Our Mary Stoiana vs Dominika Salkova prediction for Thursday, September 17, 2026 at 1:30 PM ET starts with a clear lean and ends with an empty bet slip. The model's read of the fundamentals favors Stoiana, and the market agrees hard, pricing her at -294. That gap between what the model sees and what it's willing to stake is the whole story here.
This is a PASS. No side goes on the official card, and no unit gets risked. That's worth saying plainly up front so nobody mistakes the analysis below for a play.
Why the model leans Stoiana
Start with recent hard-court form, since that's the surface both players have actually been tested on. Stoiana is 7-3 on hard since August 11, with qualifying runs through Cincinnati and the US Open mixed in, and she's sitting at a new career-high ranking of 116. Salkova is 2-3 on hard over the same stretch and has slipped to 125.
The clearest head-to-head proxy came a day apart in Barranquilla. Both players faced Julia Riera on hard court within 24 hours of each other. Stoiana took a set off her and won roughly 49% of total points. Salkova won four games and roughly 38% of points. Same opponent, same surface, same week, and Stoiana came out well ahead of Salkova on that shared measuring stick.
Their first-round wins this week point the same direction. Stoiana beat Claire Liu, who had won the Barranquilla title just three days earlier and carries a strong recent-form rating. Salkova's first-round win came against a much lower-rated opponent. Neither result is enormous on its own, but stacked with the hard-court record and the Riera comparison, it's why the model has Stoiana ahead here.
The break-point math
Both of these are shaky servers, and that's exactly where the match gets decided. Stoiana holds serve 63% of the time on hard court, and it's not clean: she was broken four times by Liu and five times by Riera in her last two matches. Salkova's return game is built for exactly that kind of target. Against lower-level servers on hard, she's winning 57% of second-serve return points and breaking 42% of the time.
Salkova's own serve is the shakier of the two, though. Over the last twelve months on hard she's winning 58% of first-serve points and 44% of second-serve points, both behind Stoiana's 61% and 48%. Her double-fault rate has also climbed from 5.9% to 10.2% over the last month. Stoiana's second-serve return, meanwhile, is winning 58% of points, which lines up well against that trend.
The wildcard is Salkova's own first-round serving line: she held 8 of her 9 service games. If that carries over, this turns into a match decided by whoever survives more of the other player's return pressure, and the model still gives Stoiana the edge in that exchange given the season-long gap between their serves.
Why this stays a pass
None of that adds up to a bet we're making. The model's lean toward Stoiana is real, but this candidate did not clear what it takes to become an official play, and that's as far as we're going to characterize it. We're not naming a reason, because guessing at one risks stating something we haven't actually confirmed.
What we can say is that PASS here means exactly that: no side, no stake, no play. It isn't a soft lean dressed up as something more, and it isn't a knock on the analysis above. It's a straightforward call that this one doesn't get backed today.
The verdict
The fundamentals point to Stoiana on hard-court form, the Riera comparison and the serve-versus-return math. The market has already priced her a heavy favorite at -294. Between those two facts, there's no bet here. Call it PASS and move to the next match.
- Event · Sao Paulo · WTA (2026) · Thursday, September 17, 2026 · 1:30 PM ET
- Market · Mary Stoiana -294 — 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.

