Quoted By:
So, in a pokemon's data there's two bytes that detemine their size and weight. From what it looks like, 0 is XS and 255 is XL, or in this case alpha. Naturally it takes extra math to convert this into a multiplier (which would be a float), and you're limited to the range of a byte.
It would be easier to just put the floats in the data directly, which would remove size limitations, it would remove the need to convert shit, and it would be six extra bytes in a 350 byte structure, which is essentially a rounding error today.