TennisEdge's AI model finds no betting value in T. Fritz vs K. Majchrzak (Washington on 2026-07-29) — the market is efficient, with T. Fritz favoured at 1.27.
T. FRITZ (A) vs K. MAJCHRZAK (B) — ATP Washington, hard
Round field empty; both played 1/16-finals (07-27 / 07-28) → this is the 1/8-finals.
═══ MODELS vs MARKET ═══
ELO A 72.8% (B 27.2%)
ML A 71.1% (B 28.9%)
Sim/model A 72.6% (B 27.4%)
Book A 1.27 / B 3.97 (implied A 75.8%)
BETFAIR A 1.29 / B 4.20 ← executable, all edges priced here
Commission-adjusted (x0.95 on the win part):
A eff 1.2755 → implied 78.40% | model A 72.58% → edge −5.8pp (no bet A)
B eff 4.0400 → implied 24.75% | model B 27.42% → edge +2.7pp
ELO B 27.20% → edge +2.5pp
ML B 28.89% → edge +4.1pp (best lone model)
min_odds for B at ML's 28.9% = 1.05/0.2889 = 3.63 eff = 3.77 raw Betfair.
4.20 clears min_odds, but the absolute edge (+4.1pp) sits UNDER the 5% bar.
Direction: all three models lean B (every model prob for A is below the market's
~76.5% fair). Real disagreement — just too small to stake.
═══ RECENT FORM (5 most recent each) ═══
T. FRITZ (28, rank 9)
07-27 W Bergs (1766, Strong) 6-3 6-4 DOM — Washington R1, hard
07-08 L Zverev (1913, Elite) 4-6 4-6 2-6 — Wimbledon QF, heavy
07-06 W Bublik (1856, Elite) 7-6 6-4 6-4 GRIND
07-04 W Sonego (1675, Strong) 4-6 6-3 6-4 7-6 GRIND
07-02 W Kypson (1550, Avg) 6-2 6-2 7-5 DOM
Deeper window: Wimbledon QF, Halle final, Stuttgart final — a full grass season
at a high level, then a clean hard-court re-entry.
K. MAJCHRZAK (30, rank 59)
07-28 W T. Paul (1914, Elite) 7-5 7-6 GRIND — Washington R1, hard
07-02 L Svajda (1769, Strong) 5 sets — Wimbledon R1 exit
06-30 W Tabilo (1460, Weak) 6-3 7-5 7-5 DOM
06-16 L Lehecka (1783, Strong) 5-7 6-7 — London R1 exit
06-14 W De Minaur (1910, Elite) 6-3 2-6 7-6 GRIND — TITLE, Hertogenbosch
Hertogenbosch title run also took out Medvedev and Auger-Aliassime.
═══ SHAPE / PATH (Step 1b) ═══
Fritz: STABLE-to-SHARPENING. DOM over a Strong opponent (Bergs 1766) in the
Washington opener is over-performing vs bracket, and it is on THIS surface —
the grass-to-hard transition is already made and made cleanly. The Zverev QF
loss is a neutral-to-negative marker (not narrow), but it is 3 weeks stale.
Majchrzak: SHARPENING on this surface, FADING before it. Two straight R1 exits
(London, Wimbledon) after the Hertogenbosch title = a fade. Then 7-5 7-6 over
Tommy Paul (1914 Elite) on hard — a genuine bracket-jump win and hard
independent evidence from THIS event. Live form has re-set upward.
Path: neither is match-sharp relative to the other — both are exactly one match
deep at the venue (Fritz 07-27, Majchrzak 07-28). No rust on either side
(both played within 2 days). Path is a wash.
Surface transition: NONE for either — both already played and won on Washington
hard. No transition risk to price.
═══ FATIGUE (Step 2) ═══
Fritz: 1 match / 14d, 85 min, last 07-27 (2 days ago), 2-set win.
Majchrzak: 1 match / 14d, 112 min, last 07-28 (yesterday), 2-set win.
(Brief reports total_sets 3 for both; the scorelines are 2-setters — using the
scorelines.)
No extreme spot: no 3-set match yesterday, no same-day double, no 3-in-3, no
best-of-5. The calibrated −4.5pp 3-set-hangover does NOT apply — Majchrzak's
yesterday win was 2 sets, which is ~neutral by the measured study.
Age modulation: Majchrzak is 30 (≥30, slow recovery) → harden one tier, so the
yesterday-vs-two-days-ago gap reads as a slight Fritz edge rather than equal.
Cumulative load is 27 minutes apart — noise. Fatigue is CONTEXT here, and by
framework it cannot create or enlarge an edge. It does not enter the thesis.
═══ NARRATIVE (Step 3) ═══
Let-down: NONE. Majchrzak's Hertogenbosch Final win was 06-14 — 45 days ago,
far outside the 0-7d window. Fritz's last Final (Halle, lost 06-21) is 38d.
TITLE: Majchrzak — Hertogenbosch 45d ago (GRASS). Live hard-court form (the
Paul win) is what matters now, not the grass title.
Venue momentum: no prior Washington history in either window.
ELO/ranking sanity: no material divergence to discount. Majchrzak's ELO is
built on Paul / De Minaur / Medvedev / Auger-Aliassime scalps — that is the
opposite of a farmed rating. Fritz ELO ≈ rank 9. Both ELOs credible.
═══ MARKETS ═══
alt_markets.handicap = null, alt_markets.totals = null, alt_markets.model = null.
No priced margin market → ml is the only stakeable market. No margin thesis
possible and none attempted.
GATES
G1 edge-origin: FAIL — pre-narrative, the best model-vs-market gap on Majchrzak
is ML 28.9% vs exchange-effective 24.75% = +4.1pp, under the 5% bar; the
consensus figure is only +2.7pp and ELO +2.5pp. Fritz's side is negative
(−5.8pp). No side clears the numeric bar without leaning on narrative.
G2 consensus: PASS (stake-tier only, never blocks) — all three models lean B,
plus hard independent this-event evidence (Majchrzak beat Paul here on hard).
Full-ladder tier would have applied had G1 cleared.
G3 fatal-risk: PASS — ML market, G3 never blocks ML; no margin bet in play.
═══ VERDICT: PASS ═══
The disagreement is real but small. Every model prices Fritz below the market,
and Majchrzak's Washington win over Tommy Paul is exactly the kind of
independent, on-surface, this-event evidence that would validate the dog — but
the biggest edge available at the executable Betfair price is +4.1pp off a
single model, and +2.7pp on consensus. That is under the 5% bar, so G1 fails
and the bet is dead by the numbers regardless of how good the story reads.
The qualitative read is also genuinely split rather than one-directional:
Majchrzak is the one who played yesterday (112 min, two tight sets, age 30),
while Fritz is two days rested off an 85-minute DOM over a Strong opponent on
the same surface. Washington's fast hard court suits Fritz's serve, and a
top-10 hard-courter at ~76% against #59 is not an obviously stale price.
Would become live if Majchrzak drifts: at 4.20 he needs ~29.8% to clear the
bar; roughly 4.55+ makes it a bet at the ML's 28.9%. Not there. Pass.