A. Ruzic vs M. Keys — Prediction & Analysis

TennisEdge's AI model finds no betting value in A. Ruzic vs M. Keys (Toronto on 2026-08-04) — the market is efficient, with M. Keys favoured at 1.25.

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. 3, 2026.

Head to head

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

Recent form

All Aug. 4, 2026 results & picks →

Full analysis

A. Ruzic vs M. Keys — WTA Toronto, hard, 1/32-finals

═══ MODELS vs MARKET ═══
Book: Ruzic 3.90 (raw 25.6%) / Keys 1.25 (80.0%); de-vigged implied_prob_a 24.3%.
ELO A 34.2% | ML A 32.1% | blended model A 30.8%.
App value_pick A, value_edge +6.5% (vs de-vigged).
Edge vs the RAW BOOK price (the only quote available): 30.8% − 25.6% = +5.1%.
No Betfair exchange market (alt_markets.handicap/totals/betfair_ml/model all null),
no Pinnacle reference. Pricing Discipline #4 therefore applies: an ML bet with no
executable exchange price needs edge >= 8% vs the book quote. 5.1% does not reach it.
Only the single most generous model (ELO, 34.2%) would scrape 8.6%; the two others
and the blend do not.

ELO/ranking sanity: Ruzic ELO ~1617 vs rank 59, Keys ELO ~1869 vs rank 7 — both
coherent, no farmed-rating distortion, no reason to distrust either rating.

═══ RECENT FORM (last 5 each) ═══
A. Ruzic (23, #59)
  03-Aug Toronto hard  W  vs L. Boisson (1559)   6-4 1-6 3-6   GRIND, 3 sets, Avg opp
  20-Jul Prague clay   L  vs D. Salkova (1577)   6-7 4-6
  01-Jul Wimbledon gr  L  vs J. Ostapenko (1764) 6-2 6-0 (straight-sets loss)
  29-Jun Wimbledon gr  W  vs D. Semenistaja (1536) 6-3 3-6 6-3  GRIND
  22-Jun Eastbourne gr L  vs P. Marcinko (1458)  6-7 6-4 6-7   loss to Weak opp
B. M. Keys (31, #7)
  27-Jul Washington hd L  vs L. Samsonova (1677) 3-6 6-4 6-4   competitive 3-setter
  06-Jul Wimbledon gr  L  vs L. Noskova (1874)   4-6 6.2-7.7   narrow loss to Elite
  04-Jul Wimbledon gr  W  vs A. Anisimova (1887) 3 sets        WIN over ELITE
  02-Jul Wimbledon gr  W  vs K. Swan (1637)
  30-Jun Wimbledon gr  W  vs K. Day (1549)
  (27-Jun Eastbourne FINAL W vs T. Maria (1824) — title)

═══ SHAPE / PATH (Step 1b) ═══
Ruzic: FADING-to-mixed. The June grass titles (Nottingham 14-Jun, London 07-Jun)
are 50+ days stale; since then L Nottingham 1/16, L London, L Wimbledon R2 (6-2 6-0
blowout), L Prague to a 1577, and a scrappy 3-set escape vs a 1559 yesterday. A
GRIND over an Avg opponent = UNDER-performing vs bracket. Evidence: 6-7 loss to
Marcinko (1458) and the 3-setter vs Boisson.
  Path: POSITIVE — one match already in the bank at this venue on this surface,
  so she is conditions-adjusted and off the clay/grass block; the transition risk
  is already spent. This is her genuine plus in the matchup.
Keys: STABLE. Eastbourne title over Maria (1824), Wimbledon 4R with a win over
Anisimova (1887 — Elite), narrow loss to Noskova (1874). Only blemish is a
competitive 3-set loss to Samsonova. Nothing here is fading.
  Path: 8 days since last match, first match at this venue — mild rust risk, the
  one genuine argument for the dog, and it is a weak one for a 31-year-old #7
  who has been playing continuously all season.

═══ FATIGUE (Step 2) ═══
Ruzic: 1 match/14d, 3 sets, 98 min, last match 03-Aug (YESTERDAY), back_to_back 0.
Keys: 1 match/14d, 110 min, last match 27-Jul (8 days ago).
Cumulative load is trivial on both sides — not an edge either way per the 2026-07-27
calibration.
BUT the calibrated 3-set-hangover applies and it lands on the DOG: Ruzic WON a
3-set match yesterday → −4.5pp haircut (surface-agnostic; age 23 = baseline band,
no softening of the haircut itself, applied once, not stacked). That takes the
blended 30.8% to ~26.3% — i.e. essentially the raw book price of 25.6%. The one
numeric adjustment available here removes what little edge existed rather than
adding to it.

═══ NARRATIVE (Step 3) ═══
Let-down: Ruzic's most recent Final = Nottingham 14-Jun (Won), 51d ago → None.
  Keys' most recent Final = Eastbourne 27-Jun (Won), 38d ago → None.
Title recency: TITLE Nottingham 51d ago (grass) — stale, and on a different surface.
Venue momentum: Ruzic 1-0 at Toronto in window; Keys no matches at Toronto in window.
Surface transition: neither player flagged — Ruzic already played hard yesterday,
Keys' last match was hard court in Washington.
H2H (independent, hard evidence, AGAINST the dog): Keys beat Ruzic 28-May-2026 at
Roland Garros in STRAIGHT SETS. It is clay rather than hard, so it is not decisive,
but a straight-sets loss 10 weeks ago is not a foundation for a live dog thesis;
if anything hard court flatters the bigger hitter here.

═══ MARKET CHOICE ═══
alt_markets: none at analysis time
AH: not quoted
No exchange handicap, total or match-odds market, and no Pinnacle line to
corroborate the book quote. ML is the only market and it is book-priced only.
Probe tier (3–5% band) explicitly requires an EXECUTABLE exchange price and is
therefore unavailable here — never book-only. No refutation check run (PASS
candidates get no web access).

GATES
  G1 edge-origin: FAIL — pre-narrative edge vs the only available (book) price is
    +5.1%, short of the 8% bar that Pricing Discipline #4 imposes when no exchange
    ML price exists; probe tier is barred without an executable price. Nothing but
    the single most bullish model reaches the bar, and the −4.5pp 3-set hangover
    on the dog erases even that.
  G2 consensus:   PASS(tier) — all three models lean the same way (A underpriced),
    so this would have been full-ladder eligible on tier grounds; G2 never blocks.
  G3 fatal-risk:  PASS — n/a, ML bet, G3 no longer blocks ML (and no margin bet is
    on the table since no handicap/total is quoted).

═══ VERDICT: PASS ═══
The model-vs-market gap is real but small and it exists only against a book quote
we cannot execute and cannot corroborate. Every piece of qualitative evidence runs
against enlarging it: Ruzic is the one carrying the measured 3-set hangover, her
shape is under-performing her bracket (grinding past 1559s, blown out by Ostapenko,
beaten by a 1458), her titles are 50-day-old grass, and Keys beat her in straight
sets 10 weeks ago while otherwise beating Elite opposition. Her only genuine plus —
a match already played at the venue against Keys' 8-day layoff — is precisely the
kind of soft rust argument that cannot manufacture 3 points of missing edge. The
disciplined outcome is no bet.