TennisEdge's AI model finds no betting value in J. Munar vs R. Hijikata (Montreal on 2026-08-03) — the market is efficient, with J. Munar favoured at 1.72.
Independent estimates of who wins, each from a different method. The blend is what the bot prices against the market.
| Method | J. Munar | R. Hijikata |
|---|---|---|
| Elo ratings-based, surface-weighted |
61.2% | 38.8% |
| Machine learning 29-feature gradient-boosted model |
41.0% | 59.0% |
| Serve / return point-by-point simulation |
53.8% | 46.2% |
| Blend what the bot actually prices off |
61.2% | 38.8% |
| Market implied de-vigged from the opening odds |
55.0% | 45.0% |
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. Munar | clay | 1693 | 115–79 | 194 |
| R. Hijikata | hard | 1725 | 183–115 | 298 |
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 13, 2026.
J. Munar leads 2–0 over 2 meetings. Full record.
| Date | Event | Surface | Winner | Score |
|---|---|---|---|---|
| Aug. 2, 2026 | Montreal | hard | J. Munar | 2 - 0 |
| May 9, 2024 | ATP Rome | clay | J. Munar | 2 - 0 |
| W | R. Hijikata | 2 - 0 | Aug. 2, 2026 |
| L | D. Stricker | 2 - 0 | July 13, 2026 |
| L | J. Lehecka | 3 - 1 | July 4, 2026 |
| W | J. Fearnley | 0 - 3 | July 2, 2026 |
| W | F. Cerundolo | 3 - 0 | June 30, 2026 |
| L | J. Munar | 2 - 0 | Aug. 2, 2026 |
| L | D. Shapovalov | 2 - 1 | July 29, 2026 |
| L | J. De Jong | 3 - 2 | June 30, 2026 |
| L | U. Humbert | 2 - 0 | June 19, 2026 |
| W | J. Lehecka | 2 - 1 | June 18, 2026 |
All Aug. 3, 2026 results & picks →
J. Munar (A, 29, #38) vs R. Hijikata (B, 25, #96) - Montreal, hard, 1/64-finals
MODELS vs MARKET
ELO: Munar 61.2% (high - likely overstates Munar on hard)
ML: Munar 41.0% (lone outlier, favors Hijikata)
Sim: Munar 53.8% / Hijikata 46.2% (point-by-point, surface=hard)
Betfair ML: Munar 1.80 (55.6%) / Hijikata 2.14 (46.7%)
Model split is stark, BUT the hard-court SIM converges with the market:
sim Hijikata 46.2% vs market implied 46.7% - essentially identical. ELO is
the high dissenter, ML the low dissenter; the sim sits with the price.
RECENT FORM (last 5 each)
Munar (A):
07-13 Gstaad clay: LOST Stricker(1537) 6-2 6-4 - UNDER-perf loss to Weak
07-04 Wimb grass: LOST Lehecka(1826) 4s - competitive vs Elite
07-02 Wimb grass: WON Fearnley(1747) GRIND - over Strong
06-30 Wimb grass: WON Cerundolo(1736) 6-1 6-4 6-3 DOM - over Strong
06-22 Eastbourne grass: LOST Bergs(1586) 6-2 6-4
NOTE: ZERO hard-court matches in 90d window (clay/grass only). Clay specialist.
Hijikata (B):
07-29 Los Cabos HARD: LOST Shapovalov(1798) 2-6 6-3 6-2 - competitive, took a set
06-30 Wimb grass: LOST De Jong(1474) 5s - under-perf loss to Weak
06-19 London grass QF: LOST Humbert(1693) 6-1 6-2 - blowout loss
06-18 London 1/8: WON Lehecka(1819) GRIND - over Elite
06-14 London Final grass: WON Giron(1720) - TITLE (grass, ~49d ago, stale)
STEP 1b SHAPE / PATH
Munar: FADING+STALE - last outing straight-set loss to Weak(1537) on clay,
20 days ago. First match at Montreal = rust; 20d off + clay->hard transition.
Hijikata: mixed/FADING off grass-title peak (Humbert blowout, De Jong,
Shapovalov losses) but match-sharp on HARD (played 07-29, took a set off
Shapovalov). On-surface and match-active vs Munar's rust.
FATIGUE (Step 2)
Munar well-rested (0 matches 14d) - but that IS the rust side.
Hijikata 1 match/7d, 4 days ago, 122min 3-setter - not extreme; played 4d ago
(not yesterday) so no 3-set haircut. Roughly neutral.
NARRATIVE (Step 3)
Surface transition: Munar clay/grass -> hard = TRANSITION (weakest surface,
depresses level, widens margins). Hijikata already on hard = no transition.
Title: Hijikata London grass 06-14 (~49d, stale, wrong surface) - no carry.
No recent Final let-down in window.
ELO/RANK SANITY
ELO (61%) agrees with rank direction (#38 vs #96) but both are SURFACE-BLIND.
Munar's overall ELO is clay-inflated, ignores zero hard reps + rust + transition.
The sim (hard-specific) correctly discounts to 53.8%.
EDGE CHECK (all markets, commission-adj, one coherent sim distribution)
Munar ML @1.80: eff 1.76, breakeven 56.8%; sim 53.8% -> NEGATIVE.
Hijikata ML @2.14: eff 2.083, breakeven 48.0%; sim 46.2% -> NEGATIVE (sim==market).
Handicap A -1.5 @1.96: be 51.0%; sim A covers 49.5% -> neg. B +1.5 @1.95: be 51.3%; sim 50.5% -> neg.
Total 22.5: over 2.02 be 50.8% vs sim 48.8% -> neg; under 1.93 be 53.1% vs sim 51.2% -> neg.
AH shadow: N/A (no dog ML staked >=3.0).
GATES
G1 edge-origin: FAIL - no market clears the 5% bar off the coherent sim
distribution. The lone Hijikata/ML edge evaporates when the surface sim is
the anchor; it converges with the market.
G2 consensus: n/a (stake tier only; no qualifying edge)
G3 fatal-risk: n/a
VERDICT: PASS. Genuine model split, but the hard-court simulation resolves it in
the market's favour - sim ~= Betfair on every line. ML is a lone outlier
contradicted by the sim; ELO is clay-inflated. Munar's rust+transition risk and
Hijikata's match-sharpness roughly cancel into the ~coin-flip already priced.
No bettable edge in ML, handicap, or total.