User blog comment:Hyp cos/SCG(n) and some related/@comment-10429372-20140812170828/@comment-1605058-20140812200832

@Deedlit Yes, I also consider that intuitively true. I'm not saying this is not true, I'm just saying it might not be true. With my last statement I meant that f_e_2(...) (value predicted by bound) is vast larger than f_e_1(...) (real value).

@Wyth That's true, I used to also use these things. At the time of writing that article I didn't know mamy things, and I was taking many things for granted (e.g. that tree(n)>f_SVO(n)). Right now, after I learned that such bounds use alternative FSes, I'm trying to avoid all such claims about bounds.

(btw, I've pointed out that division by constant might not be fully preserving bounds when constants get too large)