Previous research on modified numerals established some widely accepted contrasts between comparative modifiers (CM) and superlative modifiers (SM) (see \citealt{buring_least_2008,geurts_at_2007,nouwen_modified_2015,mayr_implicatures_2013,schwarz_consistency_2016} a.o.), such as: (i) CM don't but SM do give raise to obligatory ignorance implicatures; (ii) CM can scope over or under existential modals (EM) but SM have to outscope them. A \textit{no more than Num} construction (NMC), where negation and comparison are combined in a way exemplified by an English sentence \Next from \cite{nouwen_upper-bounded_2008}, to this day the most developed formal treatment of NMC, is then claimed to allow both scopes w.r.t. EM (\citealt{nouwen_upper-bounded_2008}) and to have scalar bounding inference, signalling speaker's well-informedness (ibid), since English \textit{no more} construction (unlike class B modifiers) give raise to equality readings like (=50) for \Next. Such claims seem to be supported by the comparative morphology of NMC. I bring new experimental and corpus evidence against both claims, showing that (cross-linguistically) (i) NMC can be interpreted only with wider scope then EM; (ii) NMC is compatible (mostly) only with speaker's insecurity (or so-called variation) readings as SM. The experimental and corpus evidence comes from Czech as it was observed before that Slavic languages (unlike English) generally don't support speaker's well-informedness NMC interpretation.