TennisEdge's AI model finds no betting value in C. Osorio vs E. Alexandrova (Toronto on 2026-08-04) — the market is efficient, with E. Alexandrova favoured at 1.89.
Independent estimates of who wins, each from a different method. The blend is what the bot prices against the market.
| Method | C. Osorio | E. Alexandrova |
|---|---|---|
| Elo ratings-based, surface-weighted |
56.7% | 43.3% |
| Machine learning 29-feature gradient-boosted model |
50.3% | 49.7% |
| Serve / return point-by-point simulation |
56.5% | 43.5% |
| Blend what the bot actually prices off |
52.7% | 47.3% |
| Market implied de-vigged from the opening odds |
48.8% | 51.2% |
We publish how well these score against the closing market, including where they lose: how accurate are tennis predictions?
| Player | Surface | Elo | Record | Matches |
|---|---|---|---|---|
| C. Osorio | hard | 1807 | 72–65 | 137 |
| E. Alexandrova | hard | 1760 | 103–77 | 180 |
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 3–0 over 3 meetings. Full record.
| Date | Event | Surface | Winner | Score |
|---|---|---|---|---|
| May 25, 2026 | French Open | clay | C. Osorio | 2 - 0 |
| Oct. 18, 2022 | Guadalajara 2 | hard | C. Osorio | 2 - 0 |
| June 30, 2021 | Wimbledon | grass | C. Osorio | 2 - 0 |
| 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 |
| L | K. Liutova | 1 - 1 | July 28, 2026 |
| L | I. Jovic | 1 - 2 | July 3, 2026 |
| W | L. Tararudee | 2 - 0 | July 1, 2026 |
| W | P. Udvardy | 2 - 0 | June 29, 2026 |
| L | N. Osaka | 2 - 0 | June 25, 2026 |
All Aug. 4, 2026 results & picks →
C. Osorio vs E. Alexandrova — WTA Toronto, hard, 1/32-finals (R2)
═══ MODELS vs MARKET ═══
ELO : Osorio 56.7%
ML : Osorio 50.4% (break-even, no signal)
Composite : Osorio 52.7%
Sim (base): Osorio 56.5%
Betfair ML: A 2.00 / B 1.93 -> devigged Osorio 49.1%
Pinnacle : 1.962 / 1.901, no_vig_prob_a 49.2% (fetched 2026-08-03 22:55Z, ~fresh)
Book quote: A 1.98 / B 1.89 (context only; exchange is the baseline)
Two independent sharp sources agree on 49% for Osorio. Only ELO and the sim
disagree, and both are compromised (see below).
═══ SIM VALIDATION & RE-RUN (logged) ═══
Baseline inputs: Osorio spw 0.580 / rpw 0.434 / hold 62.4%; Alexandrova spw 0.579
/ rpw 0.422 / hold 72.2%. Baseline sim = Osorio 56.5%.
Two input problems:
(1) Internal incoherence — a 58.0% SPW does not sit with a 62.4% career-window
hold; Alexandrova holds 72.2% off an almost identical 57.9% SPW.
(2) Field/surface contamination — Osorio's 50-match window is dominated by clay
and sub-1650 opposition (Rabat, Parma, RG, Memphis quali vs 1500s), which
fattens serve/return; Alexandrova's window is grass + top-level draws against
Sabalenka 1895, Jovic 1899, Osaka 1694, Andreeva 1669, which depresses hers.
Re-run with the one input I can justify: spw_a 0.580 -> 0.555 (only override,
not clamped). Result: Osorio 43.4%, expected margin flips from +1.1 to -0.6.
A 2.5pp serve shade swings the sim 13 points — the sim is input-dominated here,
NOT a robust anchor. The market's 49% sits almost exactly between the baseline
and the corrected run. Conclusion: no side of this sim is usable as an edge.
═══ ELO / RANKING SANITY ═══
Osorio ranked 83 yet ELO-favourite over a world No.10 -> classic "high ELO,
modest rank" farmed-rating pattern (her ELO window is clay/125-level). Discount.
Alexandrova ranked 10 on a 1734 ELO -> "strong rank, low ELO": ELO understates
her. Both divergences point the same way — the ELO edge for Osorio is not real.
