Donna Vekic vs Lin Zhu prediction: a rest edge meets a rusty favorite in Beijing, and the model likes what Zhu is doing right now without liking it enough to bet it.
Vekic and Zhu meet Friday, October 2, 2026 at 11:00 PM ET in the first round in Beijing, and the market price on Zhu is plus-120. The model's read favors Zhu, but this one lands as a PASS, an official-card no-play. We are not backing it.
Why the model likes Zhu
Zhu's only completed match in this window is a 6-4 6-2 win over Alycia Parks on October 1, and the match holds up under the numbers. She made 32 of 47 first serves, 68.1 percent, won 78.1 percent of her first-serve points and 60.0 percent of her second-serve points, hit zero double faults, and converted 4 of her 11 break chances while facing only one herself. Parks is a top-70 player with 19 tour main-draw wins and three quarterfinals already this season, so beating her clean is a real signal rather than a soft result. That win took 75 minutes on October 1, which leaves Zhu a full rest day before tonight's match. The other side of that coin is that her previous listed singles match was all the way back on August 24, a 38-day gap, so a single strong outing is still a thin sample to build a case on.
The case against Vekic
Vekic's most recent result is a 1-6 2-6 loss to Xinyu Wang on September 22, in a match that lasted only 66 minutes. The service numbers were rough across the board: 58.8 percent first serves in, just 60.0 percent of first-serve points won and 33.3 percent of second-serve points won, two double faults, 0 of 1 on break points, and only 36 of 96 total points won. Wang is a top-40 player who went on to reach that event's quarterfinals, so the loss came against a legitimate opponent, not a trap draw. Vekic also played a September 21 doubles match alongside Maria Sakkari that she lost, adding to her workload on top of that 66-minute singles total, and she arrives in Beijing off a first-round bye with zero completed matches at this event so far, meaning no live reps on this specific draw heading in.
The number that matters
Across the supplied form windows, Vekic's first-serve-points-won percentage has sat above Zhu's in every stretch except the most recent month. Year-to-date it's 68.2 percent for Vekic against 60.1 for Zhu, over 12 months it's 68.8 against 60.2, over six months 69.7 against 61.1, over three months 67.3 against 54.0. Only in the past month does Zhu move ahead, 78.1 percent to Vekic's 67.0, and that one figure is built entirely on the single Parks match, so it reads more like a spike than an established trend. Zhu's path to the upset is to keep pressuring Vekic's second serve and lean on a return game that has generally outperformed Vekic's, while Vekic's counter path is first-serve-led aggression, the same formula that worked when she beat Zhu 6-4 6-3 on a hard court back in January 2026, a result that carries more weight here than their older, surface-mismatched 2023 meeting given how recent and how closely matched the conditions are.
The verdict
Put the two pictures together and neither is clean. Zhu's verified win chips away at her recent poor-form read without proving a sustained turnaround, and Vekic's latest match confirms the same shaky service numbers that have shown up for her for most of this year. That keeps the case for Zhu narrow rather than convincing. The model favors Zhu at plus-120, but the tier on this one is PASS, an official no-play, and we are not putting money on this match tonight. It's still worth watching how Zhu handles the extra rest and whether Vekic's second serve holds up under pressure, but for now it's a match to watch, not a match to bet.
- Event · Beijing · WTA (2026) · Friday, October 2, 2026 · 11:00 PM ET
- Market · Lin Zhu +120 — recorded when the pick was posted
- Model · T Money, Botcappers' AI Tennis model — how it is built and what it does not know
- Record · every pick is timestamped before the event and graded in public, win or lose
- Statistics · the model's own match/game database, current to the posting time above. Figures a reader cannot open here are internal to that database, not published third-party data.

