TennisEdge's AI model finds no betting value in D. Shnaider vs L. Samsonova (Washington on 2026-07-31) — the market is efficient, with D. Shnaider favoured at 1.80.
Independent estimates of who wins, each from a different method. The blend is what the bot prices against the market.
| Method | D. Shnaider | L. Samsonova |
|---|---|---|
| Elo ratings-based, surface-weighted |
63.5% | 36.5% |
| Machine learning 29-feature gradient-boosted model |
39.7% | 60.3% |
| Serve / return point-by-point simulation |
66.2% | 33.8% |
| Blend what the bot actually prices off |
63.5% | 36.5% |
| Market implied de-vigged from the opening odds |
52.9% | 47.1% |
We publish how well these score against the closing market, including where they lose: how accurate are tennis predictions?
| Player | Surface | Elo | Record | Matches |
|---|---|---|---|---|
| D. Shnaider | hard | 1841 | 98–62 | 160 |
| L. Samsonova | hard | 1750 | 30–32 | 62 |
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. 1, 2026.
D. Shnaider leads 2–1 over 3 meetings. Full record.
| Date | Event | Surface | Winner | Score |
|---|---|---|---|---|
| July 31, 2026 | Washington | hard | D. Shnaider | 2 - 0 |
| July 2, 2026 | Wimbledon | grass | L. Samsonova | 2 - 1 |
| Aug. 11, 2024 | Toronto | hard | D. Shnaider | 2 - 1 |
| L | J. Pegula | 2 - 0 | Aug. 1, 2026 |
| W | L. Samsonova | 2 - 0 | July 31, 2026 |
| W | A. Potapova | 2 - 0 | July 30, 2026 |
| L | L. Samsonova | 1 - 2 | July 2, 2026 |
| W | E. Lys | 2 - 0 | June 30, 2026 |
| W | J. Tjen | 0 - 2 | Aug. 3, 2026 |
| L | D. Shnaider | 2 - 0 | July 31, 2026 |
| W | Xin. Wang | 0 - 2 | July 29, 2026 |
| W | M. Keys | 2 - 1 | July 27, 2026 |
| L | M. Bouzkova | 1 - 2 | July 4, 2026 |
All July 31, 2026 results & picks →
D. Shnaider (A, 22, #20) vs L. Samsonova (B, 27, #21) - WTA Washington QF, hard.
MODELS vs MARKET
Market: A 1.83 / B 2.08 (Betfair ML same) -> A ~53% implied.
ELO: A 0.635 | ML: A 0.373 (i.e. B 0.627) | model=ELO 0.635 | value_pick A, edge 0.103.
Sim (alt_markets.model): A covers -2.5 @ 57.9% -> sim also favours Shnaider outright.
=> 3 of 4 indicators (ELO, sim, market) lean A; ML is the lone dissenter favouring B.
RECENT FORM / PATH / SHAPE
Shnaider (A): Washington R16 6-1 6-2 DOM over Potapova (1724, Strong) - sharp. But that
is her ONLY hard match; transitioning off grass (Wimbledon 07-02) and her big ELO is
CLAY-FARMED at RG (beat Sabalenka 2047, Keys 1895, SF run) - surface mismatch. Fresh
(57 min, 1 match/7d). Age 22, recovers fast. Shape STABLE but hard form unproven.
Samsonova (B): MATCH-SHARP on hard - 2 wins at THIS event: DOM Xin Wang (1735, Strong)
and beat Keys (1925, ELITE) 3-6 6-4 6-4. SHARPENING on surface. 170 min /2 matches,
last 07-29 (2d rest) - no extreme fatigue spot, no 3-set-yesterday haircut.
H2H: Samsonova beat Shnaider 4-6 6-4 2-6 at Wimbledon 07-02 (grass, 4 wks ago).
FATIGUE: neither in an extreme spot. Samsonova heavier load but match-sharp beats fresh-
but-transitioning; Shnaider 22 (fast recovery) - gap is context only, not an edge.
EDGE DECOMPOSITION
A side: apparent +10% value is ELO clay-inflation; market has already discounted ELO
63% -> 53% because rating was farmed on clay and she is off grass onto hard. Discount
is correct -> no real edge on A. Backing A -1.5 also self-contradicts (doubt she wins
comfortably on hard).
B side: only ML supports B, and against ELO+sim+market. Qualitative read (hard sharpness,
H2H) leans B but is contrarian to 3 indicators and rests on least-reliable model. B
eff-impl 49.4%; I put Samsonova ~52-55% -> edge thin (3-6%), low-confidence pick'em.
Handicap A -1.5 (sim 57.9% cover -2.5): sim inputs likely clay/weak-field-biased and
contradict ML; laying a margin in a coin-flip = high variance -> pass.
Total 22.0: sim under 53% vs eff-impl under 52.8% -> ~0% edge -> pass.
GATES
G1 edge-origin: FAIL - A numeric edge is clay-farmed ELO market already discounts; B
edge is lone-ML contrarian to ELO+sim+market. No reliable model-vs-market edge.
G2 consensus: stake-tier only - indicators split; any bet would be a capped lone signal.
G3 fatal-risk: n/a - no margin bet staked.
VERDICT: PASS. Rank-equal hard QF; models split violently (ELO/sim vs ML). Decomposed, A
edge is illusory (surface-inflated ELO) and B edge is lone-model contrarian. Market's
near-pick'em is defensible across ML/handicap/total. No stakeable edge.