TennisEdge's AI model finds no betting value in A. Falei vs U. Radwanska (Warsaw on 2026-08-03) — the market is efficient, with A. Falei favoured at 1.09.
Independent estimates of who wins, each from a different method. The blend is what the bot prices against the market.
| Method | A. Falei | U. Radwanska |
|---|---|---|
| Elo ratings-based, surface-weighted |
87.6% | 12.4% |
| Machine learning 29-feature gradient-boosted model |
87.2% | 12.8% |
| Serve / return point-by-point simulation |
55.7% | 44.3% |
| Blend what the bot actually prices off |
79.2% | 20.8% |
| Market implied de-vigged from the opening odds |
86.0% | 14.0% |
We publish how well these score against the closing market, including where they lose: how accurate are tennis predictions?
| Player | Surface | Elo | Record | Matches |
|---|---|---|---|---|
| A. Falei | hard | 1659 | 144–76 | 220 |
| U. Radwanska | hard | 1313 | 75–98 | 173 |
Our own ratings, computed from match results since 2018. Players with fewer than 5 rated matches are omitted rather than shown at a provisional starting value. Last updated Aug. 3, 2026.
A. Falei leads 1–0 over 1 meeting. Full record.
| Date | Event | Surface | Winner | Score |
|---|---|---|---|---|
| Aug. 3, 2026 | Warsaw | hard | A. Falei | 2 - 0 |
| W | U. Radwanska | 2 - 0 | Aug. 3, 2026 |
| L | V. Valdmannova | 2 - 0 | July 17, 2026 |
| W | K. Penickova | 1 - 2 | July 15, 2026 |
| W | Y. Yuan | 2 - 1 | July 14, 2026 |
| L | A. Kubareva | 2 - 0 | June 25, 2026 |
| L | A. Falei | 2 - 0 | Aug. 3, 2026 |
| W | M. Podlinska | 0 - 2 | Aug. 2, 2026 |
| L | V. Ryser | 2 - 0 | July 30, 2026 |
| W | A. Sorska | 2 - 0 | July 29, 2026 |
| L | E. Adeshina | 2 - 0 | July 14, 2026 |
All Aug. 3, 2026 results & picks →
A. Falei (A) vs U. Radwanska (B) -- Warsaw, hard, 1/16-finals
MODELS
ELO Falei 87.6 pct | ML Falei 87.2 pct | Sim(model) Falei 79.2 pct
Market: odds_a 1.09 (impl ~86 pct), odds_b 6.70 (impl 14.9 pct). value_pick B edge 6.8 pct -- driven ENTIRELY by the sim.
ELO + ML + market form a tight consensus on Falei ~86-88 pct. Only the sim dissents, giving B 20.8 pct.
RECENT FORM (last 5)
Falei (A, 22, rank 231): ITF level. 07-17 Lost Valdmannova(1500); 07-15 W Penickova(1500) 3-set grind;
07-14 W Yuan(1500) 3-set grind; 06-25 Lost Kubareva(1436); 06-24 W Rey Garcia(1251) DOM. Mixed, grindy vs avg/weak
fields. Shape STABLE-to-FADING. matches_14d 0 -- last match 07-17 (16d ago), mild rust, under the 21d threshold.
Radwanska (B, 35, rank 693): match-sharp. 08-02 W Warsaw final vs Podlinska(1186) 6-1 6-3 DOM (title);
07-30 Lost Ryser(1500) 6-2 6-0; 07-29 W Sorska(1500) 6-1 6-3; 07-14 Lost Adeshina(1391). SHARPENING but opposition
very weak (elo 1186-1500). Won an ITF title today.
FATIGUE
Falei well-rested (0/14d). Radwanska 3 matches/7d, 9 sets, 173min, won a 2-set final TODAY (08-02), back_to_back 1.
2-set win today ~ roughly neutral toll (~-2pp); age 35 hardens slightly. Net small fatigue disadvantage, not extreme.
Market over-prices visible fatigue, so this is context only, not an edge maker.
NARRATIVE
Radwanska TITLE: Warsaw ITF final won 08-02 (hard). No surface transition (all hard). ELO/ranking: Falei rank 231 and
ELO-fav is credible; Radwanska rank 693 with weak-field wins -- her raw serve/return stats are inflated by elo under
1500 opposition, exactly what biases the opponent-unadjusted sim UPWARD for her. Sim likely understates Falei; true B
prob is nearer ELO/ML ~12.5 pct, not 20.8 pct.
PRICING
No exchange price: betfair_ml null -- book-quote path (needs edge >=8 pct + book flag).
B edge off sim = 20.8 - 14.9 = +5.9 pct -- below the 8 pct no-exchange bar AND rests on the least reliable model.
B edge off ELO/ML consensus = ~12.5 - 14.9 = -2.4 pct (negative). Falei ML unbettable: min odds 1.21 > 1.09.
MARKET CHOICE
alt_markets: none at analysis time. ML only. AH: not quoted.
GATES
G1 edge-origin: FAIL -- only positive edge comes from the sim, which two math models + the market contradict; sim
inputs are weak-field, opponent-unadjusted stats that inflate the dog. No robust model-vs-market edge.
G2 consensus: stake-tier only -- lone_signal (sim), contaminated.
G3 fatal-risk: n/a (ML, no margin bet).
VERDICT: PASS. Full ELO+ML+market consensus on a heavy favorite; the lone dissent is the opponent-unadjusted sim biased
by Radwanska weak-field stats; no exchange liquidity and the book-quote edge (5.9 pct) is below the 8 pct bar. No bet.