TennisEdge's AI model finds no betting value in P. Kudermetova vs A. Alexander (Toronto on 2026-08-01) — the market is efficient, with P. Kudermetova favoured at 1.02.
Independent estimates of who wins, each from a different method. The blend is what the bot prices against the market.
| Method | P. Kudermetova | A. Alexander |
|---|---|---|
| Elo ratings-based, surface-weighted |
95.2% | 4.8% |
| Machine learning 29-feature gradient-boosted model |
92.1% | 7.9% |
| Serve / return point-by-point simulation |
62.6% | 37.4% |
| Blend what the bot actually prices off |
86.5% | 13.5% |
| Market implied de-vigged from the opening odds |
95.4% | 4.6% |
We publish how well these score against the closing market, including where they lose: how accurate are tennis predictions?
| Player | Surface | Elo | Record | Matches |
|---|---|---|---|---|
| P. Kudermetova | hard | 1665 | 127–77 | 204 |
| A. Alexander | hard | 1151 | 3–11 | 14 |
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.
P. Kudermetova leads 1–0 over 1 meeting. Full record.
| Date | Event | Surface | Winner | Score |
|---|---|---|---|---|
| Aug. 1, 2026 | Toronto | hard | P. Kudermetova | 2 - 0 |
| L | A. Korneeva | 1 - 0 | Aug. 3, 2026 |
| W | A. Alexander | 2 - 0 | Aug. 1, 2026 |
| L | E. Svitolina | 0 - 2 | July 31, 2026 |
| W | C. Bucsa | 1 - 2 | July 28, 2026 |
| W | V. Lepchenko | 2 - 1 | July 26, 2026 |
| L | P. Kudermetova | 2 - 0 | Aug. 1, 2026 |
| L | M. Kuramochi | 2 - 0 | July 28, 2026 |
| W | M. Okalova | 2 - 1 | July 25, 2026 |
| L | M. Kuramochi | 0 - 2 | July 23, 2026 |
| W | M. Ostrzygalo | 0 - 2 | July 22, 2026 |
All Aug. 1, 2026 results & picks →
═══ P. Kudermetova (A) vs A. Alexander (B) — Toronto, hard, WTA Final ═══
DATA-INTEGRITY FLAG (dominant): Player B "A. Alexander" listed age 56 / rank 5 — nonsense.
Her real recent history is ITF W35/W75 (Vancouver, Saskatoon, Granby), opponents ELO 1161-1459,
lost the Vancouver final to a 1448 and got 6-0 6-2 by a 1352. This is a low-level ITF player,
NOT a WTA world #5. "Ranking 5" is corrupted/name-collision data. Player A (Kudermetova, rank 123,
age 23) is a genuine WTA main-tour player (slam-final references in her window).
MODELS vs MARKET
ELO A 95.2% | ML A 92.1% | model_prob_a 86.5% | market implied A 95.4% (odds 1.02) / B 21.0
Value model flags B edge 0.088 off model_prob_a (B 13.5%). ELO & ML both put B at 5-8%.
The class reality (WTA tour vs ITF W35) says B's true prob sits AT/BELOW the ELO estimate.
RECENT FORM (last 5)
A (Kudermetova): L Svitolina(2025) 6-7 4-6 (competitive vs elite) | W Bucsa(1761) | W Lepchenko(1610)
Final | W Yamaguchi(1518) SF | L Sakkari(1500). WTA-level throughout, STABLE.
B (Alexander): L Kuramochi(1448) Vancouver Final | W Okalova(1344) | L Kuramochi(1459) |
W Ostrzygalo(1161) | L Sema(1352) 6-0 6-2. ITF W35 level, FADING.
FATIGUE
A: 4 matches/7d, 416min, back_to_back 1; last match 07-31 was a 2-SET LOSS (6-7 4-6) -> no
3-set hangover haircut; age 23 recovers fast anyway. Context only, no edge.
B: 2 matches/7d, last 07-28. Not material.
NARRATIVE: No usable let-down/venue signal. No surface transition (all hard).
PRICING
Back A @ 1.02: win ~95%, min_odds = 1.05/0.95 = 1.105. Executable 1.02 < 1.105 -> AUTO-PASS.
Back B @ 21.0: commission-adj effective 20.0, break-even 5%. Credible true prob ~5% (ELO) ->
edge ~= 0, min_odds ~= 21 = current price. Favorite-longshot trap; "value" is a model artifact
off corrupted opponent data. No real edge.
MARKET CHOICE
alt_markets: none at analysis time (handicap/totals null; only betfair_ml present)
AH: not quoted
VERDICT: PASS. Favorite fails min-odds; longshot edge is illusory (model_prob_a underweights a
class mismatch built on corrupted rank/age data). Nothing bettable on either side.