Bahia 1–1 Corinthians: AI predicted 2–1 ✗

Bahia's structural edges, but a closer contest than the cluster suggests

All five experts land on a Bahia home win, and four of them independently arrive at the same 2-1 scoreline. The decisive question is whether the compound situational edge (rest differential plus attacking absences plus travel) is fully captured in that line, or whether there is residual value.

The rest-and-depletion stack is the strongest argument

The most persuasive thread, raised by Grok, GLM, and Kimi, is that Corinthians face a compressed Thursday-to-Sunday turnaround with a ~1,500 km trip to Salvador's heat, while Bahia sit on four clear days of rest at home. Layered on top, Memphis Depay (rested), Kayke (knee), and Vitinho (hip) strip three attacking options from a side already short on recovery. GLM's framing is the sharpest: absences and fatigue are often priced additively by models, but when they all hit the same functional area — attacking output — they interact multiplicatively. That is a genuine, specific inefficiency, not a vague form read. Corinthians' recent 3-0 over Remo, which several models flag as flattering, came at home against a side that managed zero shots on target — it is weak evidence of attacking sharpness under Sunday's conditions.

Where GPT-5.6 (the strongest 30-day model) tempers the consensus

Its counter-argument is tactically specific: Bahia are without Ramos Mingo and Caio Alexandre, two first-pass build-up players, which could produce cheap turnovers against a Diniz side that regains bodies around the ball. It also notes Willian José's seven-match scoring drought and Corinthians' fifth-best away defensive record. This is not a case to flip the pick, but it is a real reason to dampen the goal expectation and to resist treating 2-1 as high-probability.