lukestein’s Twitter Archive—№ 1,470

@CookieSci Totally agreed that there isn’t really much advantage. Both are ad hoc normalizations of the 0→x change relative to a percentage change. The ln(1+•) shows its arbitrariness in the “1,” while asinh hides it in functional form. 🤷🏻‍♂️ @lukestein/1091908743292280832