TennisEdge's AI model finds no betting value in A. Blinkova vs E. Kalieva (Toronto on 2026-08-01) — the market is efficient, with A. Blinkova favoured at 1.72.
Independent estimates of who wins, each from a different method. The blend is what the bot prices against the market.
| Method | A. Blinkova | E. Kalieva |
|---|---|---|
| Elo ratings-based, surface-weighted |
43.0% | 57.0% |
| Machine learning 29-feature gradient-boosted model |
42.0% | 58.0% |
| Serve / return point-by-point simulation |
37.0% | 63.0% |
| Blend what the bot actually prices off |
44.5% | 55.5% |
| Market implied de-vigged from the opening odds |
53.8% | 46.2% |
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. Blinkova | hard | 1646 | 101–95 | 196 |
| E. Kalieva | hard | 1684 | 90–84 | 174 |
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.
E. Kalieva leads 1–0 over 1 meeting. Full record.
| Date | Event | Surface | Winner | Score |
|---|---|---|---|---|
| April 21, 2026 | Madrid | clay | E. Kalieva | 2 - 1 |
| L | R. Sramkova | 2 - 0 | Aug. 3, 2026 |
| W | H. Watson | 2 - 0 | Aug. 1, 2026 |
| L | S. Bejlek | 1 - 2 | July 21, 2026 |
| L | A. Sasnovich | 1 - 2 | July 14, 2026 |
| L | A. Zantedeschi | 0 - 2 | July 10, 2026 |
| L | K. Liutova | 2 - 0 | Aug. 1, 2026 |
| W | P. Stearns | 2 - 1 | July 31, 2026 |
| W | Z. Sonmez | 1 - 2 | July 29, 2026 |
| W | K. Sebov | 0 - 2 | July 28, 2026 |
| L | C. Dolehide | 2 - 1 | July 24, 2026 |
All Aug. 1, 2026 results & picks →
BLINKOVA (A) vs KALIEVA (B) — WTA Toronto Final, hard MODELS vs MARKET ELO 0.43 / ML 0.42 / blended 0.445 for Blinkova -> all three favor KALIEVA (~55%). Book 1.72/2.0 ; Betfair 1.78 (A) / 2.20 (B). value_pick B, edge 9.2%. The EXCHANGE — sharpest single estimate — firmly favors BLINKOVA. Models disagree on who is favorite. ELO/RANKING SANITY Kalieva ELO-favored yet ranked 140 vs Blinkova 97 = "high ELO, modest rank". Her recent slate is W100 ITF (Evansville, Cary, Indian Harbour Beach) + Memphis qualification vs ~1500-level fields -> farmed-rating suspicion. Discount the model edge; back only if live form independently confirms. RECENT FORM / SHAPE Blinkova (27): FADING. Lost 4 of last 5 — L Bejlek 6-7 6-3 1-6 (clay), L Sasnovich(1500) 6-3 2-6 2-6 on HARD, L Zantedeschi 4-6 6-7 (clay); wins were grindy clay (Andreeva 7-6 7-5). Rested but RUSTY (last match 07-21, 11d). Kalieva (23): SHARPENING on hard but MESSY. W Stearns(1781) 6-1 2-6 6-3 (genuine quality win), but 3-set scrapes vs 1500-level (Sonmez, Sebov). Match-sharp (3-deep at venue) yet running on fumes. FATIGUE Kalieva: WON a 3-set match YESTERDAY (07-31) -> calibrated -4.5pp hangover haircut, age 23 baseline (no modulation). 6 matches/14d, 18 sets, 576 court-min, back-to-back run; this would be a 4th match in days. Extreme spot, but haircut applied ONCE (no stacking). This is a genuine dog priced on quality, not a fatigue-drifted strong player, so the "value on the tired side" exception does NOT apply. Blinkova: rested (11d off) — rust risk, not fatigue. PRICING Model B 0.5548 raw. Post-hangover 0.5098. Effective Betfair (x0.95 comm) = 2.14 -> implied 0.4673. Edge raw 8.75%; edge post-legitimate-fatigue 4.25% -> BELOW 5% bar. GATES G1 edge-origin: PASS — 8.75% raw edge exists from model-vs-market before any narrative. G2 consensus: stake-tier only — 3-model support = full ladder, BUT signal contaminated (farmed ELO) -> would cap 0.5-1u lone_signal. G3 fatal-risk: n/a (ML market; no margin bet). VERDICT: PASS. The only edge is model-driven off a likely-farmed ELO (Kalieva's field is W100/quali vs 1500s); the calibrated 3-set-hangover haircut alone drops the edge to 4.25%, below bar; and the sharpest estimate (exchange) sits firmly on the other side with no specific info it fails to price — if anything I have MORE reason to fade Kalieva (fatigue) than the market shows. No qualitative validation to back either side. Pass. MARKET CHOICE alt_markets handicap/totals null -> ml only. betfair_ml present (1.78/2.20). AH: not quoted.