TennisEdge's AI bot is backing D. Vidmanova (moneyline) at 1.66 in S. Stephens vs D. Vidmanova (Memphis (Usa) - Qualification on 2026-07-30). Result: WON.
Independent estimates of who wins, each from a different method. The blend is what the bot prices against the market.
| Method | S. Stephens | D. Vidmanova |
|---|---|---|
| Elo ratings-based, surface-weighted |
27.6% | 72.4% |
| Machine learning 29-feature gradient-boosted model |
44.0% | 56.0% |
| Serve / return point-by-point simulation |
19.9% | 80.1% |
| Blend what the bot actually prices off |
27.6% | 72.4% |
| Market implied de-vigged from the opening odds |
43.0% | 57.0% |
We publish how well these score against the closing market, including where they lose: how accurate are tennis predictions?
| Player | Surface | Elo | Record | Matches |
|---|---|---|---|---|
| S. Stephens | hard | 1620 | 53–68 | 121 |
| D. Vidmanova | hard | 1784 | 68–22 | 90 |
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. 4, 2026.
D. Vidmanova leads 1–0 over 1 meeting. Full record.
| Date | Event | Surface | Winner | Score |
|---|---|---|---|---|
| July 30, 2026 | Memphis (Usa) - Qualification | hard | D. Vidmanova | 2 - 0 |
| W | S. Sierra | 0 - 2 | Aug. 4, 2026 |
| W | H. Dart | 1 - 2 | Aug. 1, 2026 |
| L | D. Vidmanova | 0 - 2 | July 30, 2026 |
| W | Y. Starodubtseva | 1 - 2 | July 28, 2026 |
| L | R. Zarazua | 2 - 1 | July 23, 2026 |
| L | K. Liutova | 2 - 1 | Aug. 2, 2026 |
| W | K. Volynets | 2 - 1 | Aug. 1, 2026 |
| W | R. Zarazua | 2 - 0 | Aug. 1, 2026 |
| W | S. Stephens | 0 - 2 | July 30, 2026 |
| W | M. Bassols | 2 - 1 | July 28, 2026 |
All July 30, 2026 results & picks →
S. Stephens (A) vs D. Vidmanova (B) — WTA Memphis Qualification, hard, 1/8-finals (qualy)
MODELS
ELO: B 72.5% | ML: B 56.0% | Model(=ELO): B 72.5%
Sim: B 80.1% (SPW/RPW A .567/.401 [21 mch] vs B .593/.439 [50 mch]; matchup-adj serve A 52.8% / B 59.2%)
Market: Betfair B 1.66 -> impl ~60.2%; commission-adj executable 61.5%. Book A 2.2 / B 1.66; Betfair ML A 2.46 / B 1.66.
Read: THREE models (ELO, model, sim) cluster 72-80% on B; ML (56%) is the lone dissent, almost certainly
name-anchored on Stephens' former-champion pedigree. Sim inputs favour B on a fair sample (50 mch) and
Stephens' stats are built partly on a lower-tier French Open (ITF) title run vs 1500-1683 fields -> if
anything inflated, so her true level is no higher than modeled.
RANKING/ELO SANITY
Stephens ranked 1098, ELO win 27.5% -> rank and ELO AGREE she is weak now (33yo, grinding ITF/qualy vs
1500-level fields). No farmed-rating divergence; ELO credible. Vidmanova ranked 106, ELO/sim favourite,
consistent. No adjustment.
RECENT FORM (last 5)
A Stephens (33, #1098): Won Memphis-Q R1 vs Starodubtseva 1500 (4-6 7-5 2-6, GRIND); Lost vs Zarazua 1500;
Won vs Jeong 1500; Lost FO(itf) vs Bejlek 1732; Won FO(itf) Final vs Romero Gormaz 1607. Level = grinding
1500-ELO fields, wins ugly. Shape FADING/UNDER-performing at her tier.
B Vidmanova (23, #106): Won Memphis-Q R1 vs Bassols 1500 (6-2 3-6 6-4); Lost Evansville vs Ku 1500; Won vs
Hosogi; Won Newport vs Ansari 6-3 6-1; Lost Newport vs Minnen 1714. Deeper body of work vs 1600-1725 ELO
(Figueira Q-title June beating Teichmann/Aksu; Ilkley SF vs Krueger 1725). Clear level above A.
PATH/SHARPNESS
Both match-sharp: both played & won Memphis-Q R1 on 07-28 (both 3-set wins). Neither rusty.
FATIGUE
fatigue block: A matches_14d 3 / 277min, back_to_back 1, last 07-28; B matches_14d 3 / 280min, b2b 1, last 07-28.
Both WON a 3-set match yesterday -> calibrated -4.5pp 3-set-hangover haircut applies to BOTH -> cancels out.
Age mod: A 33 (harden) vs B 23 (soften) -> if anything nets slightly AGAINST A, but treated as neutral. No
extreme spot. Fatigue = roughly equal, not an edge.
NARRATIVE
Let-down: none current (Stephens' FO(itf) Final win was May, out of window).
Surface transition: A last non-null surface = clay (French Open May) but two hard qualy matches since -> no
live transition. B mixed grass/hard recently, played hard at Memphis -> no transition. Neither penalised.
Motivation: neither shows a specific data signal; nothing invented.
MARKET CHOICE
ML (B): my B win prob 0.69 (weighted to the ELO/sim/model cluster, discounting ML as name-anchored, held
below sim's 80% for two-sided uncertainty). Executable 1.66 -> comm-adj 1.627, impl 61.5%. Edge = 7.5%.
Min odds 1.05/0.69 = 1.52; Betfair 1.66 clears. -> BET, 1.5u (5-10% tier).
AH B -2.5 @ 1.9 (Betfair): sim cover 69.5%, exp margin only -3.4g. REJECT — laying favourite at -2.5 is the
high-variance tier AND Vidmanova is a grinder (Memphis 6-2 3-6 6-4; loads of 3-setters/tiebreaks in her
log) -> her own scoreline distribution argues against a reliable blowout. G3 margin-veto: cover thesis
self-negates. PASS the handicap.
Total under 21.5 @ 1.92: sim under 59%, comm-adj impl 53.4%, edge ~5.6% — clears but ML expresses the same
edge at lower variance (favourite -> ML preferred), so ML chosen.
AH: B -2.5 @ 1.90 (shadow n/a — B is favourite, not a dog ML).
refutation_check: web search returned NO_RESULTS (search unavailable) and brief has no web_insights; no
injury/withdrawal finding. Vidmanova won Memphis-Q R1 on 07-28 (2d ago) -> confirmed active/healthy. clean.
GATES
G1 edge-origin: PASS — 7.5% exchange edge from sim/ELO (72-80%) vs executable 61.5%, pre-narrative. Fatigue neutral.
G2 consensus: PASS — 3 models (ELO/model/sim) support B; all 4 have B as pick -> full stake ladder.
G3 fatal-risk: PASS — ML bet, G3 non-blocking. Upside risk (Stephens' pedigree) is normal variance, already
partly in the 1.66 price; not a fatal negating risk. (Applied as hard veto to the -2.5 AH -> passed there.)
VERDICT: BET Vidmanova ML @ 1.66, 1.5u. Winner edge from ELO/sim consensus vs a market anchored to Stephens'
faded name; lowest-variance expression of the edge (favourite -> ML over -2.5 lay).