M. Uchijima vs H. Sakatsume — Prediction & Analysis

TennisEdge's AI model finds no betting value in M. Uchijima vs H. Sakatsume (Toronto on 2026-08-01) — the market is efficient, with M. Uchijima favoured at 1.69.

Model probabilities

Independent estimates of who wins, each from a different method. The blend is what the bot prices against the market.

We publish how well these score against the closing market, including where they lose: how accurate are tennis predictions?

Elo ratings

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.

Head to head

M. Uchijima leads 1–0 over 1 meeting. Full record.

Recent form

All Aug. 1, 2026 results & picks →

Full analysis

═══ M. Uchijima vs H. Sakatsume — WTA Toronto (hard), Final ═══

MODELS vs MARKET
  Book: A 1.69 (impl 55.3% devigged) | B 2.09
  ELO:   A 46.1%  (favours B)
  ML:    A 54.9%  (favours A)
  Blend: A 51.8%  — essentially a coin flip
  ELO and ML DISAGREE on who is favourite; market splits them. No model consensus,
  and the blend sits right on the market's number. No side has a numeric edge vs the
  EXECUTABLE (book) price: A model 51.8% < 1/1.69=59.2% (negative); B model 48.2% vs
  1/2.09=47.8% = +0.4% (nil). value_edge 3.45% is measured against the devigged implied,
  not the executable price, and even that is below the 5% bar.

EXCHANGE / PRICING
  alt_markets: none at analysis time  (betfair_ml null, no handicap/total quoted)
  No exchange liquidity → ML bet would require edge >= 8% vs book quote (Pricing #4).
  Best available edge (B) is ~0.4% vs book. Fails the book-only bar by a wide margin.
  AH: not quoted

DATA QUALITY (material caveat)
  Uchijima's match list contains parallel same-day entries at Vancouver (hard) AND
  W15 Brisbane 3 (Australia) on 07-29/07-30 — physically impossible; a name-collision
  artifact mixing two players. This inflates her fatigue block (5 matches/7d, 9/14d,
  763 court-min) and corrupts the recent-form read. Any signal derived from her history
  is unreliable → an additional reason not to lean on the metrics here.

RECENT FORM (most-recent 5, classification uses all)
  Uchijima (22, rank 107): mixed/contaminated. Genuine hard-court Vancouver run shows
    W vs Kuramochi (6-1 6-4 DOM/Weak), W vs Kinoshita (2-6 6-4 7-5 GRIND/Avg),
    L vs Inglis (4-6 7-6 3-6 GRIND/Avg). Clay results before were poor (several straight
    losses). Shape STABLE-to-FADING, but read is muddied by data mixing.
  Sakatsume (24, rank 131): L vs Hontama 6-2 6-2, W vs Arseneault, W vs Wong (grind),
    L vs Crawley. Grass-season Nottingham title (06-14) but nothing dominant on hard.
    Shape STABLE/unremarkable.

FATIGUE (Step 2) — CONTEXT ONLY
  Both played 07-30. Uchijima's raw load looks heavy but is corrupted by the collision
  (both players 22 and 24, near-baseline recovery). No extreme spot cleanly established
  for either; cumulative minutes alone is never an edge. Not decisive.

NARRATIVE
  No clean let-down (no recent genuine Final win in-window that is trustworthy).
  Surface: both coming onto hard from mixed/hard recent play — no clear transition edge.

GATES
  G1 edge-origin: FAIL — no numeric model-vs-market edge at the executable price
                   (best is B +0.4% vs book; A negative). Below 5% / 8%-book bar.
  G2 consensus:   PASS (stake-tier only) — but moot; would be lone/contaminated anyway.
  G3 fatal-risk:  N/A — no margin bet contemplated.

VERDICT: PASS. Models disagree with each other, market prices the split correctly, no
exchange liquidity, and the only positive edge (B) is ~0.4% vs the book — far under the
8% book-only bar. Corrupted history removes any qualitative tiebreaker. Clear pass.