TennisEdge's AI model finds no betting value in M. Uchijima vs M. Inglis (Vancouver on 2026-07-30) — the market is efficient, with M. Uchijima favoured at 1.61.
Independent estimates of who wins, each from a different method. The blend is what the bot prices against the market.
| Method | M. Uchijima | M. Inglis |
|---|---|---|
| Elo ratings-based, surface-weighted |
49.4% | 50.6% |
| Machine learning 29-feature gradient-boosted model |
56.5% | 43.5% |
| Serve / return point-by-point simulation |
53.2% | 46.8% |
| Blend what the bot actually prices off |
54.8% | 45.2% |
| Market implied de-vigged from the opening odds |
58.5% | 41.5% |
We publish how well these score against the closing market, including where they lose: how accurate are tennis predictions?
| Player | Surface | Elo | Record | Matches |
|---|---|---|---|---|
| M. Uchijima | hard | 1671 | 139–105 | 244 |
| M. Inglis | hard | 1671 | 154–102 | 256 |
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. 2, 2026.
Level at 1–1 over 2 meetings. Full record.
| Date | Event | Surface | Winner | Score |
|---|---|---|---|---|
| July 30, 2026 | Vancouver | hard | M. Inglis | 2 - 1 |
| May 2, 2024 | ITF W100 Gifu Women | hard | M. Uchijima | 2 - 0 |
| W | C. Bucsa | 2 - 0 | Aug. 2, 2026 |
| W | H. Sakatsume | 2 - 0 | Aug. 1, 2026 |
| L | M. Inglis | 1 - 2 | July 30, 2026 |
| L | L. Khan | 0 - 2 | July 30, 2026 |
| W | H. Kinoshita | 2 - 1 | July 29, 2026 |
| L | T. Preston | 0 - 2 | Aug. 1, 2026 |
| W | M. Hontama | 2 - 0 | July 31, 2026 |
| W | M. Uchijima | 1 - 2 | July 30, 2026 |
| W | S. Costoulas | 0 - 2 | July 29, 2026 |
| W | G. Minnen | 0 - 2 | July 29, 2026 |
All July 30, 2026 results & picks →
M. Uchijima (A) vs M. Inglis (B) — Vancouver, hard, Quarter-finals MODELS vs MARKET ELO_A 49.4% | ML_A 56.5% | model_A 54.8% Betfair ML: A 1.80 (adj implied 56.8%) / B 2.10 (adj implied 48.9%) Book: A 1.61 / B 2.27 (soft) — system value_pick B @3.7% is vs the BOOK 2.27; that edge evaporates at the executable Betfair 2.10. Models split: ELO ~even (B 50.6%), ML/model lean A (B 43-45%). Near coin-flip, market efficient. DATA-QUALITY FLAG (material) Uchijima's match list is contaminated by a name-collision: W15 Brisbane (Australia) ITF records are interleaved with her Vancouver (Canada) results on the SAME dates (e.g. 07-29 shows both a Vancouver win vs Kinoshita AND a W15 Brisbane win vs Lopez; 07-30 a W15 Brisbane loss). Physically impossible — two different players. This inflates the fatigue block (642 min / 8 matches / back_to_back 4), which is therefore NOT trusted. Real Vancouver load for Uchijima: R32 07-28 (won 6-1 6-4), R16 07-29 (won 2-6 6-4 7-5, a 3-set GRIND), QF today. Legit 3-set-win-yesterday -> -4.5pp haircut, softened for age 22 (fast recovery) -> ~-3pp. Pushes ML_A ~53.5%. RECENT FORM (last 5 credible each) Uchijima (A) — at Vancouver 2-0: DOM vs Kuramochi (1452, weak), GRIND over Kinoshita (1613, Strong = over-performing). Clay season before was poor (losses to Bondar, Avanesyan, Burillo). Transitioned clay->hard but 2 hard matches absorbed it. Shape: SHARPENING on hard, match-sharp path. Inglis (B) — 36-day layoff (Wimbledon 06-23) then Vancouver: won 2 on 07-29, both DOM incl. over Minnen (1684, Strong) = over-performing. London grass TITLE 06-07 (52d ago, off-surface). Shape: SHARPENING dog, rust shaken off. Light load (~128 min, 2 quick wins). FATIGUE Uchijima: real toll = 3-set win yesterday; age 22 softens to ~-3pp (context, applied to numeric read). Cumulative-minutes figure discarded as contaminated. Inglis: 2 quick straight-set wins yesterday, ~128 min, age 28 baseline. Roughly equal-to-slightly-fresher. Not an edge either way. MARKET CHOICE (edges vs executable Betfair, comm-adj) A ML @1.80: need 56.8%; ML_A 56.5% (pre-haircut) -> -0.3%, worse after 3-set haircut. NO edge. B ML @2.10: need 48.9%; only ELO (50.6%) clears -> +1.7%, below 5% bar; ML/model negative. NO edge. B +2.5 @1.82: need 56.2%; model cover 55.9% -> -0.3%. NO edge. A -2.5 @2.06: need 49.8%; model cover 44.1% -> -5.7%. NO edge. Total O21.5 @1.95: need 52.6%; model 49.8%. Under @1.96: need 52.3%; model 50.2%. NO edge either side. AH shadow: B +2.5 @ 1.82 (passed — no edge). GATES G1 edge-origin: FAIL — no market produces a >=5% numeric model-vs-market edge at executable Betfair prices before any narrative. Best is B ML +1.7% (ELO-only, below bar); everything else negative. G2 consensus: stake-tier only (n/a — no bet). G3 fatal-risk: n/a (no margin bet staked). VERDICT: PASS. Near coin-flip that the exchange has priced efficiently; the apparent "value on B" is a stale book quote (2.27) contradicted by Betfair (2.10). No market clears the 5% bar. Contaminated fatigue data further argues against manufacturing a fatigue-based B lean. Predicted winner: M. Uchijima (slim).