Nauhany Vitoria Leme Da Silva vs Astra Sharma prediction time, and this one is messier than the plus-money price suggests. Leme Da Silva is the live underdog at +106 for Wednesday's 12:00 PM ET match in Sao Paulo, and the model's fundamentals read favors her side of the draw. The problem is what happens when she misses her first serve.
Sharma has lost five straight tour-level matches in 2026, so this is not a case of fading a hot player. It is a case of two players who both bleed points behind a shaky second delivery, and whoever controls that exchange wins the match.
Why the model leans Silva
The teenager has been busier and sharper than her name recognition suggests. Over four days in late August at Barueri she beat three players in a row, including wins over opponents ranked roughly No. 498 and No. 1125, before a three-set loss to a player ranked near No. 255. She followed that with a run to the US Open juniors quarterfinals, taking out the No. 11 seed along the way before losing in three sets in the round of eight. That is four singles matches and nine sets played in the two weeks before this one, plus two doubles matches, which argues against any read of her as rusty or inactive.
Sharma's log tells the opposite story. Her five recorded 2026 singles results are all losses, most recently a straight-sets defeat to a player ranked near No. 517 on August 25, and she has not played a match since, 22 days and counting. She is also working back from an October 2025 lower-back injury, and nothing in her return has flipped a result yet.
The second-serve math that worries the desk
Here is the catch. Leme Da Silva's own service numbers show the same weakness Sharma is built to attack. In her Barueri win over Cabezas Dominguez she won 80.6% of first-serve points but just 30.4% on the second serve, with six double faults. Against Dourado the gap was 88.0% to 68.2%. Against Barros it was 66.7% to 36.7%, with eight double faults in that match alone. Every one of those numbers drops hard once the first serve misses, and Sharma's game plan is built specifically to make her live on that second ball.
Sharma has the same issue, which is part of why this stays close on paper. Against Herea she won 75.0% of first-serve points but just 31.8% on the second, with seven double faults of her own. Against Ku she won 48.1% of first-serve points and only a third of her second-serve points. Neither player protects her second serve well, so the match likely comes down to whichever player gets more crossmatch than pressed on hers.
Break points back that up. Leme Da Silva converted 5 of 7 break chances against Cabezas Dominguez but only 6 of 16 against Barros, and she failed to convert against Brace in the loss. Sharma converted just 1 of 4 against Herea and 3 of 11 against Ku. Nobody in this match has shown she can reliably close out the break chances she creates, which is exactly why the desk is not willing to lean hard on either side's raw form.
Why this is a pass, not a bet
To be plain about it, this is a PASS. The model's read favors Leme Da Silva's side of the matchup, driven mainly by her recent match count and the quality of her Barueri and US Open results against Sharma's five straight losses and 22-day layoff. That is not the same as an official bet, and none is going out here. The reasoning above is the actual basis for the lean, not a hedge on a play we are also making elsewhere on the card.
The verdict
Leme Da Silva's activity and recent scorelines give her the better story heading into Wednesday, but her second serve is a live weakness and Sharma is specifically equipped to test it. Two shaky second serves and a thin break-point sample on both sides is not a number the desk is willing to back at +106. No play here.
- Event · Sao Paulo · WTA (2026) · Wednesday, September 16, 2026 · 12:00 PM ET
- Market · Nauhany Vitoria Leme Da Silva +106 — 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.

