TennisEdge's AI model finds no betting value in M. Podlinska vs V. Valdmannova (Warsaw on 2026-08-04) — the market is efficient, with V. Valdmannova favoured at 1.21.
Independent estimates of who wins, each from a different method. The blend is what the bot prices against the market.
| Method | M. Podlinska | V. Valdmannova |
|---|---|---|
| Elo ratings-based, surface-weighted |
3.7% | 96.3% |
| Machine learning 29-feature gradient-boosted model |
8.4% | 91.6% |
| Blend what the bot actually prices off |
10.4% | 89.6% |
| Market implied de-vigged from the opening odds |
20.1% | 79.9% |
We publish how well these score against the closing market, including where they lose: how accurate are tennis predictions?
| Player | Surface | Elo | Record | Matches |
|---|---|---|---|---|
| M. Podlinska | clay | 1263 | 60–55 | 115 |
| V. Valdmannova | hard | 1730 | 63–30 | 93 |
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 30, 2026.
V. Valdmannova leads 1–0 over 1 meeting. Full record.
| Date | Event | Surface | Winner | Score |
|---|---|---|---|---|
| Aug. 4, 2026 | Warsaw | hard | V. Valdmannova | 2 - 0 |
| L | V. Valdmannova | 0 - 2 | Aug. 4, 2026 |
| L | U. Radwanska | 0 - 2 | Aug. 2, 2026 |
| L | K. Kajabova | 2 - 0 | July 30, 2026 |
| W | A. Kmiecik | 1 - 2 | July 29, 2026 |
| L | S. Iliev | 2 - 1 | July 23, 2026 |
| W | M. Podlinska | 0 - 2 | Aug. 4, 2026 |
| W | A. Rajecki | 2 - 0 | Aug. 2, 2026 |
| W | V. Ryser | 2 - 1 | Aug. 1, 2026 |
| W | V. Allen | 2 - 1 | July 31, 2026 |
| W | I. L. Van den Broek | 2 - 1 | July 30, 2026 |
All Aug. 4, 2026 results & picks →
M. Podlinska (A) vs V. Valdmannova (B) — Warsaw (WTA 125), hard, 1/16-finals
═══ MODELS vs MARKET ═══
ELO: A 3.7% / B 96.3%
ML: A 8.4% / B 91.6%
Blended: A 10.4% / B 89.6%
Book: A 4.80 / B 1.21 (implied A 20.1%)
Betfair ML: A 5.80 / B 1.19 <- executable
Pinnacle: A 5.05 / B 1.177, no-vig A 18.9% / B 81.1% (fetched 2026-08-04 05:55Z, ~fresh)
Executable B price commission-adjusted: 1 + 0.19*0.95 = 1.1805 -> implied 84.71%.
Pinnacle fair price for B = 1/0.811 = 1.233. We would be taking 1.19 on a side the
sharpest market in the world prices at 1.233 — the exchange quote is ~3.6pp WORSE
than sharp fair. All three of our models sit 8.5–15pp ABOVE Pinnacle on B.
═══ WHY THE MODELS ARE THE OUTLIER ═══
The resimulate endpoint returns "Insufficient serve/return data": Podlinska has only
3 usable matches (spw/rpw null), so no sim probability exists for this match at all.
Her ELO is built on a thin ITF W15 sample — an ELO of the "farmed-low, data-starved"
kind. ELO's 96.3% for B is not a credible WTA125 R1 number (that is a top-10-vs-
qualifier price). ML at 91.6% is more sober but still well clear of the sharp line.
ELO/ranking cross-check: A ELO-dog and ranked 651, B ranked 218 — direction agrees,
so no farmed-high flag, but the magnitude rests on almost no data.
Honest number for B: between Pinnacle's 81% and ML's 92%, call it ~87–88%.
At 0.88 -> edge 3.3%; at 0.87 -> 2.3%. Both under the probe floor once measured
against the executable price, and against a sharp line that says the price is bad.
═══ RECENT FORM (5 most recent) ═══
Podlinska (A), 27, rank 651 — FADING
08-02 L Warsaw quali Final 1-6 3-6 vs U. Radwanska (1306) — beaten badly by a
sub-1350 opponent; she is in the main draw as a lucky loser
07-30 L W15 Bielsko Biala (clay) 6-3 6-3 vs Kajabova (1500)
07-29 W W15 Bielsko Biala (clay) 2-6 7-5 2-6 vs Kmiecik (1500) — GRIND vs Avg
07-23 L W15 Nogent (clay) 6-2 1-6 7-6 vs Iliev (1500)
07-22 W W15 Nogent (clay) 4-6 1-6 vs Soulie (1500) — GRIND vs Avg
Shape: FADING. Grinding at W15 level against 1500s, then a 1-6 3-6 quali-final
loss to a 1306. TRANSITION: her body of work is clay; this is hard (last match
was hard, so the switch is already made — noted, not weighted).
