The Oleksandra Oliynykova vs Mai Hontama prediction for Friday, October 23, 2026, in the Tokyo WTA qualifying draw (11:00 PM ET) comes down to which collapse looks more survivable. The model lands on Hontama, but this one is marked PASS, an official-card no-play, so there is no action behind it.
That split matters. The lean is real: Oliynykova's serve has gone somewhere over the last month, and Hontama's history says she can blow a lead as fast as she builds one. The model worked through both flaws and still came down on Hontama's side of the line. It just is not a bet being made.
Why the model leans Hontama
Start with the serve. Over Oliynykova's last four hard-court matches she won about 37% of total serve points, 40.7% on first serve and 27.5% on second, and against No. 114 Mary Stoiana on Oct. 20 she held zero of nine service games. Eight days before that, she retired trailing Aliona Falei 6-7, 0-4, her third injury-related stoppage of the year. That is not one bad match. That is a serve that has not worked in over a month, attached to a body that has broken down more than once.
Hontama's level, meanwhile, has held where it should. Accounting for opponent strength, her year-long hard-court baseline is close to Oliynykova's: she has held 59% and broken 40% of the time against a typical ranked-49 opponent, against Oliynykova holding 56% and breaking 36% against a typical ranked-70 opponent, both edges running Hontama's way. The two also share a hard-court opponent this year: against Diane Parry, Hontama won 6-4, 6-0, while Oliynykova lost 4-6, 3-6. Hontama also gets a rest edge, five days off against Oliynykova's four, and this one is a home event for her, for whatever that is worth in a qualifying match.
The results underneath confirm it is not a fluke read. Hontama's wins and losses this year sort cleanly by opponent rank: she beats players in the 150-250 range and loses to the ones ranked 60-120. Oliynykova, at No. 62, sits right in the band that has given Hontama trouble, except Oliynykova is not showing up this month as a No. 62 player. She is showing up as a server who cannot hold.
The case against Hontama
The flip side is Hontama's own serve and her tendency to fade. Against Viktorija Golubic on Oct. 19 she won just 33% of second-serve points, and she did it after building a 3-1 lead with two breaks, then lost 11 of the next 12 games. If Oliynykova's return is still finding its range (she broke five of eight return games against Stoiana even while her own serve cratered), that loopy, moonball-heavy game could be built to exploit exactly that kind of second-serve weakness and exactly that kind of letdown from ahead. A server who cannot hold is still dangerous if she can flip the match into a return contest.
The number that matters
The gap that keeps this from being a clean case either way is surface-adjusted record. Oliynykova's No. 62 ranking is built mostly on clay; she is 8-10 on hard courts at all levels this year. Hontama has no such ranking to point to, sitting around No. 164-168, but her hard-court record this year is roughly 24-20, mostly against ITF and 125-level fields. One player has the better ranking and the worse hard-court season. The other has the opposite. That is most of the match in one line, and it is why the model's lean toward Hontama does not clear the bar to be a play.
The verdict
Put plainly: this is a PASS. The model favors Hontama on the balance of the evidence, but it is not backing her with real money here, and that is not a soft call dressed up as cautious. It is simply not a bet being made on this one. Whichever side of the net finds its serve first on Friday night is probably going to win it.
- Event · Tokyo · WTA (2026) · Friday, October 23, 2026 · 11:00 PM ET
- 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.

