TennisEdge's AI model finds no betting value in K. Majchrzak vs G. Monfils (Montreal on 2026-08-03) — the market is efficient, with K. Majchrzak favoured at 1.41.
Independent estimates of who wins, each from a different method. The blend is what the bot prices against the market.
| Method | K. Majchrzak | G. Monfils |
|---|---|---|
| Elo ratings-based, surface-weighted |
44.7% | 55.3% |
| Machine learning 29-feature gradient-boosted model |
73.6% | 26.4% |
| Serve / return point-by-point simulation |
51.5% | 48.5% |
| Blend what the bot actually prices off |
44.7% | 55.3% |
| Market implied de-vigged from the opening odds |
67.8% | 32.2% |
We publish how well these score against the closing market, including where they lose: how accurate are tennis predictions?
| Player | Surface | Elo | Record | Matches |
|---|---|---|---|---|
| K. Majchrzak | hard | 1740 | 93–57 | 150 |
| G. Monfils | hard | 1806 | 92–54 | 146 |
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 Aug. 4, 2026.
G. Monfils leads 1–0 over 1 meeting. Full record.
| Date | Event | Surface | Winner | Score |
|---|---|---|---|---|
| Aug. 4, 2026 | Montreal | hard | G. Monfils | 2 - 0 |
| L | G. Monfils | 0 - 2 | Aug. 4, 2026 |
| L | T. Fritz | 2 - 0 | July 29, 2026 |
| W | T. Paul | 2 - 0 | July 28, 2026 |
| L | Z. Svajda | 3 - 2 | July 2, 2026 |
| W | A. Tabilo | 3 - 0 | June 30, 2026 |
| W | K. Majchrzak | 0 - 2 | Aug. 4, 2026 |
| L | H. Gaston | 3 - 2 | May 25, 2026 |
| L | C. Ugo Carabelli | 2 - 0 | April 23, 2026 |
| L | A. Bublik | 0 - 2 | April 7, 2026 |
| W | T. Griekspoor | 2 - 1 | April 5, 2026 |
All Aug. 3, 2026 results & picks →
K. Majchrzak (A, 30, #59) vs G. Monfils (B, 39, #68) — Montreal R1, hard.
MODELS
ELO A 44.7% | Sim A 51.5% (exp margin 0.5g) | ML A 73.6% | Market A ~73-76% (Betfair 1.32, Pinnacle no-vig 72.9%).
Three camps. ELO + sim make A a coin-flip/underdog; ML + sharp market make A a heavy favourite. The models that
disagree with the market both rate Monfils on STALE inputs: sim SPW 0.595/hold 0.825 over 42 matches, ELO on his
peak-career rating. Neither knows Monfils (39) has played ONE match since 25 May — a 2+ month layoff.
RECENT FORM
Majchrzak (A) — match-sharp, in form. Washington last week: beat T. Paul (1914) 7-5 7-6, lost to Fritz (1934) 6-3 6-4
(competitive vs elite). Won Hertogenbosch grass title in June (beat De Minaur in final, Medvedev in semi). Last match
29 Jul → 4 days rest. SHARPENING/STABLE, well-rested.
Monfils (B) — one match in window: French Open 25 May, 5-set loss to Gaston (1584) on clay. RUST RISK: >21d layoff
(>2 months), age 39, off-surface + thin sample. Sim/ELO inputs are contaminated by his healthy era.
FATIGUE
A: 2 matches/7d but last was 29 Jul (4d rest) — no extreme spot, well-rested. Age 30 = baseline.
B: "well_rested" flag is really RUST, not freshness — 2+ month gap. Not a fatigue edge for B; it's a level depressant.
NARRATIVE
Surface: A on hard (Washington) — no transition. B last played clay 2+ months ago — rust + off-surface.
No let-down/title-recency signals in play.
MARKET CHOICE
ML A: Betfair 1.32, eff odds 1.304 (implied 76.7%). My A estimate ~73-75% (with ML/market vs stale ELO/sim).
Edge = 0.73 - 0.767 = negative. No value on A. min 1.44 not met.
ML B: Betfair 3.65 (implied 27.4%). ELO's 55.3% is stale; true Monfils ~24-27% given rust/age. No edge. Backing the
rusty side against the market's information = trap.
Handicap A -3.5 @1.64: sim covers 37% — no. B +3.5 @2.02: sim covers 62.9% → nominal 13% edge, BUT built on
contaminated Monfils inputs; sim wrongly thinks match is 50/50. A rusty 39yo blown out widens the margin AGAINST
the cover — the +3.5 thesis is self-negating. ±0.05 SPW clamp can't capture a 2mo layoff, and any correction moves
the cover DOWN, not up. AH: B +3.5 @ 2.02
Totals 21.5: over @1.8 eff implied 0.568 vs sim 0.537 (neg); under @1.85 eff implied 0.553 vs sim 0.463 (neg). No edge.
refutation_check: not run (PASS decision — no bet/lean candidate clears the gates).
GATES (candidate B +3.5 handicap, the only nominal edge)
G1 edge-origin: PASS — 13% nominal edge exists numerically from sim vs price BEFORE narrative.
G2 consensus: stake-tier only — would be contaminated (stale/off-surface Monfils inputs) → capped tier at best.
G3 fatal-risk: FAIL — margin bet; cover thesis rests on sim treating Monfils as competitive, but the fatal risk
(39yo, 2+ month layoff, run over → margin widens) directly negates the +3.5 cover. Hard veto.
VERDICT: PASS. The only apparent edges (B ML value 23%, B +3.5 cover 13%) both originate from models that are blind
to Monfils' 2+ month layoff. The sharp market and ML agree Majchrzak ~73%; A's price offers no value and the B/margin
edges are stale-input mirages. Discipline says pass.