Halo was planned as a Live Service game, but as an initial release, it was pretty much a fully self-contained package as content-rich as any other non-GaaS game. So I don't think pointing to this is a great example because Halo didn't fail because of their Live Service plans. It failed because it was just disappointing as a game and underperformed because it was basically given away for free on GamePass.
Marvel Avengers also didn't fail because it was a Live Service game. It failed because it was just a bad game.
The beauty of Live Services, is you can build a game as a fully self-contained package at launch, e.g. Destiny 1/2, GTAV, Warframe, and then continue delivering a consistent flow of content updates to keep people engaged.
If at launch you deliver a quality product, people will be itching to play more content, and keeping them fed with regular updates is a net positive. Destiny, GTAV, Fortnite, and others have done this.
If at launch you end up with a lackluster product, it takes a longer-term time and resource investment to turn the ship around, but if you persist you can still build something very successful. Warframe, Sea of Thieves, etc are great examples of this.
Unlike the OP, I'm not against GaaS games at all. But I do think that devs that plan for GaaS, without figuring out how their initial release is going to freaking slap, will be fighting an uphill battle trying to convince gamers to buy in in the longer term. Best to nail the first impression as otherwise you're left with trying to catch lightning in a bottle.