Which is more garbled?
For English-speaking human readers, it’s fairly easy to allocate a ‘Garbledness Index’ to the passages above. For a machine though, things are considerably more complicated.
P.K. Saxena, Pratibha Yadav, Girish Mishra of the Indian Government’s Scientific Analysis Group (SAG) in Delhi, clarify things in their 2010 paper ‘Index of Garbledness for Automatic Recognition of Plain English Texts’ (Defence Science Journal, Vol. 60, No. 4, pp. 415-419)
“In this paper, an Index of Garbledness (IG) has been defined for automatic recognition of plain English texts based on linguistic characteristics of English language without using a dictionary.”
Therfore, when dealing with Cryptography
“ … [the Index of Garbledness] can be used for identification of the solution while going for exhaustive trial, especially in those systems where either the correct text is formed during a key trial or it is totally garbled.”