Our Alina Charaeva vs Elina Avanesyan prediction for Thursday's second-round match in Sao Paulo starts with a lopsided number: the market has Avanesyan at -185, first ball at 11:00 AM ET. That price says the book expects a comfortable favorite. Our read is closer than that suggests.
This is a PASS: no official play, no side backed on the card. The fundamentals split down the middle. Avanesyan comes in fresher and her recent serve trend is trending up. Charaeva has a specific, verified way of beating her, and she already used it once this year.
Why Avanesyan is the short price
Start with the schedule. Avanesyan needed 90 minutes to beat Julia Riera 6-2, 6-4 on September 15 and has had a full rest day since. Charaeva needed three sets and roughly 186 minutes to get past Xiaodi You, a match that spilled from September 15 into a finish on September 16. She is walking onto court on shorter turnaround.
The serve numbers back up the price, but only recently. Across the longer windows, Charaeva has been the better server: her first-serve-points-won rate sits at 61.7 percent year to date and 61.8 percent over both the 12-month and 6-month windows, versus 58.0, 58.0 and 57.9 percent for Avanesyan over those same spans. But the trend flips in the shorter windows. Over the last three months Avanesyan is at 60.8 percent to Charaeva's 59.8. Over the last month, Avanesyan is at 61.2 percent to Charaeva's 53.6. That is a real crossover, even on a small recent sample.
The opener stat lines point the same way. Avanesyan won 65.8 percent of her first-serve points and 48.0 percent of her second-serve points against Riera, with six double faults. Charaeva won 55.6 percent of her first-serve points and 41.5 percent of her second-serve points against You, with nine double faults. Riera's own serving was weak, which inflates Avanesyan's numbers somewhat, but the gap is still there.
The case against
Here is the specific risk the model flags, not a generic caution. Charaeva already beat Avanesyan this year, 6-4, 6-1 at Antalya on March 12, taking 62 of 102 points played. That result came on clay, so it counts as secondary evidence rather than a same-surface anchor for a hard-court match. It does not decide this pick on its own.
What makes it relevant is the pattern behind it, not just the scoreline. The read is that Charaeva sustains her first-serve advantage and repeatedly attacks Avanesyan's weaker service games, and that is the same plan available to her again today. The hard-court break numbers support the idea that Avanesyan's serve is the softer target: Avanesyan is converting break points at 48 percent against a 56 percent hold rate, while Charaeva is converting at just 31 percent against a 58 percent hold rate of her own. Charaeva also just saved 7 of 15 break points in her three-setter, a sign her own service games can get pressured, but the larger gap favors her finding break chances against Avanesyan, not the reverse.
The number that matters
Total points won is the number that is easy to miss. Charaeva won her three-set match against You but finished with exactly 101 of 202 points, 50.0 percent, the definition of a match with no overall point superiority. Avanesyan finished her straight-set win at 73 of 126 points, 57.9 percent. Different opponents, so it is not a clean apples-to-apples read, but it lines up with everything else here: Avanesyan looks like the more efficient player right now, even if the sample is thin and Charaeva's schedule this season has generally been the tougher one.
The verdict
This one is a PASS. The recent serve trend, the rest edge and the opener efficiency numbers lean toward Avanesyan, but Charaeva has a specific, verified path back into this match and has already executed it once this year. That is enough tension that we are not backing either side at today's price. No stake, no official play. Kickoff is 11:00 AM ET Thursday.
- Event · Sao Paulo · WTA (2026) · Thursday, September 17, 2026 · 11:00 AM ET
- Market · Elina Avanesyan -185 — 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.

