TennisEdge's AI model finds no betting value in L. Stefanini vs C. Osorio (Toronto on 2026-08-02) — the market is efficient, with C. Osorio favoured at 1.50.
Independent estimates of who wins, each from a different method. The blend is what the bot prices against the market.
| Method | L. Stefanini | C. Osorio |
|---|---|---|
| Elo ratings-based, surface-weighted |
30.4% | 69.6% |
| Machine learning 29-feature gradient-boosted model |
46.5% | 53.5% |
| Serve / return point-by-point simulation |
38.6% | 61.4% |
| Blend what the bot actually prices off |
30.4% | 69.6% |
| Market implied de-vigged from the opening odds |
36.6% | 63.4% |
We publish how well these score against the closing market, including where they lose: how accurate are tennis predictions?
| Player | Surface | Elo | Record | Matches |
|---|---|---|---|---|
| L. Stefanini | hard | 1644 | 116–80 | 196 |
| C. Osorio | hard | 1807 | 72–65 | 137 |
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. 2, 2026.
C. Osorio leads 1–0 over 1 meeting. Full record.
| Date | Event | Surface | Winner | Score |
|---|---|---|---|---|
| Aug. 2, 2026 | Toronto | hard | C. Osorio | 2 - 0 |
| L | C. Osorio | 0 - 2 | Aug. 2, 2026 |
| W | S. Kenin | 1 - 2 | Aug. 1, 2026 |
| L | Y. Ku | 0 - 2 | July 25, 2026 |
| W | M. Sawangkaew | 2 - 0 | July 24, 2026 |
| W | A. Akli | 0 - 2 | July 23, 2026 |
| W | L. Stefanini | 0 - 2 | Aug. 2, 2026 |
| L | A. Zakharova | 0 - 2 | July 29, 2026 |
| W | S. Hunter | 2 - 0 | July 28, 2026 |
| L | L. Noskova | 1 - 2 | July 2, 2026 |
| W | S. Waltert | 2 - 0 | June 30, 2026 |
All Aug. 2, 2026 results & picks →
═══ L. Stefanini (A) vs C. Osorio (B) — WTA Toronto, hard, 1/64-finals ═══
MODELS vs MARKET (player A = Stefanini)
ELO A 30.4% | ML A 46.5% | Market implied A 36.6% (odds 2.6 / 1.5)
Backend value_pick B (Osorio) +6.2% — but that edge is ELO-only; ML flatly
contradicts it (ML has Osorio at just 53.5%, BELOW market's 63.4%). Models
point in OPPOSITE directions on who holds value → no clean signal.
No exchange price: alt_markets/betfair_ml are null → any ML edge must clear
the 8% book-priced bar, which neither side does off a trustworthy number.
RANKING/ELO SANITY
Osorio rank 83, ELO ~70% fav; Stefanini rank 154, ELO 30% → ELO roughly
tracks rank, credible in the abstract. BUT Osorio's ELO is clay-weighted
(FO wins over Alexandrova 1734 / Putintseva 1731) and stale on hard given
her current hard form (see path). Stefanini's ML is the classic
LEVEL-DROPPER inflation: her win prob is fattened by a W100 streak vs
1500–1710 fields, not tour-level competition.
RECENT FORM (last 5 each, most recent first)
Stefanini: 08-01 Toronto Final W vs Kenin(1649) 3 sets GRIND; 07-25 L vs
Ku(1500); 07-24 W vs Sawangkaew(1500) DOM; 07-23 W vs Akli(1500) GRIND
(6-7 in a set); 07-22 W vs Mandlik(1500). Hot streak but almost entirely
Weak/Avg opposition — inflated by weak fields.
Osorio: 07-29 Memphis QUALIFYING L vs Zakharova(1500) 4-6 3-6; 07-28 Memphis
quali W vs Hunter(1500) 7-5 7-5 GRIND; 07-02 Wimb L vs Noskova(1855) 3s;
06-30 W vs Waltert(1403) DOM; 06-20 Eastbourne SF L vs Tomljanovic(1712).
SHAPE & PATH (Step 1b)
Stefanini: STABLE-but-inflated. Extremely match-sharp (title run culminating
in a 3-set final yesterday), but the level is W100 and today is a sharp
class step-up. Facing rust: none — she is played-in.
Osorio: FADING on hard — a rank-83 player forced to QUALIFY at Memphis and
losing the quali 1/8 to a 1500 in straights (4-6 3-6) is a genuine red
flag on current hard-court level. Her class edge is real but her live
hard form does not look like a 70% favourite.
FATIGUE (Step 2)
Stefanini WON a 3-set match YESTERDAY (08-01 final) → calibrated 3-set
hangover −4.5pp haircut applies (age 28 = baseline, no modulation). 366
court-min/14d is context only. Adjusted ML A ≈ 42%.
Osorio: last match 07-29 (4 days rest), no extreme spot → neutral.
NARRATIVE (Step 3)
LET-DOWN: Stefanini won a Final ~1 day ago → LETDOWN spot (fades A).
Surface: both last on hard (Osorio Memphis, Stefanini Toronto) — no
transition. Fatigue+letdown both push AGAINST Stefanini, i.e. toward the
market's existing lean, so they add no NEW numeric edge.
GATES
G1 edge-origin: FAIL — the only positive edge (ELO +6.2% on Osorio) is
below the 8% book-priced bar, has NO exchange price to execute against,
and is directly contradicted by ML. Stripping fatigue/letdown there is no
5%+ model-vs-market edge that both a trustworthy model and the executable
price support. No numeric edge to stake.
G2 consensus: stake-tier only (never blocks) — models split, would be
lone_signal/contaminated tier at best.
G3 fatal-risk: n/a (ML market; no margin bet).
VERDICT: PASS. Models disagree on direction (ELO→Osorio, ML→Stefanini), the
lone positive number is an ELO-only 6.2% below the book-priced threshold with
no exchange liquidity, and the qualitative picture is muddy — a red-hot but
weak-field, hungover-and-let-down level-dropper against a higher-class but
poor-hard-form favourite. Nothing here is priced wrongly enough to stake.
alt_markets: none at analysis time