TennisEdge's AI model finds no betting value in A. De Minaur vs S. Tsitsipas (Washington on 2026-07-27) — the market is efficient, with A. De Minaur favoured at 1.41.
═══ A. De Minaur (A) vs S. Tsitsipas (B) — Washington, HARD, 1/16-finals ═══
MODELS vs MARKET
ELO : De Minaur 71.9% ML : De Minaur 47.3% Sim: De Minaur 58.0% / margin +1.1g
Book : A 1.41 (implied ~71%) — STALE/soft, contradicted by exchange
Betfair (executable): A 1.86 (~53.8%) | B 2.14 (~46.7%)
The two math models STRADDLE the exchange (ELO high, ML low); the point-by-point sim
(58%) converges with the exchange (54%) — the market is pricing efficiently between them.
RECENT FORM (last 5)
De Minaur (rust risk — LAST MATCH 2026-07-06 = 21d ago, all grass, first match at venue):
06-07 Wimb 1/8 L Cobolli (1724) 5-7 6.4-7.7 3-6 — narrow 3-set loss to Strong
04-07 Wimb 1/16 W Svajda (1799) 6-2 5-7 6-2 6-4 — win over Strong, mixed
02-07 Wimb 1/32 W Mannarino (1790) 6-3 6-2 6-2 — DOM over Strong
30-06 Wimb 1/64 W Burruchaga(1515) 7-6 6-1 6-0 — GRIND-start, rout over Weak
19-06 Lon QF L Nakashima (1779) 5-7 3-6 — straight-set loss to Strong
Shape: STABLE, rust-dominant. Grass form fine but 21d layoff + grass->hard transition reset it.
Tsitsipas (TITLE Gstaad clay 8d ago; path = 5 matches at Gstaad -> MATCH-SHARP, but on clay):
19-07 Gstaad F W Collignon(1865) 6-4 6-7 6-3 — GRIND title win over Strong
18-07 Gstaad SF W Shevchenko(1704) 4-6 6-3 3-6 — GRIND, 3 sets over Strong
17-07 Gstaad QF W Rinderknech(1734)3-6 6-3 3-6 — GRIND, 3 sets over Strong
16-07 Gstaad 1/8 W Kym (1642) 4-6 7-6 6-7 — GRIND (TBs) over Avg
14-07 Gstaad 1/16 W Buse (1823) 6-4 6-4 — DOM over Strong
Shape: sharp on CLAY (lots of grinding 3-setters) but clay->hard transition; sharpness does not carry.
FATIGUE (Step 2 calibration)
Tsitsipas fatigue block is heavy (810min/14d, b2b 3, 15 sets) but that load ran THROUGH 07-19 —
he has had 8 days rest since the title, so NO current extreme spot; cumulative load 8d old = noise.
De Minaur fully rested but RUSTY. Both age 27 -> baseline recovery. No fatigue edge either way.
NARRATIVE
Let-down: Tsitsipas won Final 8d ago — just past the 7d window -> None (borderline note only).
Surface transition: BOTH players transitioning (A grass->hard, B clay->hard) -> widens margin
uncertainty on both sides, argues AGAINST any tight cover.
ELO/ranking: A rank 6 vs ELO 72% -> agree, credible (but ELO ignores the 21d layoff/transition,
so it overstates this spot). B rank 75 but ex-top-4; sim(42%)/exchange(47%)/ML(53% for B) all
rate him well above raw rank — genuine uncertainty, not a clean edge.
EDGES (commission-adj vs EXECUTABLE Betfair price; sim is the single distribution driving all)
ML A @1.86 : +2.9% (min 1.81) below 5% bar
ML B @2.14 : -6.0% no edge
AH A -0.5 @1.97 : +3.6% (cover~0.556) below 5% bar (only ±0.5 offered; model -1.5 unpriced)
Tot UNDER 23.0 @1.96: +0.4% no edge
Tot OVER 23.0 @2.00: -4.0% no edge
No market clears the 5% bar. Sim inputs not shaded: surface-transition arguments cut both ways
(A's hard-court class UP but rust DOWN; B's clay-inflation DOWN but match-sharpness UP) -> net
offsetting, no justified numeric correction, trust baseline.
GATES
G1 edge-origin: FAIL — best pre-narrative model-vs-market edge is +3.6% (AH -0.5), below the 5% bar.
G2 consensus: FAIL — ML sits against De Minaur (47%); a live-form signal (B match-sharp title run)
also cuts against; no on-surface H2H or this-event evidence to override.
G3 fatal-risk: FAIL — De Minaur 21d layoff + grass->hard first match is a real level-depressor that
would negate any thin A edge; units cannot buy it back.
VERDICT: PASS. Models straddle the exchange, the sim converges with the executable price, and no
market clears the 5% edge bar. Both players carry offsetting transition/rust-vs-sharpness concerns.
Textbook efficient-market pass.