TennisEdge's AI model finds no betting value in P. Carreno-Busta vs V. Royer (Montreal on 2026-08-03) — the market is efficient, with P. Carreno-Busta favoured at 1.44.
Independent estimates of who wins, each from a different method. The blend is what the bot prices against the market.
| Method | P. Carreno-Busta | V. Royer |
|---|---|---|
| Elo ratings-based, surface-weighted |
65.7% | 34.3% |
| Machine learning 29-feature gradient-boosted model |
56.4% | 43.6% |
| Serve / return point-by-point simulation |
55.9% | 44.1% |
| Blend what the bot actually prices off |
65.7% | 34.3% |
| Market implied de-vigged from the opening odds |
65.6% | 34.4% |
We publish how well these score against the closing market, including where they lose: how accurate are tennis predictions?
| Player | Surface | Elo | Record | Matches |
|---|---|---|---|---|
| P. Carreno-Busta | hard | 1742 | 81–51 | 132 |
| V. Royer | clay | 1705 | 144–71 | 215 |
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 Feb. 25, 2026.
| L | L. van Assche | 2 - 1 | July 23, 2026 |
| W | V. Gaubas | 0 - 2 | July 21, 2026 |
| L | C. Ugo Carabelli | 1 - 2 | July 15, 2026 |
| W | M. Dodig | 0 - 2 | July 14, 2026 |
| L | R. Jodar | 3 - 2 | July 2, 2026 |
| L | A. Molcan | 2 - 0 | July 14, 2026 |
| L | T. Droguet | 0 - 2 | July 11, 2026 |
| W | T. Daniel | 2 - 1 | July 10, 2026 |
| W | O. Wallin | 2 - 1 | July 9, 2026 |
| W | M. Neuchrist | 2 - 0 | July 7, 2026 |
All Aug. 3, 2026 results & picks →
P. Carreno-Busta (A) vs V. Royer (B) — Montreal, HARD, 1/64-finals
MODELS vs MARKET
Market: A 1.44 (impl 65.6%) / B 2.75. value_edge ~0.08% (nil).
ELO A 65.7% == market (efficient). ML A 56.4% (lone dissenter, favors B).
ELO/rank sanity: CB ELO-fav yet ranked 91; Royer higher-ranked (71) but ELO dog.
CB's ELO leans on historical hard-court pedigree (ex-top10) — but he is 35 and
declining; his recent clay results are ordinary. Do not over-trust either extreme.
RECENT FORM (last 5, all clay/grass — NEITHER has a recent hard-court match)
CB (A): L van Assche(1706) 6-3 4-6 6-2 clay; W Gaubas(1635) DOM clay;
L Ugo Carabelli(1809) 6-3 5-7 5-7 clay; W Dodig(1640) DOM clay;
L Jodar(1666) 5-set grass. Shape FADING — ugly losses to mid-ELO clay fields.
Royer (B): L Molcan(1786) 6-4 7-5 clay; L Droguet(1783) 3-6 4-6 clay;
W Daniel(1725) GRIND clay; W Wallin(1554) GRIND clay; W Neuchrist(1448) DOM clay.
PATH / SHAPE / SURFACE
BOTH players are transitioning clay -> hard (Step 3 TRANSITION on both sides):
depressed level + wider margin variance for each — no directional edge either way.
CB last match 2026-07-23 (~10d gap); Royer last match 2026-07-14 (~19d layoff =
rust risk). Neither in an extreme fatigue spot (Step 2): cumulative load is context
only, not an edge here.
FATIGUE
No extreme spot. CB matches_14d 2, last 07-23. Royer well_rested, long layoff.
No 3-set-hangover (no next-day win-and-play). Roughly equal, no haircut applied.
EDGE MATH (Pricing Discipline — no exchange market)
Back CB @ 1.44: implied 69.4% vs ELO 65.7% -> edge -3.7%. FAIL (market > model).
Back Royer @ 2.75: implied 36.4% vs ML B 43.6% -> +7.2% BUT book-priced only
(alt_markets null, no betfair_ml) -> bar is >=8%; 7.2% < 8%. And this rests on
ML alone, contradicting ELO+market, with NO independent qualitative validation
(CB hard-court pedigree + Royer rust/transition argue the other way). FAIL.
GATES
G1 edge-origin: FAIL — no side clears the numeric bar at the executable price
(CB negative; Royer 7.2% < 8% book-priced bar, lone-ML, unvalidated).
G2 consensus: PASS (stake-tier only; moot — no bet).
G3 fatal-risk: n/a (ML market; no margin bet).
MARKET CHOICE
alt_markets: none at analysis time
AH: not quoted
ML only available; no side clears its bar.
VERDICT: PASS. ELO agrees with market (efficient), ML is a lone dissenter with no
qualitative backing, no exchange liquidity, edge below the book-priced bar. Both
players cold onto a surface transition. No priced opinion worth staking.