TennisEdge's AI model finds no betting value in A. Vukic vs L. Musetti (Washington on 2026-07-29) — the market is efficient, with L. Musetti favoured at 1.37.
═══════════════════════════════════════════════
A. Vukic (30, #99) vs L. Musetti (24, #7)
ATP Washington — hard — 1/8-finals
═══════════════════════════════════════════════
MODELS vs MARKET (Vukic = player A)
ELO 27.4%
ML 28.7%
Blend 34.0%
Sim (raw) 49.1% <-- outlier, see input audit
Book 3.10 / 1.37 (context only)
Betfair 3.00 / 1.42 (executable)
Commission-adj implied: Vukic 34.5% (eff 2.90) | Musetti 71.5% (eff 1.399)
SIM INPUT AUDIT (validate-by-convergence)
Sim gives Vukic 49.1% — diverges from ELO (27.4), ML (28.7) AND market (33-34).
Three of four estimates cluster 27-34; the sim is the lone outlier, so the inputs
are suspect, not the consensus. Inputs: SPW 0.613/0.613, RPW 0.421/0.422,
hold 86.5% vs 82.1% — i.e. the sim thinks these two serve identically and that
Vukic HOLDS BETTER than a world #7. That is an opponent-quality artifact:
- Vukic's 50-match window is Challenger-heavy (Granby: opponents 1372-1813) plus
Washington qualifying (Bolt 1641, Dougaz 1606). Serve/return fattened on weak fields.
- Musetti's window is clay + elite (Ruud 1974, Cerundolo 1907, Mpetshi Perricard),
plus a 76-day gap, so his hold% is depressed by field strength, not by ability.
No re-run staked: any justified correction shades Vukic's SPW/RPW DOWN, i.e. moves
him further from a bet, not toward one. Baseline rejected as the anchor; ELO/ML/market
consensus used instead. Handicap/totals are unpriced (alt_markets.handicap = null,
totals = null), so no alt market is stakeable regardless.
RECENT FORM — Vukic (5 most recent)
07-27 W Svajda 1755 6-3 6-1 DOM (Washington R1, main draw)
07-26 W Bolt 1641 6-0 6-1 DOM (Washington QUALI final)
07-25 W Dougaz 1606 6-3 6-3 DOM (Washington QUALI semi)
07-19 W Gea 1813 6-4 1-6 6-1 GRIND (GRANBY TITLE)
07-19 W Holmgren 1657 6-4 3-6 6-3 GRIND
Streak: 8 straight wins on hard, all since 07-14. Last loss 07-08 (grass).
RECENT FORM — Musetti (all 4 in window)
07-27 W Arnaldi 1702 6-0 3-1 ret. 45 min, opponent retired
05-12 L Ruud 1974 3-6 1-6 clay
05-10 W Cerundolo 1907 7-6 6-4 clay, DOM-adjacent vs Elite
05-08 W Mpetshi P. 1682 6-4 6-4 clay
*** 76-day gap between 05-12 and 07-27. Missed French Open, entire grass swing,
Wimbledon. Returning player. ***
SHAPE / PATH / SHARPNESS (Step 1b)
Vukic: SHARPENING — margins widening vs bracket (6-3 6-1 over a 1755 is DOM over
Strong; 6-0 6-1 in the quali final). Path: 3 matches at THIS venue this
week (2 quali + R1) => maximally match-sharp, conditions-adjusted.
Musetti: UNREADABLE — one 45-minute walkover-ish win is not evidence of anything.
Rust risk is severe by the Step 1b rule (last competitive match >21 days:
76 days). Pre-layoff shape was good but it was CLAY (Cerundolo 1907 in 2).
Surface transition: Musetti clay -> hard, first real hard match in months. Real, but
his most recent match is now hard, so partially discounted.
FATIGUE (Step 2)
Vukic: 8 matches/14d, 830 court min, back_to_back = 3, played 07-27.
Today is the 4th match in 4 consecutive days => EXTREME SPOT (3rd-in-3 rule
triggered and exceeded). Age 30 => recovers slowly => harden one tier.
Real disadvantage, not context.
No 3-set-hangover haircut: yesterday was a 2-set win (6-3 6-1), and the
calibrated -4.5pp applies only to 3-set wins. Not stacked with the above.
Musetti: 45 min in 14d. Physically fresh; the cost is rust, not fatigue.
Net: the two "conditioning" stories point in OPPOSITE directions and roughly cancel —
Vukic is sharp but cooked, Musetti is fresh but rusty. Neither is a tiebreaker.
NARRATIVE (Step 3)
Let-down: literal rule fires (Won "Final" 07-26, 2d ago = LETDOWN) but that is the
Washington QUALIFYING final, not a title — do not price it as a peak-and-crash.
The real title is Granby 9d ago => WEAK letdown at most.
TITLE: Granby 9d ago (hard) — genuine Challenger title, live hard-court form.
Venue momentum: Vukic 3-0 at Washington this week (quali + R1).
Musetti: no venue history in window; motivation unknown, no speculation.
ELO vs RANKING SANITY
Vukic: ELO consistent with #99. Caveat — the last 3 weeks of ELO gain is farmed at
Challenger/qualifying level (1372-1813 fields). Discount the streak.
Musetti: #7 with a thin, layoff-shortened ELO sample => ELO likely UNDERSTATES him.
Per the rule, that means do NOT back Vukic on the ELO gap either. The
divergence cuts against the underdog, not for him.
PRICING
Vukic @ 3.00 BF (eff 2.90 = 34.5%). My prob 32% — below ELO/ML would allow much
upside, above pure ELO for the sharpness/path. min_odds = 1.05/0.32 = 3.28.
Executable 3.00 < 3.28 => AUTOMATIC PASS.
Musetti@ 1.42 BF (eff 1.399 = 71.5%). My prob 68% (ELO 72.6% haircut for 76-day
rust). min_odds = 1.05/0.68 = 1.54. Executable 1.42 < 1.54 => AUTOMATIC PASS.
Even at un-haircut ELO 72.6%, min_odds = 1.45 > 1.42. Fails either way.
Book 3.10 on Vukic is the "soft" quote; Betfair contradicts it at 3.00. Per pricing
discipline that KILLS the bet, it does not confirm it. The brief's value_edge of
+3.4% is computed against the book snapshot and is not executable.
GATES
G1 edge-origin: FAIL — stripped of all path/fatigue/narrative, blend 34.0% vs
executable 34.5% = -0.5% edge; ELO/ML are 6-7pp WORSE. No numeric
model-vs-market gap exists on either side before adjustment.
G2 consensus: FAIL — ELO, ML and market all agree Musetti; the only dissent is a
sim whose inputs are demonstrably weak-field-inflated. That is a lone
discredited split, not hard independent evidence.
G3 fatal-risk: FAIL — the Vukic thesis is "match-sharp vs rusty," but he arrives on
a 4th-match-in-4-days at 30 years old, and every win in the streak is
vs 1372-1813 opposition. Class gap vs a top-7 is real if Musetti's
timing shows up at all — that directly negates the thesis.
VERDICT: PASS
A genuinely interesting shape — the sharpest player in the draw against a top-7
returning from 2.5 months out — and the market has already priced it: a #7 is only
1.42 against a #99, which is exactly the rust discount. The qualitative story that
would make this a bet is the same story the price already reflects. Both sides fail
the min-odds test, both fail on the executable exchange price, and all three hard
gates fail. No stake.