TennisEdge's AI model finds no betting value in J. Tjen vs R. Sramkova (Washington on 2026-07-27) — the market is efficient, with J. Tjen favoured at 1.31.
J. Tjen (A, rank 40, age 24) vs R. Sramkova (B, rank 118, age 29) — Washington, hard. ═══ MODELS ═══ ELO A 63.9% | ML A 42.0% | blend A 57.0% | point-sim A 48.9% (B 51.1%). Market: book 1.31/3.25 stale (76% overround). EXCHANGE (truth) A 1.61 / B 2.60 → devig A ~61.8% / B ~38.2%. Sim expected margin 0.1 games; matchup-adj serve 0.574 (A) vs 0.576 (B) — structurally a coin-flip. MODEL CHAOS: ELO+market (~62% A) vs ML+sim (~49% A). No clean consensus. ═══ RECENT FORM ═══ Tjen (last 5): L Kasatkina (grass, 1652, 3-set TB); W Fernandez (grass, GRIND); L McNally (grass); L Maria (grass 6-3 6-2); L Minnen (grass). Mostly losses across grass/clay. LAST MATCH 2026-07-01 = 26 DAYS AGO → rust risk. Last surfaces grass → hard = TRANSITION. Shape: FADING/UGLY on off-surfaces. But she is a legit rank-40 hard-courter whose sample is suppressed by a losing grass swing. Sramkova (last 5): W Knutson (hard, Washington Final, DOM); W Ansari (hard, Washington SF, DOM); L Kostovic (grass SF, close 3-set); W Hruncakova (grass, 7-6 7-6 GRIND); L Primorac (clay). MATCH-SHARP on this surface (2 hard wins in 2 days). Path: 2 matches deep at venue = played-in. Blowout losses (6-2 6-2 Grabher, bagel sets Urhobo/Werner) were all on CLAY, not hard. ═══ FATIGUE (Step 2 calibration) ═══ Sramkova: matches_7d 2, back_to_back 1, 132 court-min over 2 matches (~66 min avg — quick DOM wins), played 2026-07-26. Would be 3rd match in 3 days today (borderline extreme spot) BUT prior two were 66-min routine wins, not grinds → context, not extreme. Age 29 = baseline recovery. Per calibration the market over-prices this visible fatigue; not a bettable edge either way. Tjen: 0 matches 14d, fully rested but that = 26-day layoff rust, not an advantage. ═══ NARRATIVE ═══ Sramkova won a Final 1d ago → LETDOWN signal — but she is the UNDERDOG here, so let-down argues against backing her, not for. Tjen: grass→hard TRANSITION + 26-day rust depress her level and WIDEN margins (cuts both ways on any cover). ELO/ranking: Tjen ELO≈rank (credible). Sramkova rank 118 but recent strong results at ITF/low-ELO fields (opp 1452–1517) → sim serve inputs likely inflated on the hard sample. ═══ MARKET SCAN ═══ Tjen ML @1.61 (eff 1.58): needs 63.3% — blend 57%, sim 49%, ML 42% all below; market already ≈ fair → NO edge. Sramkova ML @2.60 (eff 2.52): edge off blend 42.95% = 0.4295 − 0.397 = +3.3% → below 5% bar. Sramkova +3.5 @1.86 (eff 1.817): edge off SIM 66.3% = +11.3%. BUT sim diverges from ELO+market and under-rates Tjen's class (suppressed losing-sample serve stats). Discount sim toward market view (Tjen ~62% fav → +3.5 cover ~59%) → edge collapses to ~4%, below bar. Totals 21.0: sim over(20.5) 55.3% but market line higher (21.0) → over ~53% vs breakeven ~53% = ~0 edge → PASS. Handicap variance tier: dog +3.5 = tight (5% bar); edge is the binding fail, not min-odds. ═══ GATES ═══ GATES G1 edge-origin: FAIL — 11% cover edge is real off the sim, but the sim diverges from ELO+market and under-rates Tjen's hard-court class; discounting to the reliable view drops the edge to ~4%, below the 5% bar. G2 consensus: PASS — ML + live Washington hard results independently support B's competitiveness; not fighting a full consensus. G3 fatal-risk: FAIL — class gap (rank-40 hard-courter vs 118) means if Tjen's level shows once rust clears she wins by 4+; plus a possible 3rd-match-in-3-days that could run Sramkova over. A live negating risk to the +3.5 cover. ═══ VERDICT: LEAN (Sramkova side) ═══ Genuine directional value on Sramkova (+3.5, secondarily ML): ML + point-sim both favor her, she is match-sharp on hard while Tjen is rusty and transitioning off grass, and her blowout losses were clay-only. But the fat cover edge lives entirely on a sim that diverges from ELO+market and plausibly over-rates her; discounted to the reliable view the edge is only ~4% and a real class-gap/fatigue fatal-risk sits on the cover (G1 + G3 fail). Directional lean, insufficient conviction to stake → LEAN, no position.