A. Kalinskaya vs D. Kasatkina — prediction

The model's baseline split (61% vs 45%) is anchored in a substantial ranking and Elo gap: Kalinskaya sits at #20 with a 1785 Elo rating, while Kasatkina is ranked #65 with a 1645 rating. A 140-point Elo differential and a 45-spot ranking gap represent a real quality disparity that shows up directly in the model's probability, not just as a market narrative.

Kalinskaya's ranking trend (+4) also moves in the right direction, while Kasatkina's (-12) suggests recent erosion in results relative to the field. This reinforces the class gap rather than offsetting it.

Kalinskaya's 63% serve-points-won rate is the standout number here, nine points clear of Kasatkina's 54%. That gap matters more than it might appear because Kasatkina's return game (46% return points won) is only five points better than Kalinskaya's (41%) — meaning Kalinskaya's serve edge outweighs Kasatkina's return edge in net terms.

In practice, this suggests Kalinskaya should be able to hold serve more comfortably than Kasatkina can convert return chances, giving her control of a larger share of service games over the match.

Both players carry a scheduling risk, but of different kinds. Kalinskaya returns from a 25-day layoff with zero matches in the last two weeks — the model itself flags possible rustiness. Kasatkina, by contrast, arrives with match rhythm (2 matches in 14 days) but is only 2 days removed from a Washington final, which raises legitimate deep-run fatigue concerns.

These two risks point in opposite directions and are not fully resolved by the data — one player risks being undercooked, the other overplayed. Neither number allows a confident lean, so this factor is best read as a wash rather than a swing point.

The model gives Kalinskaya a 66% win probability against a market-implied 62% (odds of 1.61), producing a modeled 7% expected value. That's a modest, not dramatic, edge — the model and the market are largely in agreement on Kalinskaya as the favorite, with only a small divergence in degree.

Being the favorite here is not the same as the bet carrying strong value: a 4-point gap between model and market is within the normal noise band for a WTA factor model with ~64% out-of-sample accuracy. The serve advantage and ranking/Elo gap support the lean toward Kalinskaya, but the rest/fatigue picture is mixed and the head-to-head is tied, so this should be treated as a fair, not exceptional, positive-EV opportunity.