Dissecting the Web with Opera’s MAMA
Detailed information about MAMA’s findings is available in a series of (long) articles published on Opera Developer Community. I suggest you start by reading the introductory document titled, well, MAMA.
I doubt it will surprise anyone that the quality of the HTML used on the Web is – in general – very poor. For instance only 4.13% of the examined 3.5 million URLs passed validation.