Shanghai's early rounds love dropping landmines in the moneyline, and the Botic Van De Zandschulp vs Daniel Merida Aguilar prediction for Thursday's 4:30 AM ET first-rounder is exactly that kind of match. Merida Aguilar is priced at -120, the shorter favorite on paper, but the numbers underneath that price are messier than the line suggests.
The model's read favors Merida Aguilar's side of this draw, built on how each man's serve and return actually perform rather than on name recognition. That does not mean a bet got made here, more on that below, but the reasoning behind it is worth laying out in full.
Why the model's math favors Merida Aguilar
Start with the serve matchup, because that is where the case lives. Merida Aguilar wins 53% of points when he gets to swing at a second serve, a return number that climbed as high as 64% in his US Open win over Fucsovics. Botic Van De Zandschulp, by contrast, wins only 45-47% of the points on his own second serve and double-faults on 5.5% of his total service points. That gap, Merida's return against Botic's weakest shot, is the core of the case for this side.
It shows up in the year-long numbers too. Over the last twelve months on hard courts, Botic has held serve more often than Merida, 82% to 79%, but Merida has broken serve more often, 25% to Botic's 18%. Botic has faced tougher competition in that stretch, the 64th percentile of opponents to Merida's 51st, yet even adjusting for that, Merida's combined hold-and-break total still runs about one to two points ahead. Merida's own serve rarely implodes, with a 3.7% double-fault rate that blunts the bite of Botic's typical 18% break rate against the field.
The case against: Botic's blueprint for hittable serves
Botic has a specific, repeatable way of beating a mild server on hard courts, and it is not a small sample. Against weaker servers on hard, he has held serve 85% of the time and broken 25% of the time across 201 service games. His own first serve produces an ace on 9% of points, which keeps a returner like Merida, whose serve generates aces on just 4.8% of points, away from the second-serve looks he needs to do damage. That 25% break rate against this type of server is the real counterweight to Merida's case, and it is built on hard-court data, not a hunch.
The rust question
Form complicates both sides. Since his US Open quarterfinal run, Botic is 2-2, including a clean loss to a player ranked outside the top 140 and a blown 5-2 first-set lead against Cerundolo, where he missed break points early and piled up double faults in the deciding set. Two qualifying wins in Beijing kept him on court, but the win over Royer needed saving five break points, hardly a statement result, and the form line reads closer to his twelve-month level than to his US Open peak.
Merida Aguilar's picture is thinner. He has not played a hard-court match in five weeks, and his only result since the US Open is a clay win over Alejandro Tabilo in Davis Cup qualifying, worth noting for health and form in general but not for hard-court sharpness, since it came on a different surface. His last hard-court match was a straightforward three-set loss to Alex Michelsen at the US Open. A five-week hard-court layoff against a player who holds serve behind a 9% ace rate and a proven record versus mild servers is a real risk, and it is the main reason this stays off the card as a bet.
The verdict
This match is marked PASS: no official action here. The numbers point toward Merida Aguilar at the -120 moneyline for Thursday's 4:30 AM ET first-rounder in Shanghai, but that case does not clear the bar to become a bet. This is simply not a play we are making.
- Event · Shanghai · ATP (2026) · Thursday, October 8, 2026 · 4:30 AM ET
- Market · Daniel Merida Aguilar -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.

