TennisEdge's AI model finds no betting value in T. Svajda vs J. Mensik (Washington on 2026-07-28) — the market is efficient, with J. Mensik favoured at 1.18.
T. Svajda (A, #387, age 20) vs J. Mensik (B, #18, age 20) — Washington, hard.
MODELS / MARKET (Svajda win prob)
ELO 18.5% | ML 28.1% | sim-baseline 37.5% | Betfair 23.8% (4.2) | book 20% (5.0)
Mensik: Betfair 1.27 (78.7%), book 1.18. value_pick B, edge +0.6% (nil).
ELO (81.5% Mensik) and market (78.7%) agree tightly; ML a touch softer (72%).
SIM DIAGNOSIS (why the raw sim is untrustworthy here)
Baseline sim gives Svajda 37.5% — 10-19 pts above ELO/ML/market. Cause = opponent-
unadjusted serve/return inputs, both biased toward Svajda:
- Svajda spw 0.598 built over 50 matches vs M25/Challenger fields (opp ELO 1261-1712,
weak returners) -> INFLATED.
- Mensik spw 0.624 built vs elite ATP/Slam fields (De Minaur, Rublev, Zverev, Dimitrov,
Mannarino; elo 1771-2050 on clay/grass) -> DEFLATED, and off-surface.
Re-ran bounded (spw_a 0.598->0.58, spw_b 0.624->0.64) to reproduce the ELO/ML/market
consensus win prob: Svajda -> 21.6% (matches market 24% / ELO 18.5% / ML 28% band).
At that coherent anchor: Svajda +3.5 covers ~36% (NOT 55.8% baseline / 50% brief),
+2.5 ~31%, +4.5 ~51%; Mensik -3.5 covers ~64%; expected margin Mensik ~3.3 games.
The baseline "coin-flip cover" was a pure input-bias artifact.
RECENT FORM (last 5 shown; all fetched used for classification)
Svajda: 07-26 Washington FINAL Won vs McDonald 1712 (6-4 6-4, DOM over Strong);
07-25 SF Won vs Ymer 1605 (6-0 7-6, GRIND); 07-15 Lincoln Lost vs Lajal 1658 (3set);
07-12 M25 Dallas title Won vs Kim; 07-11 Won vs Zhu 1374 (6-4 6-4). SHARPENING, red-hot,
match-sharp on hard (won a title yesterday). But titles/stats built on M25 fields.
Mensik: 07-02 Wimbledon R64 Lost vs Dimitrov 1780 (4set); 06-30 R128 Won vs Samuel (5set);
Boodles exos; 06-16 Queen's Lost vs Mannarino; French Open SEMI (bt De Minaur, Rublev,
Fonseca). Genuine top-20 class. STABLE. Last competitive match 07-02.
PATH / NARRATIVE
Svajda: match-sharp, but LETDOWN spot — won a Final 1 day ago (07-26). Slightly depresses
the title-winner; pushes marginally toward Mensik.
Mensik: RUST (25 days idle, last match 07-02 >21d) + surface TRANSITION (grass->hard) +
first match at venue. Supports the dog's competitiveness — but transition WIDENS margins
(blowout tail cuts both ways) so it does NOT cleanly support any tight cover.
Fatigue: Svajda 2 matches in 2 days but straight-setters (~90-100min each), no 150min+/3-set
yesterday -> not an extreme spot; age 20 (<=22) recovers fast -> softened to context only.
Market over-prices visible fatigue; Svajda is the dog anyway so no edge either way.
GATES (evaluated on strongest candidate — Svajda +3.5 @ 2.16 / Svajda ML @ 4.2)
G1 edge-origin: FAIL — the only numerical edge came from the biased baseline sim; once
inputs are corrected to the ELO/ML/market consensus, Svajda +3.5 cover ~36% vs implied
47.6% = NEGATIVE edge. Svajda ML edge <=3% at ML's own 28% (<5% bar). No pre-narrative edge.
G2 consensus: FAIL — ELO+market+corrected sim all agree Mensik ~78%; only ML (72%) splits,
and Mensik went to a French Open SEMI = live-class evidence AGAINST a Svajda upset.
G3 fatal-risk: FAIL for Mensik -3.5 (~64% cover, +6.2% raw but HIGH-VARIANCE line needs
>=10%; laying a big number with a RUSTY favourite in transition vs a red-hot grinder is a
negating risk) and for Svajda ML (387 vs 18 class gap, Slam-semi form).
MARKET SCAN (all at Betfair/commission-adjusted)
Mensik ML 1.27: implied 78.7% ~ estimate 78%; min_odds 1.34 > 1.27 -> auto-PASS + rust risk.
Svajda ML 4.2: implied 24.8%; estimate 22-28%; edge 0 to +3% (<5%). PASS/at best lean.
Svajda +3.5 @ 2.16: cover ~36% vs 47.6% -> negative. PASS.
Mensik -3.5 @ 1.77: cover ~64% vs 57.8% = +6.2% but wide-AH needs >=10% + fatal risk. PASS.
Total 22.0: expected ~21-22 games, input-sensitive, no robust 5% edge. PASS.
VERDICT: PASS. Winner consensus is efficient (ELO 81.5% / market 78.7% / corrected sim ~78%);
the tempting +3.5 cover was a weak-field input artifact that vanishes on calibration. Svajda's
hot/match-sharp read vs Mensik's rust/transition is real but already inside the price and cannot
push Svajda's win prob credibly past ~28%, where no market clears its min-odds/edge bar. No bet.