TennisEdge's AI model finds no betting value in J. Faria vs C. O'Connell (Montreal on 2026-08-01) — the market is efficient, with J. Faria 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 | J. Faria | C. O'Connell |
|---|---|---|
| Elo ratings-based, surface-weighted |
44.0% | 56.0% |
| Machine learning 29-feature gradient-boosted model |
58.9% | 41.1% |
| Serve / return point-by-point simulation |
49.1% | 50.9% |
| Blend what the bot actually prices off |
55.0% | 45.0% |
| Market implied de-vigged from the opening odds |
64.5% | 35.5% |
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. Faria | clay | 1796 | 92–74 | 166 |
| C. O'Connell | hard | 1700 | 131–96 | 227 |
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 24, 2026.
C. O'Connell leads 1–0 over 1 meeting. Full record.
| Date | Event | Surface | Winner | Score |
|---|---|---|---|---|
| Aug. 1, 2026 | Montreal | hard | C. O'Connell | 2 - 1 |
| L | C. O'Connell | 1 - 2 | Aug. 1, 2026 |
| L | L. Darderi | 1 - 2 | July 24, 2026 |
| W | G. Bueno | 2 - 0 | July 23, 2026 |
| W | B. Van De Zandschulp | 0 - 2 | July 21, 2026 |
| L | C. Ruud | 1 - 2 | July 16, 2026 |
| L | J. Duckworth | 2 - 0 | Aug. 3, 2026 |
| W | J. Faria | 1 - 2 | Aug. 1, 2026 |
| L | A. Martin | 1 - 2 | July 25, 2026 |
| L | M. McDonald | 3 - 0 | June 25, 2026 |
| W | B. Tomic | 0 - 2 | June 24, 2026 |
All Aug. 1, 2026 results & picks →
J. Faria (A, 22, #119) vs C. O'Connell (B, 32, #150) - Montreal, hard, Final. RE-CHECK of earlier LEAN with fresh Betfair prices now live. MODELS vs MARKET ELO: A 0.44 (favors B/O'Connell); ML: A 0.59 (favors A/Faria); Blend: A 0.55 -> B 0.45. Book: A 1.44 / B 2.62 ; Betfair ML now live: A 1.54 / B 2.66 (A implied ~65%, B ~38%). Models DISAGREE sharply; ML aligns with market, ELO dissents. RECENT FORM (last 5) Faria: L Darderi(1929) 6-2 5-7 2-6 clay; W VdZ(1715) DOM clay; W Bueno(1624) DOM clay; L Ruud(2008) GRIND clay; W Wawrinka(1561) GRIND clay. Good clay shape, DOM wins over Strong. BUT all recent play is CLAY. O'Connell: L A.Martin(1528) 3-6 6-4 2-6 hard (UNDER-perf vs weak); month gap before to Wimbledon(06-25). June grass run good, but current hard-court level shaky. PATH / SHAPE / TRANSITION Neither shows Montreal matches (data gap / both rusty; Faria 8d off, O'Connell 7d off). SURFACE TRANSITION: Faria clay -> hard (timing untested, level depressed). O'Connell on-surface (Washington hard). Supports B qualitatively - but already in a 65% market price. FATIGUE Faria 3 matches_14d / 360min; O'Connell 1 / 131min. No extreme spot for either. Cumulative gap = context only. Roughly neutral. Faria 22 recovers fast; O'Connell 32. ELO/RANKING SANITY (decisive) Faria strong rank (#119) but ELO-DOG (44%) = classic ELO UNDER-statement of a rising 22yo. O'Connell weaker rank (#150) yet ELO-favorite = grass-INFLATED ELO (Nottingham title, beat 1816/1764) contradicted by fresh hard-court loss to a 1528 player. Rule: strong rank / low ELO -> do NOT back the opponent purely on the ELO gap. The B value is exactly that error. NARRATIVE Both off LOSSES in last match. No let-down/title-recency for backed side. Venue neutral. EDGE ORIGIN B value (blend 45% vs Betfair effective ~38.8%) comes ENTIRELY from ELO propping the blend. ML (B 41%) and sharp Betfair (B 38%) say ~1-2% or none. Discount the suspect ELO -> B edge < 5%. A side: models 55-59% but market charges 66% -> A overpriced too. GATES G1 edge-origin: FAIL - staked B edge exists only via ELO-propped blend; strip suspect ELO (rank-divergence rule) and edge < 5%. G2 consensus: stake-tier only - ML+market favor A; only lone suspect ELO supports B. G1 already fails. G3 fatal-risk: N/A (ML market). MARKET CHOICE Available: Betfair ML only. alt_markets.handicap / totals / model = null. AH: not quoted. Best expression would be B ML @2.66; min_odds for B at ~0.41 prob = 2.56 (price clears) but the probability does not hold up once ELO discounted -> fails G1. No refutation check run (PASS gets no web access). VERDICT: PASS. Market (sharpest single estimate) has Faria ~65% and ML agrees; the only thing favoring O'Connell is an ELO that both the rank divergence and his fresh hard-court loss say is unreliable. Surface transition supports B but is already priced. No exchange-mispriced edge -> stand down.