THE BOOK--Playing The Percentages In Baseball

The two big questions in my mind in terms of positioning fielders based on count, the style of the pitcher, teh game situation, etc., is:

Once we have some spray charts, how much do we “regress” the data toward league averages?  My suspicion has always been that teams position their players based on small-sample, unregressed, spray charts (why not, since they do everything based on too-small samples?).

And two, even we have a good estimate of a player’s likely hit locations, what is the optimal position for fielders for each game situation?

Same concerns here - I actually cringe when thinking about the kinds of analysis/decisions that might be based on 20 or 30 batted balls that occurred in a very specific situation for a specific player.
The sample size issue is particularly of concern when using the mlb.com hit data because they record the spot where a ball is picked up by a fielder. A few unusual bounces off a wall can really change the picture when you’re looking at small subset of batted ball information.