TennisEdge's AI model finds no betting value in O. Oliynykova vs Ka. Pliskova (Toronto on 2026-08-03) — the market is efficient, with Ka. Pliskova favoured at 1.31.
Independent estimates of who wins, each from a different method. The blend is what the bot prices against the market.
| Method | O. Oliynykova | Ka. Pliskova |
|---|---|---|
| Elo ratings-based, surface-weighted |
4.8% | 95.2% |
| Machine learning 29-feature gradient-boosted model |
52.9% | 47.1% |
| Serve / return point-by-point simulation |
40.7% | 59.3% |
| Blend what the bot actually prices off |
4.8% | 95.2% |
| Market implied de-vigged from the opening odds |
27.8% | 72.2% |
We publish how well these score against the closing market, including where they lose: how accurate are tennis predictions?
| Player | Surface | Elo | Record | Matches |
|---|---|---|---|---|
| O. Oliynykova | clay | 1771 | 174–75 | 249 |
| Ka. Pliskova | hard | 1747 | 77–47 | 124 |
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 July 24, 2026.
Ka. Pliskova leads 1–0 over 1 meeting. Full record.
| Date | Event | Surface | Winner | Score |
|---|---|---|---|---|
| Aug. 3, 2026 | Toronto | hard | Ka. Pliskova | 2 - 0 |
| L | Ka. Pliskova | 0 - 2 | Aug. 3, 2026 |
| L | E. Avanesyan | Walk Over | July 24, 2026 |
| W | L. Romero Gormaz | 2 - 1 | July 23, 2026 |
| W | N. M. Schunk | 2 - 1 | July 21, 2026 |
| L | M. Sherif | 2 - 0 | July 18, 2026 |
| W | O. Oliynykova | 0 - 2 | Aug. 3, 2026 |
| L | I. Swiatek | 0 - 2 | July 2, 2026 |
| W | T. Valentova | 0 - 2 | June 30, 2026 |
| L | M. Bouzkova | 0 - 2 | June 20, 2026 |
| W | T. Gibson | 2 - 0 | June 19, 2026 |
All Aug. 3, 2026 results & picks →
O. Oliynykova (A, #68, 25yo) vs Ka. Pliskova (B, #130, 34yo) — WTA Toronto R1, hard.
MODELS vs MARKET (player A)
ELO 4.77% | ML 52.86% | Sim 40.66% | Market (Betfair 4.0) ~25% | Pinnacle no-vig 24.9%
A four-way violent disagreement = no single estimate is trustworthy here. ELO is a
stale/farmed Pliskova rating (former world #1 peak, decays slowly; she is now ranked
130) — grossly overrates B at 95%. ML is ranking-driven (A #68 > B #130) and ignores
the surface-class gap — overrates A. Sim gives A 40.7% but its serve/return inputs are
built on 5 CLAY matches for A and 14 grass/clay for B — off-surface, unreliable on hard.
RECENT FORM (most recent 5)
A (all CLAY): L Avanesyan (walkover/no data); W Romero Gormaz 7-6 3-6 6-2 (GRIND);
W Schunk 6-2 3-6 7-5 (GRIND); L Sherif (1799); W Burel 6-4 7-5. Solid clay run, #68,
but ZERO recent hard-court evidence. Non-clay history includes blowout losses
(Wimbledon 0-6 0-6 vs Kessler, Eastbourne 0-6 2-6) — she gets run over off clay.
B (grass/clay): L Swiatek 1-6 3-6; W Valentova; L Bouzkova; W Gibson 7-5 6-4;
W McNally 6-4 7-6. Hard-court specialist historically (big serve) but her recent
window has NO hard results either.
PATH / SHAPE / SURFACE
A: match-sharp (Hamburg QF 9d ago) BUT transitioning CLAY -> HARD (worst-surface risk).
B: RUST RISK — last match 2026-07-02 (Wimbledon), 31 days ago; transitioning grass->hard.
Both players are off-surface AND partly rusty/misprepared. Two-sided uncertainty, not
a clean edge for either side.
FATIGUE
A matches_14d 4, last 07-24 (9d rest) — no extreme spot, context only.
B well-rested (31d off) — rust, not an edge to fade cleanly.
No 3-set-hangover applies (A's last match 9d ago). No fatigue edge either way.
NARRATIVE
TRANSITION both sides (A clay->hard, B grass->hard). A's transition is onto the side
I'd have to back for any cover -> RISK, never support. No letdown/title signals.
ELO/ranking: B modest rank (130) but extreme ELO (95%) = STALE peak rating, discounted.
MARKET CHOICE
ML: only nominal "value" is B via the stale ELO (G1 fail — edge is not real). Backing
A at 4.0 (needs >27% true, min 1/0.27) rests on sim/ML off-surface inflation vs a hard
specialist — not trustworthy; market ~25% for a clay grinder on hard vs a big-serving
former #1 looks fair-to-generous, no fade value.
Handicap: A +4.5 @ Betfair 2.1. Sim covers +3.5 at 58.8% -> +4.5 ~62%, nominal ~13% edge,
high-variance tier. But the cover rests on off-surface sim inputs AND A's documented
blowout losses on non-clay (0-6 0-6) — a clay player on hard vs a hard specialist can
get run over. That is a self-negating margin risk -> G3 hard veto. PASS.
Total: Over 19.5 @ 1.89, sim 58.3% -> edge ~4.3% (below 5% bar) and off-surface inputs. PASS.
AH: A +4.5 @ 2.10 (shadow record; passed on G3 margin veto).
GATES
G1 edge-origin: FAIL — the only market-vs-model gap is B via a stale/farmed Pliskova
ELO; A-side gaps (ML/sim) are off-surface artifacts. No real pre-narrative numeric edge.
G2 consensus: PASS(tier) — models violently split, no coherent consensus; N/A to stake.
G3 fatal-risk: FAIL (handicap) — A +4.5 cover self-negated by clay-player-on-hard
blowout risk vs a hard specialist (0-6 0-6 non-clay losses on record).
VERDICT: PASS. Four estimates for A span 4.8%–52.9% — every model is compromised
(stale ELO, ranking-blind ML, off-surface sim). The exchange price (~25%) is the sharpest
input and I have no articulable surface-specific information to beat it. Both markets with
nominal edge fail a hard gate.