I do think that's somewhat true in software, but I think games are more of a mix of art and software, and in art it's much less common to find consistent winners outside of the mediocre category--- Thomas Kinkade makes consistent nice-looking paintings, but Picasso produced a lot of crap amidst his masterpieces. I'm not sure it's possible to take risks on game design and not end up with a reasonable number of duds, even if you're solid and professional on implementation.
It is true that most game-industry positions aren't really for designers, but these days a large number do involve some amount of design, especially if you want the game to be unique in some way. For example, AI programmers are in an odd position of doing "just implementation", but implementing a system that deeply impacts what kind of gameplay and design is possible (in some cases they quite literally write the tools that the designers will use to do things like write NPC behavior).
Is disproof by counter example allowed? Blizzard is one case of a company that has consistently produced masterpiece games in all of their major ip's at least since Diablo, Starcraft I, and Warcraft 3.
Only because they scrap ones that don't work. There's a difference between consistently producing incredible work and consistently releasing incredible work. Curation is a big part of industry success.
It is true that most game-industry positions aren't really for designers, but these days a large number do involve some amount of design, especially if you want the game to be unique in some way. For example, AI programmers are in an odd position of doing "just implementation", but implementing a system that deeply impacts what kind of gameplay and design is possible (in some cases they quite literally write the tools that the designers will use to do things like write NPC behavior).