By Patrick Strelka
Man vs Machine: Why Human Instinct Beat the AIs in the MLB Wild Card Series 2025
BlogShare
Man vs Machine: Why Human Instinct Beat the AIs in the MLB Wild Card Series 2025
The machines slipped while I stayed sharp.
In this installment of Précis AI’s Man vs Machine series, I break down my MLB postseason 2025 predictions, how the Wild Card series actually played out, where the models missed, and what we can expect heading into the Division Series.
Read: Man vs Machine: Can a Cubs Fan Beat AI Databases? MLB Postseason Is Here!
MLB Wild Card Recap 2025: How the Picks Landed
I called every Wild Card game correctly with a perfect four-for-four with the Mariners, Yankees, Dodgers, and Cubs. The AI systems? Each missed at least one – hence challenging the idea that algorithms always have the edge. Let’s take a closer look:
- ChatGPT went 2-for-4, missing on the Red Sox and Guardians.
- Perplexity, Claude, and Précis AI each finished 3-for-4 — all stumbling on the Guardians.
When every model trips on the same matchup, that’s a clue: shared inputs can create shared blind spots.
Why I Think the AIs Missed
Baseball is more than numbers and charts. Team spirit, late surges, match‑up edges, nagging injuries, clubhouse energy — they’re real, and they often don’t register cleanly in model inputs. A player with a sore shoulder won’t tank his WAR overnight, but he might change a series. A team riding a hot September can outperform its season-long mean.
We’ve already seen the models wobble even before the games started: in our series kickoff, Claude penciled in the Braves and ChatGPT still had the Mets on its bracket — neither belonged in this year’s postseason picture.
MLB Division Series Picks 2025: My Calls vs the AI Models
I’m keeping it straightforward:
- American League: Mariners and Yankees
- National League: Dodgers and Cubs
The AIs mostly line up with me on the Yankees, Mariners, and Dodgers. Where we split is the NL: most models lean Brewers over my Cubs. And for an odd twist, Claude even floated the Braves in earlier DS logic — a reminder that model outputs are only as grounded as their inputs.
Read: Man vs Machine: MLB Division Series 2025 Scoreboard Update
What This Faceoff Teaches Me
AI is phenomenal at scanning mountains of data and surfacing patterns. But it can struggle to read the moment — the mood in the dugout, the way a veteran steadies a bullpen, the confidence of a kid who just found his swing. The sweet spot is obvious: let AI handle the facts and probabilities, and let human judgment interpret the pulse of October.
As the Division Series gets underway, I’ve got the early lead. Will instinct keep me in front, or will the models regroup and surge over the next rounds? Either way, this experiment proves that in a world buzzing with new tech, human insight still matters — sometimes most when the margins are thinnest.
