TennisEdge's AI model finds no betting value in W. Ewald vs J. Mikulskyte (Warsaw on 2026-08-03) — the market is efficient, with J. Mikulskyte favoured at 1.53.
Independent estimates of who wins, each from a different method. The blend is what the bot prices against the market.
| Method | W. Ewald | J. Mikulskyte |
|---|---|---|
| Elo ratings-based, surface-weighted |
47.4% | 52.6% |
| Machine learning 29-feature gradient-boosted model |
59.7% | 40.3% |
| Serve / return point-by-point simulation |
58.2% | 41.8% |
| Blend what the bot actually prices off |
50.2% | 49.8% |
| Market implied de-vigged from the opening odds |
38.0% | 62.0% |
We publish how well these score against the closing market, including where they lose: how accurate are tennis predictions?
| Player | Surface | Elo | Record | Matches |
|---|---|---|---|---|
| W. Ewald | hard | 1450 | 59–40 | 99 |
| J. Mikulskyte | hard | 1499 | 123–96 | 219 |
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.
J. Mikulskyte leads 1–0 over 1 meeting. Full record.
| Date | Event | Surface | Winner | Score |
|---|---|---|---|---|
| Aug. 4, 2026 | Warsaw | hard | J. Mikulskyte | 2 - 0 |
| L | J. Mikulskyte | 0 - 2 | Aug. 4, 2026 |
| L | C. Noel | 1 - 2 | July 30, 2026 |
| W | S. Yardley | 0 - 2 | July 29, 2026 |
| L | G. Feistel | 0 - 2 | June 27, 2026 |
| W | B. Palicova | 2 - 1 | June 26, 2026 |
| W | W. Ewald | 0 - 2 | Aug. 4, 2026 |
| L | G. T. Di | 2 - 1 | July 28, 2026 |
| L | A. Herrero Linana | 2 - 0 | July 22, 2026 |
| W | M. Bayerlova | 0 - 2 | July 21, 2026 |
| L | M. Kubka | 2 - 1 | July 19, 2026 |
All Aug. 3, 2026 results & picks →
═══ W. Ewald (A) vs J. Mikulskyte (B) — Warsaw, HARD ═══ MODELS vs MARKET ELO A 47.4% | ML A 59.7% | Blend A 50.2% | Market-implied A 38.0% (odds 2.5 / 1.53) Value pick A, book edge +12.3% (blend) / +9.5% (ELO). Models disagree with market and with EACH OTHER: ML (60%) is a strong outlier, ELO conservative at 47%, blend a coin-flip. No exchange price (betfair_ml null) — edge is book-only. RECENT FORM (last 5) Ewald (20, #393) — clay specialist: 07-30 W50 Dublin L C.Noel(1500) 6-1 3-6 4-6 (lost in 3) 07-29 W50 Dublin W S.Yardley(1500) 6-3 6-2 DOM 06-27 W50 Gdansk(clay) L Feistel(1545) SF 1-6 6-7 06-26 W50 Gdansk(clay) W Palicova(1493) QF GRIND (0-6 7-6 6-4) 06-25 W50 Gdansk(clay) W Paszun(1221) Shape: STABLE/FADING — lost last, run built on clay. Only recent hard result (W50 Lopota 05-04) was a loss. Surface TRANSITION clay→hard on the backed side. Mikulskyte (30, #250): 07-28 W50 Knokke L Di(1500) in 3 07-22 W100 Amstetten L Herrero 6-3 6-0 (blown out) 07-21 W100 Amstetten W Bayerlova 7-5 6-3 07-19 W75 Vitoria L Kubka in 3 07-18 W75 Vitoria W Savinykh Shape: FADING at W75/W100 level, but higher level of play. Won W35 Tumkur HARD title 05-10 (weak field elo 1000-1460 — farmed). Some hard pedigree Ewald lacks. FATIGUE (Step 2) Snapshot showed B heavy load (443min/14d) but last match 07-28 = 5d rest now; A last 07-30 = 3d rest. Neither in an extreme spot. Cumulative load = context only, no edge. Not used. NARRATIVE Surface TRANSITION: Ewald's recent run is clay; Warsaw is hard, her hard sample is thin and negative — a transition on the BACKED side = risk, depresses level. No let-down/title-recency signal live enough to matter. ELO/RANKING SANITY ML rates Ewald 60% yet she is ranked #393 vs #250 — high-model/modest-rank, suspect the ML input for this clay player on hard. ELO (47%) is the credible read; discount ML. GATES G1 edge-origin: PASS — numeric book edge (+9.5% ELO / +12.3% blend) exists pre-narrative. G2 consensus: stake-tier — split models (ELO 47 vs ML 60), lone-credible-model → capped tier. G3 fatal-risk: n/a (ML market). VERDICT: PASS. The 12% "value" is BOOK-ONLY (no Betfair liquidity — Pricing Discipline #4, the segment the backtest loses on). The edge is driven by an outlier ML model while the credible ELO is only mildly above market; the qualitative read actively CONTRADICTS the backed side — Ewald is a clay specialist transitioning to hard with a weak hard record, lower-ranked than an opponent who holds recent hard-court pedigree. Model-market disagreement is not independently validated; no exchange price to execute at a fair line. No bet. MARKET CHOICE alt_markets: none at analysis time AH: not quoted