"People don't finish long games, therefore you shouldn't make long games," was a specious corollary developed by publishers to give teams smaller budgets.

Most people also don't finish short games.

AFAIK no one has ever researched the correlation of "didn't finish" with "unhappy about buying/playing the game". I asked the heads of a highly reputable research group if they ever had and they admitted they hadn't even thought of asking.

Confirmation bias is epic.

Short games can be good. Long games can be good. But, "People don't finish long games," has never been a sound argument for not making them unless your ultimate goal is game completion rate (as opposed to sales and player satisfaction).

Sincerely,

A Long Game Maker