13% comes from AMD slides? Perhaps no-one tested with games and settings AMD selected? Cherry picking games and/or other software for presentation slides is normal.
But it sets expectations higher than they should be and then the product looks bad when it's tested by independent reviewers. If AMD says it's 13% faster, but most reviews find it's slower or >5% faster, it looks like AMD was trying to mislead people. It doesn't matter if other companies do it too. No one should be doing it and they should be criticized if they do.