TennisEdge's AI model finds no betting value in F. Tiafoe vs T. Atmane (Washington on 2026-07-27) — the market is efficient, with F. Tiafoe favoured at 1.29.
F. Tiafoe (A) vs T. Atmane (B) — ATP Washington, hard, 1/16-finals
═══ MODELS vs MARKET ═══
ELO 0.701 | ML 0.7244 | blended 0.7186 | sim 0.7096 — full consensus ~71% Tiafoe.
Betfair ML: A 1.38 (comm-adj 73.5%), B 3.60 (comm-adj 28.8%).
Tiafoe ML edge -1.6% | Atmane ML edge -0.7% → no ML edge either side; price is fair.
ELO/rank: Tiafoe rank 30 / ELO ~1786 (agree). Atmane rank 51 but ELO ~1480 — normally
"strong rank/low ELO" flags an understated player, but Atmane's live form (2-8 last 10)
fully justifies the low rating, so no reason the 28% understates him. ELO credible.
═══ RECENT FORM (last 5 shown) ═══
TIAFOE: Lost Bublik(1836) 5s narrow (grass, Wimb) | Won Choinski(1589) 4s | Won Atmane(1480)
4s incl TB | WON HALLE TITLE beating Fritz(2011) 6-4 6-4 DOM + AAA(1751)/Shimabukuro(1743).
Strong grass block, SHARPENING through Halle then a competitive Wimb loss. Class player.
ATMANE: Lost Tiafoe(H2H) | Lost Diallo(1684) | Lost Medvedev(1836) | Won Landaluce(1771) GRIND |
Lost Borges(1553). 2-8 in window, losing to sub-1700 opponents. FADING dog in poor form.
═══ H2H ═══ Tiafoe beat Atmane 26d ago at Wimbledon (grass) 7.8-6.6 6-1 4-6 6-4 — controlled
but competitive, not a blowout. Direct recent evidence Tiafoe is the class side (as priced).
═══ PATH / SURFACE / RUST ═══ BOTH players' last match >21d ago (Tiafoe 23d, Atmane 26d) →
match-1 rust for both. BOTH transitioning grass→clay-season→now HARD → TRANSITION both sides,
depresses level and WIDENS margin variance in match one. First match at Washington for both.
Atmane serve strong (sim hold 0.853 > Tiafoe 0.816); Washington is a fast hard court → serve
holds, tiebreak risk → pushes toward OVER/closer sets, i.e. AGAINST a tight under or -3.5 lay.
═══ FATIGUE ═══ Both well-rested (0 matches/14d). No fatigue edge either way.
═══ MARKET-BY-MARKET (sim resim: total 21.7 calibrated from raw 24.6, margin 2.5) ═══
- ML: consensus ~71%, exchange fair → no edge. PASS.
- Tiafoe -3.5 @2.04: sim cover 55.2% vs 50.3% → edge +4.9%. Laying fav at -3.5 = HIGH-VARIANCE
tier (needs edge ≥10%). 4.9% < 10% → FAIL bar.
- Atmane +3.5 @1.89: edge -9.4% → no edge.
- Under 23.0 @1.98: sim under ~58.3% vs 51.8% → edge +6.5%. BUT rests entirely on a generic
3.0-game bias correction (raw total 24.6 would make OVER the play) on serve/return inputs
built off grass/clay, not hard. Fast-court + two good servers + match-1 transition variance
all point toward MORE games, contradicting the under. Model uncertainty > edge.
- Over 23.0 @1.96: edge -10.6% → no edge.
═══ GATES (best candidate = Under 23.0) ═══
GATES
G1 edge-origin: PASS — 6.5% under edge is numeric (sim vs market), not narrative-built.
G2 consensus: FAIL — no independent on-surface evidence for under; both players off hard
court, and the surface read (fast court, big servers, transition looseness)
points AGAINST under. No hard evidence supports the pick.
G3 fatal-risk: FAIL — my own read (fast Washington hard + Atmane 0.853 hold + match-1
transition variance → tiebreaks/3-setters) directly negates the under. A
bias-corrected off-surface total is exactly the biased-input segment to skip.
═══ VERDICT ═══ PASS. ML is a full model-market consensus at a fair exchange price (no edge).
The only positive nominal edges (Tiafoe -3.5 +4.9%, Under 23.0 +6.5%) both fail their bar —
the handicap is below the wide-AH 10% threshold, and the under rests on a generic calibration
of off-surface inputs that the fast-court/big-server/transition read actively contradicts.
Nothing clears the gates.