TennisEdge's AI model finds no betting value in B. Bonzi vs Y. Hanfmann (Montreal on 2026-08-02) — the market is efficient, with Y. Hanfmann favoured at 1.47.
Independent estimates of who wins, each from a different method. The blend is what the bot prices against the market.
| Method | B. Bonzi | Y. Hanfmann |
|---|---|---|
| Elo ratings-based, surface-weighted |
52.0% | 48.0% |
| Machine learning 29-feature gradient-boosted model |
46.8% | 53.2% |
| Serve / return point-by-point simulation |
51.9% | 48.1% |
| Blend what the bot actually prices off |
47.3% | 52.7% |
| Market implied de-vigged from the opening odds |
40.7% | 59.3% |
We publish how well these score against the closing market, including where they lose: how accurate are tennis predictions?
| Player | Surface | Elo | Record | Matches |
|---|---|---|---|---|
| B. Bonzi | hard | 1746 | 163–94 | 257 |
| Y. Hanfmann | clay | 1810 | 123–82 | 205 |
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. 3, 2026.
Y. Hanfmann leads 1–0 over 1 meeting. Full record.
| Date | Event | Surface | Winner | Score |
|---|---|---|---|---|
| Aug. 3, 2026 | Montreal | hard | Y. Hanfmann | 2 - 1 |
| L | Y. Hanfmann | 1 - 2 | Aug. 3, 2026 |
| W | H. Dellien | 2 - 0 | Aug. 1, 2026 |
| L | J. D. Hara Friend | 1 - 2 | July 30, 2026 |
| W | M. Mmoh | 2 - 0 | July 29, 2026 |
| W | B. Bicknell | 2 - 1 | July 28, 2026 |
| W | B. Bonzi | 1 - 2 | Aug. 3, 2026 |
| L | Q. Halys | 0 - 2 | July 24, 2026 |
| W | S. Baez | 2 - 0 | July 23, 2026 |
| W | M. Trungelliti | 0 - 2 | July 22, 2026 |
| W | L. Neumayer | 0 - 2 | July 20, 2026 |
All Aug. 2, 2026 results & picks →
Bonzi (A) vs Hanfmann (B) — ATP Montreal R64, hard MODELS vs MARKET ELO A 0.520 | ML A 0.468 | sim A 0.473 — split, straddling market. Betfair ML A 2.08 (exec imp 0.494) / B 1.90 (exec imp 0.539). Pinnacle no-vig A 0.466. ELO/rank divergence: Bonzi ELO-favourite yet ranked 101 vs Hanfmann 59. Bonzi's ELO built on Challenger fields (Dellien 1412, Bicknell 1537) — FARMED-ELO flag, discount that edge. ML/sim both put A below market, so consensus sits with the price. RECENT FORM (last 5) Bonzi (A), hard: Won Montreal (Challenger) Final 08-01 vs Dellien 1412 7-5 6-2 (DOM, weak); Lost Vancouver QF vs Hara Friend 1644 7-5 6-7 5-7 (grind loss); Won R16 vs Mmoh 1716 7-5 6-4 (DOM over Strong — good); Won R32 vs Bicknell 1537 3-6 6-3 6-4 (grind). Match-sharp on hard, mildly SHARPENING but field is Challenger-level. Hanfmann (B), all clay: Lost Kitzbuhel SF vs Halys 1796 4-6 6-7; Won QF vs Baez 1789 6-3 6-1 (DOM over Strong — over-performing); Won R16 vs Trungelliti; Won R32 vs Neumayer. Solid clay run to SF, rank 59 credible. PATH / SHAPE Bonzi: 4 matches/7d, won a title YESTERDAY, playing-in and on-surface — match-sharp. Hanfmann: 9 days off + surface TRANSITION clay->hard (level-depressing, wider margins). Rest vs rust roughly cancels the fatigue-load story. FATIGUE (Step 2) Bonzi: heavy load (501min/14d, 4 matches/7d, back_to_back 2) but yesterday's title was a 2-SET win (7-5 6-2) -> NO 3-set hangover haircut (~neutral). Age 30 -> slow recovery, harden one tier, but this is visible fatigue the market already prices (Betfair has him only ~48%). Not an extreme spot today. Hanfmann: rested (0 matches/7d) but off-surface. NARRATIVE LETDOWN: Bonzi won Final 1d ago -> letdown spot (predicts who wins, not margin; cannot create edge). TRANSITION: Hanfmann clay->hard. GATES G1 edge-origin: FAIL — no side reaches the 5% exchange edge pre-narrative. A best = ELO +2.6% (and that ELO is Challenger-farmed); ML/sim A negative; B all negative; handicaps tiny/negative; Under only +1.8% (sim line 22.5 vs 23.0 quoted, no margin thesis). G2 consensus: n/a (stake-tier only). G3 fatal-risk: n/a (no margin bet). VERDICT: PASS. Market efficient — models straddle the price, Bonzi's only edge (ELO) is farmed on weak Challenger fields, and no market clears the 5% executable-price bar. Fatigue/letdown/transition are real but self-cancelling (Bonzi tired vs Hanfmann rusty+off-surface) and cannot manufacture a numeric edge. MARKET CHOICE ML: no side >= 5% exec edge -> pass. Handicap A -0.5 @1.97 edge +0.6%; B +0.5 @1.94 edge -5.5% -> pass. Total Under 23.0 @2.06 edge ~+1.8% (marginal, sim line differs) -> pass. AH: B +0.5 @ 1.94 (only line quoted).