The idea to force blocks bigger than 1 MB is inherently flawed.
Not exactly. There's more to it than that. They can't use the hardfork bit because they're trying to trick lite clients into following their fork, thus causing maximum disruption (they don't seem to have any other goal). In that respect, forcing >1MB blocks is a very good idea, or it would be if their code actually produced >1MB blocks. The inherent flaw is that it actually
doesn't with the default settings (even if the mempool has more than 1MB of transactions), with the result that all their miners stopped dead as soon as the fork activated, and it took over a day for anyone to figure out what the problem was (and the so-called developers spent most of that time denying that there was any problem at all and insulting anyone who tried to point it out).
It's the typical level of incompetence we've all come to expect from these clowns.