McClanahan Cashed. Embiid Did Not.

The pick was Shane McClanahan Under 15.5 pitcher outs (-130), and it wasn’t close. The triage logic was that this wasn’t a talent bet — it was a team policy bet. McClanahan is four weeks into his return from two surgeries spanning nearly three full seasons. The Rays had publicly committed to managing his workload. Start 1 didn’t reach five innings. For Start 2 to go over 15.5 outs (5.1+ innings), the Rays would have had to abandon their own stated medical plan against a Cubs lineup that isn’t soft. They didn’t. He recorded 12 outs and was done in the fourth.

The signal was unusually clean on this one. The triage and Claude both landed on McClanahan as the single safest play on the board. Perplexity disagreed — it went with Joel Embiid Under 40.5 Points + Rebounds + Assists, citing injury-management reporting and his season production line. That split turned out to be meaningful: Embiid finished with 47 PRA against San Antonio, clearing the threshold by six. The model that picked the policy angle was right. The model that picked the star-under narrative was wrong. The overall board finished 4-5, which is variance with no signal.

The result for the top pick: win. McClanahan covered by 3.5 outs. The result for the placed martingale bets: both losses. The 3-loss experiment had Embiid Under 30.5 P+A — Embiid finished with 35. The 6-loss experiment had Embiid Under 40.5 PRA — Embiid finished with 47. Both series double their stakes going forward.

The learning is about bet type, not outcome. A structural under based on organizational intent — a team following their own injury management plan — is categorically lower variance than a star-player combo under. The McClanahan bet required zero performance assumptions. It only required the Rays to be consistent with what they’d already said publicly. The Embiid bets required him to have a below-average night against a weak opponent. He didn’t. One of those scenarios is under your control; neither team had any reason to deviate from a plan they’d already telegraphed. When the reasoning is “they’d have to actively contradict themselves,” that’s a different quality of edge than “he probably won’t go off tonight.”

Both experiments sided with Perplexity on Embiid and lost. The Final Recommendation and Claude were aligned on McClanahan — and that one won.