Valdmannova (B), 18, rank 218 — SHARPENING
08-02 W W50 Dublin FINAL 6-1 7-6 vs Rajecki (1500) — TITLE
08-01 W W50 Dublin SF 6-2 6-7 6-1 vs Ryser (1500) — GRIND
07-31 W W50 Dublin QF 4-6 7-6 6-4 vs Allen (1500) — GRIND
07-30 W W50 Dublin R16 6-4 1-6 6-1 vs Van den Broek (1500)
07-29 W W50 Dublin R32 6-2 6-2 vs Reva (1500) — DOM
Shape: SHARPENING (5-0 run, closing 6-1 in the final set repeatedly). Level is
clearly above Podlinska's; hard-court pedigree fine (W50 Lopota title, Istanbul
quali run). But it is a W50 field of 1500-ELO opponents, not a level-up.
═══ PATH ═══
A: 1 match at this venue (lost the quali final) — lucky loser, home player in Warsaw,
acclimatised to the courts. Match-sharp for what she is.
B: FIRST match at this venue. Arrives from Dublin — 5 matches in 5 days ending 08-02,
then travel. Rust is not the issue; freshness is.
═══ FATIGUE (brief-supplied) ═══
A: 3 matches/7d, 5/14d, b2b 2, 473 court-min, last 08-02. Age 27 -> baseline tier.
B: 5 matches/7d, 5/14d, b2b 4, 458 court-min, last 08-02. Age 18 -> recovers fast,
soften one tier.
No 3-set-hangover haircut applies: B's last match was a TWO-set win and it was two
days ago, not yesterday. Per the measured study a 2-set win is ~neutral. Court-time
is near-identical (473 vs 458) — context only, no edge either way.
═══ NARRATIVE ═══
LETDOWN (B): won a Final 2 days ago (08-02) = the 0–2d LETDOWN band, after a 5-match
week, plus a Dublin->Warsaw move. This is the classic flat-first-round spot and it
cuts AGAINST the 1.19 favourite. It cannot create an edge on A (G1), but it removes
any comfort about paying above sharp fair on B.
Venue: A is the home player in Warsaw with courts under her belt; B is cold on site.
Title recency: TITLE: W50 Dublin 2d ago.
═══ MARKET CHOICE ═══
Handicap quoted: A +5.5 @ 2.18 / B -5.5 @ 1.75. Totals 18.5: over 1.97 / under 1.93.
alt_markets.model is null AND resimulate cannot produce a distribution (A has no
serve/return data) — so there is NO margin anchor for either the cover or the total,
and no independent margin read exists. Per Step 4b that forces ml-only; a cover
thesis here would be free-handed, and fatigue/letdown may never build one.
AH: A +5.5 @ 2.18 (not staked — no margin model available).
Market evaluated: ml only. Best available: B @ 1.19 exchange -> edge under bar.
═══ GATES ═══
GATES
G1 edge-origin: FAIL — pre-narrative edge vs the executable 1.19 (eff 1.1805,
84.71%) is 4.9% on the blended model and only ~2–3% on an honest probability
that respects Pinnacle's 81.1%; below both the 5% full bar and the 3% probe
floor. The models' gap is manufactured by a data-starved ELO (3 matches, sim
refuses to run), not by market error.
G2 consensus: PASS (tier only) — 3 models on B, but they agree with each other,
not with the sharp price; would be full-ladder tier if G1 had cleared.
G3 fatal-risk: PASS — not a margin bet; G3 does not block ML.
G1 FAIL -> decision is pass.
═══ VERDICT: PASS ═══
Podlinska is genuinely bad and Valdmannova should win this — the match-winner call is
easy and I am not disputing it. But the job is pricing, not forecasting. The only
executable price is 1.19, which after commission needs ~89.7% to be a bet, and the
sharpest reference on the planet says 81.1%. Our models say more, but the model that
says most (ELO 96.3%) is built on three matches, and the point-by-point sim declines
to price this match at all. Paying above sharp fair on a 1.19 teenager who won a
title 48 hours ago, played five matches in five days, travelled, and now faces a home
player who has already hit balls on these courts is exactly the losing shape of bet.
No side offers 5% at a price we can take. Pass.