TennisEdge's AI model finds no betting value in F. Cobolli vs A. Fery (Wimbledon on 2026-07-08) — the market is efficient, with F. Cobolli favoured at 1.40.
Independent estimates of who wins, each from a different method. The blend is what the bot prices against the market.
| Method | F. Cobolli | A. Fery |
|---|---|---|
| Elo ratings-based, surface-weighted |
50.9% | 49.1% |
| Machine learning 29-feature gradient-boosted model |
47.4% | 52.6% |
| Serve / return point-by-point simulation |
21.8% | 78.2% |
| Blend what the bot actually prices off |
47.9% | 52.1% |
| Market implied de-vigged from the opening odds |
67.4% | 32.6% |
We publish how well these score against the closing market, including where they lose: how accurate are tennis predictions?
| Player | Surface | Elo | Record | Matches |
|---|---|---|---|---|
| F. Cobolli | clay | 1896 | 132–82 | 214 |
| A. Fery | hard | 1722 | 118–49 | 167 |
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 15, 2026.
A. Fery leads 2–0 over 2 meetings. Full record.
| Date | Event | Surface | Winner | Score |
|---|---|---|---|---|
| July 8, 2026 | Wimbledon | grass | A. Fery | 3 - 0 |
| Jan. 18, 2026 | Australian Open | hard | A. Fery | 3 - 0 |
| L | R. A. Burruchaga | 0 - 2 | July 15, 2026 |
| L | A. Fery | 0 - 3 | July 8, 2026 |
| W | A. De Minaur | 0 - 3 | July 6, 2026 |
| W | K. Khachanov | 2 - 3 | July 4, 2026 |
| W | J. Duckworth | 1 - 3 | July 2, 2026 |
| L | A. Zverev | 0 - 3 | July 10, 2026 |
| W | F. Cobolli | 0 - 3 | July 8, 2026 |
| W | G. Dimitrov | 2 - 3 | July 6, 2026 |
| W | Z. Bergs | 2 - 3 | July 4, 2026 |
| W | O. Virtanen | 3 - 1 | July 2, 2026 |
All July 8, 2026 results & picks →
═══ F. Cobolli vs A. Fery — Wimbledon QF (grass) ═══
MODELS vs MARKET
Book: Cobolli 1.40 (67.4%) | Fery 2.90 (34.5%) — Betfair ML 1.40 / 3.45
ELO: 50.9% A | ML: 47.4% A | Blend: 47.9% A
Value flag: Fery (B), edge 19.6% — BOTH models disagree with market → needs qualitative validation
RECENT FORM
Cobolli (5 most recent):
07-06 W De Minaur (1907, Elite) — straight sets w/ TB → GRIND vs Elite (real edge)
07-04 W Khachanov (1788, Strong) — 5 sets → GRIND
07-02 W Duckworth (1571, Avg) — 4 sets, 2 TBs → GRIND
07-01 W Navone (1441, Weak) — 4 sets, 2 TBs → GRIND
06-27 L Etcheverry (exhibition, Hurlingham)
Context: French Open FINALIST (lost to Zverev in 5, 06-07); beat FAA, Medvedev, De Minaur in last 6 weeks. Elite opponent quality.
Fery (5 most recent):
07-06 W Dimitrov (1834, Strong) — 5 sets → GRIND
07-04 W Bergs (1770, Strong) — 5 sets, 2 TBs → GRIND
07-02 W Virtanen (1839, Strong) — 4 sets w/ TB → GRIND
06-30 W Dzumhur (1440, Weak) — 4 sets → DOM vs Weak
06-24 L J.M. Cerundolo (1396, WEAK) — Eastbourne
Context: career fortnight, but pre-Wimbledon grass = losses to 1396 (Cerundolo) and 1598 (Bu). All Slam wins are max-effort grinds.
FATIGUE
Cobolli: 5 matches/14d, 15 sets, 808 min, b2b 1
Fery: 6 matches/14d, 18 sets, 942 min, b2b 1
→ 134 min court-time gap: Fery clearly heavier load (two 5-setters + long 4-setters). Edge Cobolli.
NARRATIVE
Let-down: Cobolli lost RG Final 06-07 (31d ago) → no letdown window. Fery: none.
Venue: both 4-0 at this Wimbledon. Title recency: none live.
ALT MARKETS (Betfair): AH Cobolli -4.5 @1.86 / Fery +4.5 @2.06; Total 38.0 O 2.00 / U 1.95. Simulation probs null → no model anchor. Both players grind (Over 38 is tempting: avg court time 161/157 min), but a fatigued Fery fading late is equally plausible → no independent margin conviction. Stay ML.
VERDICT: PASS
The model-market gap is real on paper, but the qualitative read invalidates it: Fery's ELO/ML numbers are inflated by a hot fortnight — three weeks ago he was losing to sub-1600 opponents on this surface, and he now carries the heaviest workload of the QF field. Cobolli is the classier player (RG finalist, straight-set win over Elite De Minaur) and better rested. My true number on Cobolli is ~70% vs 67.4% implied — under 3 points, below the bet bar. No stakeable edge in any market.