
stale/career-inflated; discount, do not lean on.
SIM INPUT VALIDATION (resimulate)
Baseline serve inputs: Zheng spw 0.611 / hold 0.75 vs Eala spw 0.569 / hold 0.592.
Zheng's 75% hold is inconsistent with recent losses to sub-1500 players and was
built partly on a weak Athens quali field; Eala's numbers were depressed facing
elite grass returners (Swiatek/Rybakina/Paolini). Re-ran with spw_b 0.611->0.561
(clamped -0.05, justified by field-quality bias): win prob FLIPS to Eala 56.0%.
A single clamped input moves the estimate 25 pts (30.5% -> 56.0%) — the win prob
is knife-edge sensitive and straddles the exchange's 43%. The baseline Zheng 69.5%
is an unstable, opponent-quality-biased artifact, NOT a robust edge.
MARKET-BY-MARKET (executable prices, 5% commission-adj)
ML Zheng @1.71: baseline sim "edge" +9.8% is pure divergence artifact; corrected
input converges to exchange (~58%) -> edge ~ -1.7%. NO edge.
ML Eala @2.32: baseline hugely negative; even shaded scenario ~56% gives edge only
+1.6% vs eff 2.254 -> below 5% bar, and rests on rust+transition upside. NO bet.
AH Eala +2.5 @1.88: covers 0.43 (baseline) / 0.45 (shaded) -> edge -11% to -9%. NO.
AH Zheng -2.5 @2.04: covers 0.57 -> +6.7%, but HIGH-VARIANCE lay (needs blowout),
bar is 10% -> FAIL; also inflated by same divergent sim. NO.
Totals: over 21.5 @2.0 (0.457) neg; under 21.5 @1.92 (0.543) edge +1% -> below bar.
GATES
G1 edge-origin: FAIL — the only >5% "edge" (Zheng ML +9.8%) exists solely because
the sim diverges high from the ELO/ML/exchange consensus; a justified -0.05 input
correction erases and reverses it. No robust model-vs-market disagreement survives.
G2 consensus: FAIL — market AND ML both sit against a strong-Zheng thesis; her live
form (losses to 1400-1500, lost last match) is independent evidence AGAINST it.
G3 fatal-risk: FAIL — Eala carries a 21-day layoff + grass->hard transition (both
depress level and widen margins); Zheng's edge rests on stale/inflated serve stats.
Both sides carry a thesis-negating risk.
VERDICT: PASS. Converged, efficient market (exchange Zheng ~58.5% sits in the middle
of the plausible range). Every apparent edge is either below the bar or a manufactured
artifact of an unstable, field-quality-biased simulation input. No bet.