TennisEdge's AI model finds no betting value in J. M. Cerundolo vs H. Medjedovic (Montreal on 2026-08-03) — the market is efficient, with H. Medjedovic favoured at 1.50.
Independent estimates of who wins, each from a different method. The blend is what the bot prices against the market.
| Method | J. M. Cerundolo | H. Medjedovic |
|---|---|---|
| Elo ratings-based, surface-weighted |
40.6% | 59.4% |
| Machine learning 29-feature gradient-boosted model |
51.4% | 48.6% |
| Serve / return point-by-point simulation |
35.6% | 64.4% |
| Blend what the bot actually prices off |
40.8% | 59.2% |
| Market implied de-vigged from the opening odds |
36.4% | 63.6% |
We publish how well these score against the closing market, including where they lose: how accurate are tennis predictions?
| Player | Surface | Elo | Record | Matches |
|---|---|---|---|---|
| J. M. Cerundolo | clay | 1838 | 205–111 | 316 |
| H. Medjedovic | clay | 1875 | 100–43 | 143 |
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 21, 2026.
H. Medjedovic leads 2–0 over 2 meetings. Full record.
| Date | Event | Surface | Winner | Score |
|---|---|---|---|---|
| May 28, 2025 | French Open | clay | H. Medjedovic | 3 - 0 |
| May 24, 2023 | ATP French Open | clay | H. Medjedovic | 2 - 1 |
| L | M. Trungelliti | 0 - 2 | July 21, 2026 |
| L | R. Collignon | 2 - 1 | July 18, 2026 |
| W | C. Ruud | 2 - 1 | July 17, 2026 |
| W | M. Kecmanovic | 2 - 1 | July 16, 2026 |
| W | Z. Kolar | 2 - 1 | July 14, 2026 |
| L | S. Ofner | 2 - 3 | June 29, 2026 |
| L | U. Humbert | 1 - 2 | June 19, 2026 |
| W | A. Rinderknech | 0 - 2 | June 16, 2026 |
| L | C. Ruud | 3 - 0 | May 27, 2026 |
| W | Y. Hanfmann | 3 - 1 | May 24, 2026 |
All Aug. 3, 2026 results & picks →
J. M. Cerundolo (A) vs H. Medjedovic (B) — Montreal, hard, R1 (1/64)
MODELS vs MARKET (prob A wins)
ELO 40.6% | ML 51.4% | Sim 40.8% | Pinnacle no-vig 40.3% | Betfair implied ~41% (2.44)
Three sharp sources (ELO, point-by-point sim, Pinnacle) AND the exchange all converge ~40-41%.
ML at 51.4% is the LONE outlier — and it is the surface-unaware model, hitting its textbook
failure mode: a clay specialist transitioning to hard.
RECENT FORM
Cerundolo (A): last 5 all CLAY —
07-21 Kitzbuhel L Trungelliti(1732) 2-6 1-6 (blown out, DOM loss)
07-18 Gstaad SF L Collignon(1849) 1-6 7-6 7-5
07-17 Gstaad QF W Ruud(2015) 3-6 7-5 6-2 (big win vs Elite, on clay)
07-16 Gstaad W Kecmanovic(1786) 3-6 6-3 7-5
07-14 Gstaad W Kolar(1614) 4-6 6-3 6-4
Shape: FADING at end of clay swing (SF loss then 2-6 1-6 destruction). Clay->hard TRANSITION.
Medjedovic (B): last match 06-29 Wimbledon (L Ofner, 5 sets) — 34-DAY LAYOFF, rust risk.
Recent grass form weak (L Ofner, L Humbert). Hard court is his stronger surface (big serve).
PATH / SHAPE
Cerundolo: clay-courter off a bad loss, 12 days idle, now on hard — level-depressing transition,
looser margins. Not a back.
Medjedovic: rust risk from 34-day gap, but surface fit + serve suit hard; market ~60% is fair.
FATIGUE
Neither has an extreme spot. A rested 12d, B rested 34d (rusty, not tired). No haircut applies.
No fatigue edge either way.
NARRATIVE
TRANSITION on A (clay->hard) = risk on A, not support. No let-down/title/venue signals of weight.
ALT MARKETS
Handicap A +2.5 @1.79 (sim covers 0.498 -> edge -6%); B -2.5 @1.90 (covers 0.502 -> edge -2%).
Totals U22.5 @1.87 (sim 0.5378 -> edge +0.3%, negligible); O22.5 @1.86 (0.4622 -> negative).
No stakeable alt edge.
MARKET CHOICE
alt_markets present. Best expression evaluated: none clears the 5% bar.
ML A: prob ~41% vs Betfair 2.44 (comm-adj 2.368, need 42.2%) -> edge NEGATIVE.
ML B: prob ~59% vs Betfair 1.68 (comm-adj 1.646, need 60.7%) -> edge NEGATIVE.
AH: A +2.5 @ 1.79 (edge -6%, pass).
No refutation check run (PASS candidate — no web access for passes).
GATES
G1 edge-origin: FAIL — only ML (surface-unaware, lone outlier) beats the market; the three
convergent sharp sources and the exchange price agree ~41%. No executable edge pre- or
post-narrative; the qualitative read (clay specialist on hard, fading) REFUTES the ML optimism.
G2 consensus: stake-tier only — would be lone_signal/capped even if a bet, but G1 blocks.
G3 fatal-risk: n/a (no margin bet).
VERDICT: PASS. Market is efficient — clusters with ELO, sim, Pinnacle. The lone ML edge on
Cerundolo is a surface-transition artifact refuted by the form read, and shrinks below zero at the
executable Betfair price. No side, no alt market, clears the bar.