TennisEdge's AI model finds no betting value in Xin. Wang vs L. Samsonova (Washington on 2026-07-29) — the market is efficient, with L. Samsonova favoured at 1.51.
XIN. WANG (A) vs L. SAMSONOVA (B) — WTA Washington, hard, 1/8-finals
MODELS vs MARKET
ELO Wang 52.9% ML Wang 48.4% blend Wang 50.8%
Book Wang 2.57 / Samsonova 1.51 (implied Wang 37.0%)
Betfair Wang 2.60 / Samsonova 1.60 (executable — this is the baseline)
Commission-adj: Wang eff 2.52 -> 39.7% breakeven | Samsonova eff 1.57 -> 63.7%
Raw gap: blend 50.8% vs exchange 39.7% = +11.1pp on Wang. That gap is the ONLY
reason this match got a second look — the question is whether the ELO/ML pair
or the exchange is the stale number.
RECENT FORM (5 most recent each; all fetched matches used for classification)
Wang (24, rank 33)
07-28 Washington hard W Pareja 1612 7-6 6-3 GRIND vs Avg
07-01 Wimbledon grass L Bencic 1851 5-7 0-6
06-29 Wimbledon grass W Cocciaretto 1660 6-3 2-6 6-2
06-26 Bad Homburg grs L Osaka 1739 3-6 3-6
06-25 Bad Homburg grs W Svitolina 1500 (0-0 walkover — no data)
Also: DOM over Fernandez 1663 (6-3 6-4) and Zarazua 1579 (6-1 6-2) on grass;
clay window bad — L to Eala 1499 (4-6 3-6), L Korpatsch, L Charaeva 1675.
Samsonova (27, rank 21)
07-27 Washington hard W M. KEYS 1925 3-6 6-4 6-4 <- Elite scalp, from a set down
07-04 Wimbledon grass L Bouzkova 1811 3 sets
07-02 Wimbledon grass W Shnaider 1672 3 sets
06-30 Wimbledon grass W Kudermetova 1679 6-3 6-3 DOM
06-23 Bad Homburg grs L Svitolina 1816 3 sets
Also DOM over Siniakova 1774 (6-3 7-5); clay window bad (1 win in 4).
SHAPE / PATH (Step 1b)
Wang — STABLE/FADING on the evidence that matters here. Grass peak (BH semi) is
4 weeks stale; then a 27-day gap (07-01 -> 07-28) = rust; her one hard-court
data point in the whole window is a 7-6 6-3 GRIND over a 1612 — winning ugly
against a weak field is not shape. Zero hard-court form to validate the ELO.
Samsonova — SHARPENING, and on the right surface: beat a 1925 ELO (Keys) from a
set down at THIS venue two days ago. That is live, on-surface, same-conditions
evidence, the strongest evidence class available.
Path: both 1 match deep at the venue; neither rusty now, but Wang carried the
layoff into it.
FATIGUE (Step 2 — context, not an edge)
Wang: 1 match/14d, 83min, played YESTERDAY — but a 2-SET win (7-6 6-3), so the
measured 3-set-hangover haircut (-4.5pp) does NOT apply; a 2-set win yesterday
is ~neutral. Age 24 = baseline recovery.
Samsonova: 1 match/14d, 110min 3-setter, TWO days ago — hangover haircut applies
only to yesterday's winners, so no haircut. Age 27 = baseline.
No extreme spot on either side. Roughly equal. Fatigue plays no part below.
NARRATIVE
Let-down: none (no Final in window either side). Venue momentum: 1-0 both.
Title recency: none. Surface transition: neither — both already played this
hard court this week.
ELO vs RANKING SANITY — this is the crux
Wang is the ELO favourite (52.9%) while ranked 33; Samsonova is the ELO dog
while ranked 21. Divergence in both directions at once, and both point the same
way: Wang's rating was built on a grass fortnight plus weak-field wins, while
Samsonova's was bled on her two WORST surfaces (clay: L Teichmann/Kasatkina/
Potapova; grass: L Bouzkova/Svitolina/Mertens/Dart 1594). Fast hard is her best
surface and the rating hasn't caught up. Framework rule applies literally:
strong rank + low ELO -> ELO understates, do NOT back the opponent on the ELO
gap; high ELO + modest rank -> discount unless live form independently backs it.
Wang's live form does not back it (one grind win over a 1612; lost to a 1499 on
clay). Samsonova's live form independently backs the market price.
GATES
G1 edge-origin: [PASS] — +11.1pp is numeric model-vs-market, pre-narrative; no
fatigue/letdown anywhere in it.
G2 consensus: [capped tier] — both models nominally agree, but both lean on the
same contaminated input (surface-blended ELO inflated on Wang,
deflated on Samsonova). Not full-ladder evidence.
G3 fatal-risk: [n/a] — ML-only market (handicap and totals null in alt_markets).
PRICING TEST (my probability, not the models')
Discounting the surface-blend contamination, I put Wang at 38-40% — essentially
the exchange price. Backing Wang at 2.60 needs >=44.7% (min odds 1.05/0.40 =
2.63 — the price is already below my minimum). Backing Samsonova at 1.60 needs
>=68.7%; I have her ~61%. Neither side clears the 5% bar at the executable price.
VERDICT — PASS
G1 clears, so this is not a gate rejection: it is a pricing rejection. The
11pp "edge" is an artifact of a rating that is stale in both directions, and the
one piece of live on-surface evidence (Samsonova over Keys 1925 at this venue,
two days ago) confirms the exchange, not the models. No articulable information
the market is slow to price. No alt markets to express a margin view. Pass.