My reasoning was this:
Part of Le Chatelier theory is that the system will attempt to counteract the change, but will never fully counteract it.
Thus, I found it more likely that it will undershoot it, and thus approach it asymptotically, rather than overshoot it and come down to it.