TennisEdge's AI model finds no betting value in L. Karatancheva vs N. Hurkacz (Warsaw on 2026-08-02) — the market is efficient, with L. Karatancheva favoured at 1.09.
Independent estimates of who wins, each from a different method. The blend is what the bot prices against the market.
| Method | L. Karatancheva | N. Hurkacz |
|---|---|---|
| Elo ratings-based, surface-weighted |
73.6% | 26.4% |
| Machine learning 29-feature gradient-boosted model |
79.4% | 20.6% |
| Blend what the bot actually prices off |
73.6% | 26.4% |
| Market implied de-vigged from the opening odds |
79.5% | 20.5% |
We publish how well these score against the closing market, including where they lose: how accurate are tennis predictions?
| Player | Surface | Elo | Record | Matches |
|---|---|---|---|---|
| L. Karatancheva | clay | 1430 | 72–83 | 155 |
| N. Hurkacz | hard | 1112 | 3–6 | 9 |
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 17, 2026.
L. Karatancheva leads 1–0 over 1 meeting. Full record.
| Date | Event | Surface | Winner | Score |
|---|---|---|---|---|
| Aug. 2, 2026 | Warsaw | hard | L. Karatancheva | 2 - 0 |
| W | N. Hurkacz | 2 - 0 | Aug. 2, 2026 |
| L | B. Ricci | 0 - 2 | July 17, 2026 |
| W | G. Maduzzi | 2 - 0 | July 16, 2026 |
| W | A. Zucchini | 2 - 0 | July 14, 2026 |
| L | L. Boskovic | 2 - 1 | July 11, 2026 |
| L | L. Karatancheva | 2 - 0 | Aug. 2, 2026 |
| L | L. Petruzelova | 0 - 2 | July 28, 2026 |
| L | V. Bosio | 1 - 2 | June 18, 2026 |
| W | A. Adeikyte | 1 - 2 | June 17, 2026 |
| L | D. Glushkova | 1 - 2 | June 10, 2026 |
All Aug. 2, 2026 results & picks →
=== L. Karatancheva vs N. Hurkacz — Warsaw (hard) ===
MODELS vs MARKET
ELO A 0.736 | ML A 0.794 | Book A 1.09 / B 4.22 | Betfair A 1.11 / B 8.4
Models imply B 20.6%-26.4%. Exchange effective B = 12.45% (8.4 comm-adj 8.03).
Book B 4.22 (23.7%) vs exchange B 8.4 (11.9%) = huge gap. Per pricing discipline
the exchange is the truth; the book 4.22 is a soft/longshot-biased stale quote.
The value_pick B / edge 0.059 was computed vs the BOOK price - not executable.
RECENT FORM (last 5 shown; all classification uses full window)
Karatancheva (rank 309, age 22): W75 The Hague SF run - beat A. Rus (1560),
C. Paquet (1520), I. Primorac (1502); lost SF to Boskovic (1535) in 3.
Then W35 Torino QF (lost to Ricci 1427 6-7 1-6). Legit W75-level player,
STABLE, competes with 1500-1560 ELO fields. All on CLAY.
Hurkacz (rank 1474, age listed 56 = data anomaly): only 4 matches, W15/W35
level, mostly losses - last match 2026-07-28 LOST 3-6 2-6 to Petruzelova
(1500). Beat only a 1142. FADING, weak field, below Karatancheva's level.
All on CLAY.
SHAPE/PATH: Both transitioning CLAY -> HARD (widens margin variance both ways,
favours neither). Neither has a match at Warsaw this week. No motivation signal.
FATIGUE: A well-rested (0 matches 14d). B 1 match/3 sets in 7d, 77min, rested
enough. No extreme spot. Irrelevant to a non-edge. (Age data unreliable.)
WHY THE B "EDGE" IS ILLUSORY
1. Extreme mismatch (rank gap ~1165). ELO/ML compress the tails and
over-estimate heavy dogs - the sharp exchange (~11.7% fair for B) is the
more reliable tail estimate, not the models' 20-26%.
2. Exchange 8.4 CONTRADICTS book 4.22 -> kills the B value, doesn't confirm it.
3. Qualitative read backs the exchange: Hurkacz's form (W15/W35 losses, lost
last 3-6 2-6) supports B being a ~12% dog or worse, not 20%+. Nothing
supports B being live.
No bet on A: Betfair 1.11 < min_odds 1.167 (1.05/0.90) -> automatic PASS.
alt_markets: none at analysis time (handicap/totals/model null; only betfair_ml)
GATES
G1 edge-origin: FAIL - the numeric B edge exists only by trusting compressed
tail-models over the sharp exchange; exchange + eye test say B ~12%, giving
no real edge (0.12 - 0.1245 < 0). Manufactured, not staked.
G2 consensus: n/a (stake-tier only; no qualifying edge to size).
G3 fatal-risk: n/a (ML; G1 already blocks).
VERDICT: PASS. Apparent B value is a book-vs-exchange artifact + model tail
compression; the sharp exchange price and Hurkacz's weak form both say B is a
~12% dog with no edge at 8.4. A is unbackable below its min odds.