Easy: Among Us, for the many months that it toiled in nearly complete obscurity to the point of the development team nearly giving up the ghost and quitting, until one specific streamer picked it up and gave it the critical mass attention it needed to go absolutely huge. Had the developers had slightly less resources and shut down the servers before that crucial moment it would have never happened, even though clearly Among Us is a fantastic game in retrospect.

Quality is necessary but not sufficient for success.

Also: a while back, I wanted to empirically test the "any good game will succeed on steam, and if it doesn't succeed, it isn't a good game" hypothesis by setting up a little metagame called "Steam Prophet". We would track upcoming steam games and try to predict which ones would succeed BEFORE they came out. We eventually disbanded the project when we got so reliably good (but not perfect) at predictions that we got bored and quit. Turns out that just a few leading metrics were incredibly predictive of day 1 (and thus day 30, day 60, and day 90, and year 1) success -- chiefly the number of followers the game had on Steam on launch day. It is a rare game indeed that can succeed on Steam without first having banked tons of high quality wishlists (of which followers is a public proxy metric for).

The spirit of Steam Prophet lives on today in the form of Simon Carless' Game Discover Co newsletter, which obsessively follows upcoming and newly released Steam Games: https://gamediscover.co/