TennisEdge's AI bot is backing H. Kinoshita (moneyline) at 3.20 in M. Uchijima vs H. Kinoshita (Vancouver on 2026-07-29).
═══ M. Uchijima (A) vs H. Kinoshita (B) — Vancouver (hard), 1/8-finals ═══
MODELS vs MARKET
ELO A 54.0% | ML A 60.2% | Blend A 57.2% -> B (Kinoshita) 42.8% / 39.8% / 46.0%
Market: book A 1.39 / B 2.90; Betfair A 1.41 (impl 70.9%) / B 3.20.
Both models sit well below the market on A: blended B 42.8% vs Betfair B implied
32.4% (comm-adj) = +10.5% edge on B BEFORE any fatigue/narrative adjustment.
value_pick B, value_edge 10.4% corroborated.
RECENT FORM / SHAPE (last 5)
Uchijima (A, rank 107): 07-29 W vs Lopez(1500) 6-3 3-6 7-5 GRIND (tight 3rd vs weak);
07-28 W vs Kuramochi(1452) 6-1 6-4 DOM; 07-24 L QF Arakawa 1-6 3-6; 07-23 W Das;
07-22 W Ipunesso. Deeper window = repeated losses at higher levels (clay/hard
June-July vs 1490-1783). LEVEL-DROPPER at Vancouver: wins all vs 1450-1500 fields,
inflated stats vs a weak draw. Shape FADING at true level / grinding weak fields.
Kinoshita (B, rank 233, age 19): 07-28 W vs Okamura(1514) 6-1 6-4 DOM (sharpening);
prior window mostly losses (Kalieva, Hosogi, Shymanovich). Freshest result a clean
DOM. Path: only 2 matches at venue.
ELO/RANK SANITY
A ranked 107 but ELO ~coinflip (strong-rank / low-ELO). Checked for layoff artifact:
none, she has a full recent match load; low ELO is GENUINE poor form (1-6/3-6 losses),
so ELO is credible here, not an understatement. Edge stands.
FATIGUE
A: 5 matches/7d, matches_14d 6, 471 court-min, back_to_back 3, WON a 3-set match 07-29
(same-day/day-before this match). EXTREME spot (3rd match in 3 days + just won a
3-setter). Calibrated 3-set-hangover -4.5pp applies; age 22 (<=22) softens one tier
-> ~-2 to -3pp on A. Confirms, does not create, the B edge.
B: 2 matches/7d, 216 min, rested by comparison. Won yesterday but 2-set DOM (~neutral).
Market (1.39/1.41) is pure ranking pricing: ignores BOTH A's poor true-level form
(models capture it) AND A's fatigue (nobody captures it). Value is on the tired side.
GATES
G1 edge-origin: PASS - +10.5% on B from blend-vs-Betfair BEFORE fatigue; fatigue only confirms.
G2 consensus: PASS - 2 models (ELO+ML) both under market on A -> full stake tier.
G3 fatal-risk: PASS - ML bet, G3 non-blocking; main risk is A higher rank/match-sharp.
MARKETS
Total: over 52.9% vs 2.02 (impl 49.5%) = +3.3% < bar -> pass. Handicap model null; A -4.5
high-variance and no sim anchor for a cover -> pass. Best expression = B ML @ Betfair 3.20
(min_odds 2.45, clears).
VERDICT: BET B (Kinoshita) ML @ 3.20, 1.5u. Clean 10.5% model-market edge, independently
confirmed by A's extreme same-day fatigue and level-dropper inflation. Moderated to 1.5u
for the genuine counter (A ranked 107 & match-sharp, noisy ITF context).