I think that offsetting the biscuit in the first case just leaves a little extra meat between the biscuit and the end of the board. Imagine slicing 5/16” off the end of a board. The resulting piece would be incredibly weak because the grain runs through its thickness instead of along its length, right? You could snap it in two with hardly any effort. It’s the same situation for the top board in the vicinity of the biscuit when you cut a 1/8” biscuit slot into the face of a board 3/8” from the end. Moving that slot inward 1/8” allows just a bit more long-grain connection between the slot and the end, and makes the wood there a little bit stronger.
Offsetting the biscuit toward the bottom of a shelf seems a little more intuitive: you want more of the shelf’s thickness above the biscuit to provide extra resistance to whatever load is on the shelf. If you offset the biscuit in the other direction, toward the top, then all the weight would be carried by just 1/8” or so of wood, and that doesn’t seem good.
All that said, it’s hard to know how much stock to put in these ideas if the author didn’t cite a source or otherwise back up the claims. It might be fun to test these ideas out and see whether they really make a difference.