Okay, but these cars are really quite far from "minor cosmetic differences":
In fact they're really quite different - the Alfa most of all, since the 4C is a light, crap, 200hp road car and the 4C Gr.3 Road Car is a GT3 car (fictionally) converted from a (fictional) race car as a homologation model, which in the real world would mean almost no common components; a GT3 car only has to share an engine position with the road car it's "based" on, and engine configuration with an engine manufactured by the brand. I'd imagine (if the 4C were eligible for GT3 racing) the monocoque would be the same - the 4C is fully carbon fibre after all - but that's about it.
The 3.0CSLs aren't, for example, the same car with different years. One is a performance road car, developed from the 3.0CS (itself a development of the 2800CS) with a mild overbore to allow it to be eligible for 3.0 class racing, while the other is literally a homologation special with - despite the name - a 3.2-litre race engine. Fun fact, the rear spoiler was illegal as a factory fit item, so BMW shipped them with it removed and in the boot.
While the M4 Safety Car might be a safety car livery and light bar on an M4, it's entirely relevant to have both the regular high performance road car and the ready made motorsport safety car in a motorsport game. You could argue that perhaps there should be a way of modding the M4 into an M4 Safety Car. I'd not argue against that, so long as it was as easy to do as just buying an M4 Safety Car in the first place (which isn't anywhere near as easy as it should be in GT Sport), like "buy M4, go to GT Auto, click 'make safety car lol', done".
The two DC2 Integras only really differ if you care about the differences. In GT it'd amount to an almost imperceptible difference in performance (different gearing, different torque peak) and a handful of visual differences you'd really need to look for.
... and this is highly problematic. What's enough of a difference in performance to offset cosmetic differences - and vice versa?
Something like the DC2 Integra is pretty clear cut - they're damned close in performance and design, and name, and are fundamentally the exact same vehicle with subtle changes. The Clio RS200 is probably the clearest cut, because it's literally the same car twice only with LED headlights the second time. The Alfa is as clear cut the other way: more than twice the power and all that extra aero gives a significant difference in pace.
The 3.0CSL is very, very unclear. As well as a pretty major bundle of cosmetic differences (homologated aero kit, 3.2 engine, slightly more "leicht"), the performance differences are more noticeable due to more power and a wider torque band. Now, we're not talking about Civic VTI vs Civic Type R performance differences, but there's a difference... How different does the difference need to be before it's different enough?
... and then we have the main problem of the problem. Which do you leave out?
For the most part you might say "oh, that's easy, the slower one, because it's a racing game and gotta go fast", but it's really not that simple. I mean, there's an absolute ****load of slow cars in GT and there always has been. It's never been about only the fastest cars, which is why we've had Civic VTIs as well as Type Rs (among many, many other examples).
Sometimes the slower car is the one preferred - people want to drive the one they own. Sometimes it's both - they want to drive the one they own, and aspire to drive the faster one they don't own. A guy up thread wanted to sack off the faster one in a lot of cases because it's not a real car (as if that's ever mattered in GT). Sometimes it's super complicated, and the 3.0CSL is a perfect example: a lot of people prefer the original 3.0CSL, because it's clean and not a gaudy, bewinged monstrosity with a Jimmy Hill chin and knives up the front wings. It's the ultimate expression of the E9 in its original, pure shape. A lot of people like the Batmobile because lol wings.
Of course in principle you could have the original car and GT Auto taking care of the differences, but then you hit a licensing barrier; is BMW willing to allow you license both cars but have players modify the 3.0CSL into the 1973 homologation model? The answer to that one might surprise you (as in there's a solid chance it'd be a "no"; it's not what happened in the real world, and BMW won't want you to think that. Brands can be highly protective of their heritage stuff [or even current stuff, in the case of Maranello], so it's both or neither).
Honestly, on stuff like the two Clios, or the VGT/VGT Gr.1s, or the heinous crap pulled with the MX-5s in GT5/6 (which was well-meaning, but fundamentally broken), I'm right with you. That's straight up duplication. But in most cases cars highlighted as "duplicates" are really a lot more difficult to justify as duplicates than it look on the face of it.
Overall I agree.
But when it comes to choosing which version, it shouldn't be that hard. Well, atleast when it comes to Civics and such.
I had a Escort and a Focus Mk2 post-facelift on the family. I would gladly get the Escort RS Cosworth and Focus RS Mk2, without feeling the need for the base models (or even the ST version of that Focus). I always thought that the sportier version is a good representative of said generation of car. This applies to a Civic, to a Golf, a Focus, a Megane, 4-series, C-class, etc. The top of the range, is the better one to have in the game, even if it's not 100% what we own, but it's close enough. And these performance versions are usually the highlight of the entire model.
Even as Civics go, there are a lot more variants that should be prioritized over a slower version of any of the existing models. Be it the 2007 Sedan (premium in GT6), the EP3, 2011 Mugen 2.2, and even other models from the brand, like the CR-Z, CRX, Prelude, newer S2000...
As duplicates go, in past games things wen way past the Miatas. Many Japanese cars had duplicates, I wouldn't be surprised if the list reached almost 200 duplicates. Let's not forget that there were also 3 seasons os Nascar cars in GT6 (2010, 2011 and 2013), which comprised in 3 unique cars each, which then had multiple liveries counting as unique cars. Let's be honest, in GT5 and 6, most players when buying an R34, they would go with the Premium version, making the others pretty much pointless. And the problem is not the having that many versions in the game, is the fact that they count as unique cars. If those trim levels or whatever counted as the same car, and when the player buys them, there was the option of choosing the model, the same way we choose a paint (not counting as a new car), everything would be fine.
Btw, unless I missed it, the Porsche VGT is not on the "over 400" SOP photo. Which
could mean that the list is somewhat old. Either that, or they don't have the car fully ready for the game yet.
For whatever reason the output rate does seem to have reduced. I can't find it in searching but I seem to recall up to the end of GT Sport updates they were outputting around 70 cars per calendar year, between the cars in games at launch and updates.
In 2020 and 2021 they added only 7 cars total to GTS. That means there should have been around 133 models finished in those ywo years that weren't added to GTS. Yet as above, we're only getting around 83. That would mean per year output has gone down to around 45 per year.
With outsourcing, that number should actually go up. With the addition of external modifications to cars (which didn't happen in GTS), which certainly takes it's fair share of resources, I'd say the value should level to what it was for GTS. COVID certainly took a hit on it, if not on the modeling process, atleast on the scanning. Don't know how they do the sounds, maybe it's something that can't be done from home, so it could also slow down the process.
But yeah, most of disappointment, must come from the expectation that the game should have atleast 500 cars, judging by the rate they were adding them to GTS.