I was mainly responding to the claim that Hoskins had average to above average speed (I thought I hit reply to AF, but I guess I didn't). I, too, was surprised at how close Hoskins and Herrera's peak speed was, but that is because of Herrera's number and not Hoskins. I really didn't put much thought into the numbers. I would guess that Herrera is quicker (gets to top speed in less time) than the numbers indicate or the common perception of Herrera's speed is simply wrong.
As andyb pointed out, combined with his below average speed and his catch probability chart, it seems unlikely that he can improve much.
But I agree with your 2nd paragraph and am curious how the team is valuing that trade off. Earlier in the season it seemed that poor defense cost the team a few games. However, maybe that was variance with a number of non-3 outcome games being strung together in a small sample. It obviously hasn't been a problem recently.