═══ RECENT FORM (last 5 each) ═══
OSORIO (24, #83)
08-02 W Stefanini 1653 Strong 7-6 6-3 GRIND (Toronto R1)
07-29 L Zakharova 1500 Avg 4-6 3-6 (straight-sets loss to Avg)
07-28 W Hunter 1500 Avg 7-5 7-5 DOM
07-02 L Noskova 1855 Elite 3-6 6-4 2-6 (competitive vs Elite)
06-30 W Waltert 1403 Weak 6-2 6-1 DOM
ALEXANDROVA (31, #10)
07-28 L Liutova 1500 Avg 6-7 6-4 4-5 (loss to an Avg-level player)
07-03 L Jovic 1899 Elite 3-6 6-3 4-6 (competitive vs Elite)
07-01 W Tararudee 1533 Avg 7-5 7-5 DOM
06-29 W Udvardy 1549 Avg 6-4 6-2 DOM
06-25 L Osaka 1694 Strong 2-6 2-6 (heavy)
SHAPE
Osorio: STABLE — R1 grind past a 1653 is fine, but a straight-sets loss to a
1500 five days earlier is under-performing. Not sharpening.
Alexandrova: FADING — one match in five weeks and she lost it to a 1500
(07-28). Mild rust risk on top: first match at this venue (bye), 7 days idle.
PATH: Osorio is the match-sharp side (3 hard-court matches in 7 days, already
won at this venue). Alexandrova is played-off and untested on the surface swing.
This is a genuine qualitative point for Osorio — but it is exactly the story a
49% price for a #83 vs a #10 already reflects. The market has priced it.
H2H: Osorio won 6-2 6-4 at Roland Garros 05-25 (DOM) — but on CLAY, 10 weeks
ago. Not on-surface evidence for a hard court in Toronto.
═══ FATIGUE ═══
Osorio: 3 matches/7d, 246min, back_to_back 1, last played 08-02 (2 days ago,
straight sets). No extreme spot — not yesterday, not 3-in-3, no 150min+ match.
Age 24 -> recovers fast. Cumulative-minutes gap = context only, never an edge.
No 3-set-hangover haircut applies (her 08-02 win was 2 sets, and not yesterday).
Alexandrova: 1 match/14d, 121min, idle since 07-28. Fresh but rusty; age 31.
Net: roughly equal for betting purposes.
═══ NARRATIVE ═══
Let-down: None (no Final for either player in the window).
Venue momentum: Osorio 1-0 at Toronto; Alexandrova no matches at venue in window.
Title recency: none.
Surface transition: none — both last played on hard.
═══ MARKET CHOICE ═══
ML Osorio @ 2.00 Betfair -> effective 1.95, breakeven 51.3%. Composite 52.7%
-> edge +1.4%. Even the uncorrected sim's 56.5% gives only +5.2%, and that
number dies the moment the serve input is corrected. min_odds = 1.05/0.527
= 1.99 effective, i.e. I would need ~2.05+ raw. FAIL.
ML Alexandrova @ 1.93 -> effective 1.88, breakeven 53.1% vs my ~50%. Negative.
Handicap 0.5 (A -0.5 @1.92 / B +0.5 @1.97): brief model A covers 55.3%, but the
corrected sim has A covering 45.6%. Estimates straddle the 52.1%/50.8%
breakevens — no defensible edge, and a 0.5 line is just the ML anyway.
Totals 21.5 (over 1.91 / under 2.02): sim under 50.8% (baseline) / 51.1%
(corrected) vs 49.5% breakeven -> +1.3-1.6%. Below the 5% bar. Over is negative.
Best available edge across all three markets is under 2% on a coherent
distribution. Nothing to stake.
AH: not applicable (no dog ML >= 3.0 staked).
GATES
G1 edge-origin: FAIL — pre-narrative exchange edge is +1.4% on the composite
(best market <2%), far below the 5% bar; the only numbers that clear it come
from an ELO I have flagged as farmed and a sim that flips 13 points on a
2.5pp serve-input correction.
G2 consensus: PASS (tier-only) — would be capped 0.5-1.0u lone_signal anyway,
since the supporting sim/ELO inputs are contaminated.
G3 fatal-risk: PASS — no margin bet staked; ML would not be blocked regardless.
═══ VERDICT: PASS ═══
Two sharp sources (Betfair devig 49.1%, Pinnacle 49.2%) sit on top of each other
and the ML model (50.4%) agrees with them. The only bullish inputs for Osorio are
an ELO her ranking contradicts and a simulation whose output is dominated by a
serve figure built on clay and 1500-level opposition. The genuine qualitative
read — Osorio match-sharp, Alexandrova rusty and fading — is real, and it is
precisely why a world #83 is priced at coin-flip against a world #10. That story
is in the price, not in front of it. No edge, no bet.