Isn't that the opposite of what analytics say? The out is worth way more than the extra base.
If you're sure enough that the batter will not get a hit, you send in a pinch hitter who you have at least a little confidence in.
The analytics currently available for free to the public go into depth on shooting percentages for various players from specific situations and positions on the floor. (For example, X player hits 47% of his threes one step to the left of the top of the circle, but only on catch and shoot. The same player shoots only 29% of pick and pop from the same place.) If these new analytics guys are worth their salaries, they can do way more than that.
If they apply analytics to practice, they can figure out extremely precise things that players should or should not be doing in game action, and design plays to get players doing the things they absolutely do best, with no coaching guess work.
The analytics in baseball are simpler, because they stop in between events and batted balls are easier to track, but basketball is a wealth of information just waiting to be data mined.
I'm starting to think the analytics guys got us a winner in Carlos. We will see.