TennisEdge's AI model finds no betting value in Z. Svajda vs D. Shapovalov (Montreal on 2026-08-03) — the market is efficient, with D. Shapovalov favoured at 1.59.
Independent estimates of who wins, each from a different method. The blend is what the bot prices against the market.
| Method | Z. Svajda | D. Shapovalov |
|---|---|---|
| Elo ratings-based, surface-weighted |
35.5% | 64.5% |
| Machine learning 29-feature gradient-boosted model |
38.8% | 61.2% |
| Serve / return point-by-point simulation |
34.5% | 65.5% |
| Blend what the bot actually prices off |
37.5% | 62.5% |
| Market implied de-vigged from the opening odds |
40.4% | 59.6% |
We publish how well these score against the closing market, including where they lose: how accurate are tennis predictions?
| Player | Surface | Elo | Record | Matches |
|---|---|---|---|---|
| Z. Svajda | hard | 1741 | 147–102 | 249 |
| D. Shapovalov | hard | 1830 | 102–78 | 180 |
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 27, 2026.
| L | A. Vukic | 2 - 0 | July 27, 2026 |
| L | C. Hewitt | 1 - 1 | July 26, 2026 |
| W | S. Kozlov | 2 - 0 | July 25, 2026 |
| L | A. De Minaur | 3 - 1 | July 4, 2026 |
| W | K. Majchrzak | 3 - 2 | July 2, 2026 |
| L | A. Gea | 2 - 0 | Aug. 2, 2026 |
| W | C. Norrie | 1 - 2 | Aug. 1, 2026 |
| W | D. Svrcina | 2 - 0 | July 31, 2026 |
| W | R. Pacheco Mendez | 2 - 0 | July 30, 2026 |
| W | R. Hijikata | 2 - 1 | July 29, 2026 |
All Aug. 3, 2026 results & picks →
Z. Svajda (A, rk90, 23) vs D. Shapovalov (B, rk37, 27) — Montreal R1, hard.
MODELS vs MARKET
ELO A 0.355 | ML A 0.388 | Sim A 0.375 (=> B 0.645/0.612/0.625). Pinnacle no-vig A 0.442.
Betfair ML: A 2.24 (comm-adj impl 0.459) | B 1.76 (comm-adj impl 0.581).
All three models favour B MORE than the market does => B carries a nominal exchange edge:
ELO +6.5%, Sim +4.4%, ML +3.2%. That edge is the ONLY numeric disagreement here — and it
points straight at the exhausted side.
RECENT FORM (last 5 shown)
Svajda (A): 07-27 L Vukic 6-3 6-1 (Weak-Avg, blown out); 07-26 L Washington Final Hewitt
6-1 6-7 0-2; 07-25 W Kozlov 7-5 6-1; grass 07-04 L De Minaur; 07-02 W Majchrzak (5set GRIND).
Shape: FADING on hard — last hard match a 6-3 6-1 loss to a 1733. 6 days off since 07-27,
not at Montreal this week (mild rust, but rested). matches_14d 3.
Shapovalov (B): 08-02 L Los Cabos FINAL Gea 6-3 6-4 (TODAY); 08-01 W SF Norrie (3set);
07-31 W Svrcina 6-4 6-2; 07-30 W Pacheco 6-3 6-2; 07-29 W Hijikata (3set). Match-sharp,
deep title run, but see fatigue.
FATIGUE (decisive)
Shapovalov: 5 matches in 5 days, back_to_back 3, 428 court-min, LOST the Los Cabos final
earlier TODAY, then Mexico->Montreal (Canada) cross-country travel for R1. This is an
extreme-spot fatigue side (3rd+ consecutive match day, final same day). Age 27 = baseline.
He also WON a 3-set SF yesterday (08-01) => calibrated -4.5pp hangover haircut applies.
Applying it: Sim B 0.625 -> 0.580 => exchange edge ~0.0 (before the extra travel/final toll).
Svajda: rested 6 days, clearly the fresher man. The fatigue differential is entirely on B.
NARRATIVE
Shapovalov lost Los Cabos final 0d ago (NOTE, not a title/letdown-win). Defending champ at
Los Cabos but did not retain (press: "seeking 5th title" — lost to Gea). No surface
transition (both on hard). ELO~rank agreement both sides, ratings credible.
MARKET CHOICE
ML: B nominal +4-6% pre-fatigue collapses to ~0 after the hangover haircut and does not
survive the extreme travel/final-today spot — cannot back an exhausted favourite.
A side: sim 0.375, even with full fatigue benefit (+0.045) = 0.420 < 0.459 executable
=> negative edge; any A edge is 100% fatigue-manufactured (G1 fail).
Handicap: B -1.5 sim cover 0.601 vs 0.521 = +8.0%, BUT it is a self-negating fatigue cover
(asking the just-played-a-final side to win comfortably) => G3 hard veto for margin bets.
A +1.5 @1.99 negative. AH: B -1.5 @ 1.92 (pass, self-negating).
Totals: Over 0.483 vs 0.495 neg; Under 0.517 vs 0.521 neg. No edge.
alt_markets present; best expressions all fail their bar.
GATES
G1 edge-origin: FAIL — the only pre-narrative exchange edge (B +4-6%) evaporates once the
-4.5pp hangover is applied and belongs to the extreme-fatigue side; the A-side edge is
fatigue-manufactured and negative.
G2 consensus: PASS (stake-tier only; 3-model lean on B but unusable).
G3 fatal-risk: FAIL (margin) — B -1.5 cover thesis is self-negated by B's fatigue.
VERDICT: PASS. Models like Shapovalov, but he just lost a title final today after 5 matches
in 5 days and travels Mexico->Canada; his nominal edge is a fatigue trap and the fresher
Svajda's only edge is fatigue-manufactured. No clean, non-fatigue numeric edge => no bet.