I still think arm strength might make an appearance in your 3B's assist numbers. But I have no idea how to quantify the effect. It would make sense that arm strength contributes a certain percentage likelihood that your 3B converts a ground ball into an out. But what percentage? And how much, over the course of a season, would it really matter? How much range and glove, and/or bat, is necessary to compensate for a poor arm?
There are a lot of moving pieces involved in converting an out on a throw from third, and I have no idea how the sim engine would profile/handicap that event. Does the pitcher generate a lot of groundballs? How often do those groundballs make their way to third (lefty versus righty hitter, does the pitcher's stuff tail toward or away from the batter, how pull heavy are the batters)? What is the 3B's range and glove rating...how often to they convert chances at balls in play into chances at an out? What is the batter's speed rating? How well does he run the bases, is he likely to get a good jump out of the box? How strong is the 3B's arm? How accurate? What are the chances that the 1B digs out a one-hopper? How often does the 1B make an error? Is the missed out at first the result of a weak throw, an inaccurate throw or an error at first?
I do not know how the sim engine processes these likelihoods. But I do assume that a stronger arm (assuming at least average range and glove) results in higher than average assists. And my world that seems to be true. Sorting the season stats by assist, the leaders all have plus arm strength. Those same players also tend to be high on the double play list. And those with the fewest assists have either a weak arm or below average range/glove.
Errors do not seem to be impacted by arm strength. Nor do plus/minus plays.
Again, though, I have no idea how to quantify the results. So I don't know how much range/glove or bat is needed to make up for the noodle arm.