Using AI and baseball analytics to build a better lineup

A youth-baseball lineup has to balance player performance, development, team constraints, and the opponent. Coaches often hold that context in their heads, which makes the decision hard to inspect or revisit.

I used Claude Cowork to combine player statistics with opponent context and propose a lineup. The exercise borrows the basic Moneyball idea: use evidence to challenge instinct, then leave the coach responsible for the decision.

The useful output is not a supposedly perfect order. It is a recommendation with visible assumptions that can be questioned and changed.

This is a personal experiment for the 2026 baseball season. It does not publish player-level data or automate the coaching decision.

Analytics is most useful when it makes judgment easier to examine. The model should show its reasoning and constraints, then let the coach overrule it.