Advanced stats are useful in baseball for a few different reasons, especially compared to other sports.
1. There are just a lot more datapoints ... twice as many games as any other sport, duh
2. The fundamental exchanges in baseball are fairly simple and happen sequentially. Pitcher pitches, hitter hits and then the ball goes into play and something happens there. Compare the challenge to trying to measure the results of a football play, where 22 guys are in motion and all of them have some impact on the success of the play. You also don't have the information disadvantage you have in football or basketball (where the play was supposed to go, or what a defender's assignment is). Aside from defensive shifts, the movements are isolated, and with few exceptions, hitters and pitchers are chasing the same thing every pitch. (an out, or a non-out)
Scouting and metrics complement each other. You do need metrics for your own guys, just because the metrics are unbiased and can show things the scouts might not see. (and maybe allow you to change emphases) While metrics can identify the components of a good player, only scouts can take some of that information and make it actionable. Clearly something in the Red Sox process is squirrely there. One I have identified is I don't know if they have been able to fully capture the impact of Fenway Park on defensive measurement - because there have been some confounding results with some choices that looked like they should have worked out.