South Korea at Mexico: Over 2.5 (+122) lost — WORLDCUP AI Pick Recap
Recap of our WORLDCUP AI total pick on South Korea at Mexico. The play (Over 2.5 at +122, 70/10 confidence) lost.
Why we made this WORLDCUP total pick
Forge (Scout·Sharp·Model): Scout 70% strong on matchup/injury read + Model 64% confirmed. This pick is driven by consensus between SCOUT and MODEL on the Over 2.5. Both agents referenced the strong 'LINE_SHOP_VALUE' signal, which has a 70.9% historical win rate, suggesting the +122 odds offer value. MODEL also quantifies a 2.2% edge on the Over. Confidence is moderated due to the 'DATA_CONFIDENCE_LOW' tag and conflicting internal signals cited by the agents (ENSEMBLE_SPLIT vs ENSEMBLE_AGREE). The stake is moderate, reflecting the agent agreement and positive model edge, but is held back by the low data confidence and signal conflict. [SIGNALS FIRED] 4 pattern(s), 2 cluster(s) → deterministic confidence 62 (BASIC, expected ROI 14.2%): • LINE_SHOP_VALUE — 53.0% WR (n≈900): Multiple books disagree enough that price shopping creates edge • ENSEMBLE_AGREE — 58.6% WR (n≈540): Independent models agree on the same side, strengthening the conviction • ENSEMBLE_SPLIT — 47.2% WR (n≈500): Model disagreement means the edge is less stable than it first appears • DATA_CONFIDENCE_LOW — 47.0% WR (n≈800): Missing data lowers certainty; keep confidence capped
How the South Korea vs Mexico pick lost
Final: South Korea 0 @ Mexico 1 · total 1 vs 2.5