Obviously, you want to draft players who score points. The standard idea is that you need points to win. However, there is a slightly skewed way to look at FBB competition. Instead of looking at total points, you can argue that if you simply beat your opponent at each position, you can garner a victory that way, too. And that's where position-mean analysis comes in handy.
If you take that opinion, then guys who score drastically above the average for their position take on a new value. The theory is that you don't need a terribly high-scoring outfielder, for instance, because there are so many people who score very well in the outfield. In the course of a season, you might only miss out on 60 or 80 points with the 20th-ranked outfielder.
However, with catchers, shortstops, and second basemen, this is not so true. You can gain a massive advantage by having the top scoring player at those positions because even the 5th and 6th-best players will be over 100 points behind the leader. Your gaining points through lack of positional depth.
While the theory is sound on paper, practical application still tells you that you need real points to win, so you have to draft accordingly. I like to use position-mean analysis for the middle rounds of the draft. In early rounds, just take the best player that you need because you cannot win without points - a lot of them.