Are the Horizon (Zero Dawn, Forbidden West) games an RPG?
No.
They're action-adventure games. Like most AAA (and many indie) games today they have RPG elements, but limited ones. That's not just me saying that - the devs involved also say that, as do reviewers etc.
They're judged in a different way to RPGs - though good action adventure games are another genre with a long tail (but even mediocre RPGs have a long tail, whereas mediocre action-adventure games vanish without a trace).
Same with BG 1 and 2, all of the old Gold Box games. The details of how you proceed through the linear story are left up to you until you come to the boxed text or cut scene.
Absolutely right, but if you put them out today as new CRPGs like that, people would be confused and angry.
Very few games actually have a truly significant number of possible branches.
What exactly does this mean? There are so many qualifiers involved that, with respect, I don't this sentence actually has meaning. What standards are you judging from exactly? What counts as "truly significant"? What's "possible" doing there?
I mean, BG 3 is amazing but it also took, what 6 years of development that really stretched into 8 with all the patches and modifications? At a certain point the long tail doesn't pay for the massive investment of time, if not manpower.
Oh yes it does. Time and manpower are interchangeable here - they're money. The main dev cost for all videogames is the staff, the developers - it's the vast majority of the cost (often over 90%) every time with AAA games.
That's what you're apparently not getting. It pays for itself many times over. BG3's revenue is probably well north of $1bn by now, given it had sold 15m copies in
March. Its development costs, for all those years, was $100m.
Do you see the profit margin there? It's insane, even if we factor in tens of millions of advertising (which it didn't actually have) and the limited number of boxed copies it sold being less profitable. You can do the math on how long it would have needed to be in dev to not making money, even with Steam taking 30% of a lot of that revenue and WotC taking an estimated 5-10% (probably 8% based on what we know). It would be decades.
And my brother in Nergal, that's in less than 1.5 years! And the game is still selling. It'll still be selling next, and the year after that. It's probably already past 20m, and may well get past 25m without even having a big price cut (the deepest cut it's had was 20% off, for a week, quite recently, which rocketed it back to the top of Steam's game charts).
BG3 is the best case scenario profit-wise, pretty much. However it is also, not entirely coincidentally, the highest-effort example in terms of what the devs did to get the game made. Effort was rewarded. BG3 is pretty laser-targeted for what CRPG players actually respond to. It's not a coincidence that it has more genuine branching than most CRPGs, too (and way more fully naked elves/githyanki/vampires/druids!).
What's funny is I'm not even that keen on BG3 - it's good but Act 3 is a typical Larian "Whoops let's finish the game real quick!" deal and even Act 2 is polished but weirdly short (much more obvious on a second playthrough). It's just that it absolutely nails what CRPG fans and just a lot of gamers in general want from this kind of game.
I'm not saying this is without risk to be clear - you can screw up and put too much effort into a game, in theory, but I can't think of a single AAA CRPG or even really a story RPG where it didn't actually do at least okay compared to its budget. Like, Mass Effect Andromeda was a bit of a disaster, but as Jason Schrier reported, it only had a budget of $40m (nothing! you wouldn't even be AAA today with a budget like that! It's the same budget as 2007's ME1!), and even in initial sales, it sold 3m+ copies, which was enough to make it a financial success by EA's standards (it was up 5m later the same FY, we don't know after that). Cyberpunk 2077 was something of a car crash on launch, and cost over $400m in development and marketing costs (
inclusive of the years spent on it after release, and the expansion dev/marketing costs), but it sold 13m copies instantly, and whilst it being a disaster slowed it down, it's now north of 25m, and has never gone below half price - and most of the copies sold were full price - so it was still hugely profitable for CDPR (esp. as they managed to sell north of 5m copies of its astonishingly good expansion, Phantom Liberty).
There may be a recent (say, post-2010) AAA CRPG/story RPG I'm forgetting though, which was a genuine flop. The closest I can think of is Pillars of Eternity 2, but that was very much AA, not even close to AAA. It sold appallingly and it's not clear why apart from maybe people didn't like Pillars 1, because it was a flatly better game, and a good game by basically any standards.
EDIT - And to be clear, the reason CDPR put so much effort into fixing and rehabilitating the public opinion of 2077 is that they knew it would translate directly into sales, and keep the game selling long-term