There are plenty of old engines that have scaled better than Bethesda's. If they can't get it to modern standards after all this time, it's time to toss the engine.
i think that is a bit unfair to bethesda’s engine. all those other engines have achieved their scalability at the cost of extensibility and easy to work with game systems. this manifests most visibly in how mod support works for bethesda games versus games built on other more “optimized” engines, but it affects the core game design as well.
even if id software released their internal tooling to the public, it wouldn’t be all that useful for making the kinds of mods people make for bethesda games, because their engine isn’t built for all the systems-driven game design that bethesda’s is. that moddability is born out of how bethesda has designed their engine, the gameplay systems they built in it, and the tooling that supports all of this.
it’s truly insane just what you can do with bethesda’s engine with relatively little work. and it shows when you compare to games that try to imitate their game design on other engines. the outer worlds felt really static compared to fallout new vegas and skyrim, because it was missing so many of these systems.
bethesda games have a lot of problems, but ditching their engine for something like unreal or id tech would most likely destroy most of what makes their games unique.
Baldur’s Gate 3 has also done this very well. The build up to finally reaching the city of Baldur’s Gate really is worth the in game hype. The city is massive and the entirety of act 3 is spent within it. They use the standard trick of ensuring you can only visit part of the city, with much of the city being inaccessible but visible. That’s a great way to make the city feel like it’s actually city sized while still ensuring that the part you can explore can be explored in depth (as in, almost every building can be entered and is unique).
As contrasted with the GTA approach where the visitable area is far larger, but you can’t enter most buildings and it’s more generic.
I think it’s pretty hard for an open world game like Skyrim to achieve the way games like BG3 or TW3 do cities, though. After all, Skyrim basically lets you go everywhere, which makes it difficult to fake the size of cities. Skyrim also tries to have not one city but like a dozen cities and towns. I feel like if they wanted to make a realistic city, they’d need to really focus on a small number of cities (probably just one city and a few towns). I’m not sure of the Skyrim scale can really allow for cities as detailed as BG3 or TW3.
Skyrim also has far more in depth NPC, which have routines going all the way from waking up in the morning till they go back to bed. That surely adds scalability issues.
They could do a hybrid approach. Have many unenterable buildings and generic NPCs. But I’m not sure that’s a good idea. That’d make things look bigger, but it wouldn’t really be that much more content and it’d kinda waste our time in traveling to the good stuff. Or they could scale things down. They don’t actually need to span an entire province. They could have focused entirely on one city and surrounding area. But it does come at the cost of more limited lore options and a less varied map.
Personally, I like the Skyrim cities. They’re flawed, but very fun. Not a lot of games have the level of NPC detail that Skyrim has and none of them have the kind of massive, open world that Skyrim and Fallout have (I’d love more games like those).
The sad thing is enterable buildings is pretty achievable. They could have an algorithm to generate a generic interior based on a seed at the time that the player enters the building. It’s not exciting, and it would add complexity but it at least would make enterable buildings without massively impacting install size nor save size.
You can play with procedural generated cities and structures here for example to see the potential. Just imagine that instead of a top down map it generates an interior with generic furniture, clutter and possibly loot. Add in different sets of assets it fills in with based on building type and location and verify it generates the same interior every time it gets the same seed and you have a very efficient system for at least being able to enter every random boring building
Alternatively there is a gameplay argument to be made for only being able to enter buildings that actually do something gameplay-wise so you know what buildings you need to enter
this is why i love witcher 3, it actually has a reasonably large and detailed city! like it’s still pretty unmatched as far as i’m aware.
A reasonably large and detailed city, full of opportunities for playing Gwent.
The Witcher 3 was a really strange DLC to the critically acclaimed game of Gwent
the witcher 3 had the benefit of not needing to run on a ps3
Sounds logical, but seeing how barren cities in Starfield are makes me think it was a design decision rather than a technical limitation.
It’s very much a technical limitation since Bethesda still uses an engine that goes back all the way to 1997.
There are plenty of old engines that have scaled better than Bethesda's. If they can't get it to modern standards after all this time, it's time to toss the engine.
i think that is a bit unfair to bethesda’s engine. all those other engines have achieved their scalability at the cost of extensibility and easy to work with game systems. this manifests most visibly in how mod support works for bethesda games versus games built on other more “optimized” engines, but it affects the core game design as well.
even if id software released their internal tooling to the public, it wouldn’t be all that useful for making the kinds of mods people make for bethesda games, because their engine isn’t built for all the systems-driven game design that bethesda’s is. that moddability is born out of how bethesda has designed their engine, the gameplay systems they built in it, and the tooling that supports all of this.
it’s truly insane just what you can do with bethesda’s engine with relatively little work. and it shows when you compare to games that try to imitate their game design on other engines. the outer worlds felt really static compared to fallout new vegas and skyrim, because it was missing so many of these systems.
bethesda games have a lot of problems, but ditching their engine for something like unreal or id tech would most likely destroy most of what makes their games unique.
Baldur’s Gate 3 has also done this very well. The build up to finally reaching the city of Baldur’s Gate really is worth the in game hype. The city is massive and the entirety of act 3 is spent within it. They use the standard trick of ensuring you can only visit part of the city, with much of the city being inaccessible but visible. That’s a great way to make the city feel like it’s actually city sized while still ensuring that the part you can explore can be explored in depth (as in, almost every building can be entered and is unique).
As contrasted with the GTA approach where the visitable area is far larger, but you can’t enter most buildings and it’s more generic.
I think it’s pretty hard for an open world game like Skyrim to achieve the way games like BG3 or TW3 do cities, though. After all, Skyrim basically lets you go everywhere, which makes it difficult to fake the size of cities. Skyrim also tries to have not one city but like a dozen cities and towns. I feel like if they wanted to make a realistic city, they’d need to really focus on a small number of cities (probably just one city and a few towns). I’m not sure of the Skyrim scale can really allow for cities as detailed as BG3 or TW3.
Skyrim also has far more in depth NPC, which have routines going all the way from waking up in the morning till they go back to bed. That surely adds scalability issues.
They could do a hybrid approach. Have many unenterable buildings and generic NPCs. But I’m not sure that’s a good idea. That’d make things look bigger, but it wouldn’t really be that much more content and it’d kinda waste our time in traveling to the good stuff. Or they could scale things down. They don’t actually need to span an entire province. They could have focused entirely on one city and surrounding area. But it does come at the cost of more limited lore options and a less varied map.
Personally, I like the Skyrim cities. They’re flawed, but very fun. Not a lot of games have the level of NPC detail that Skyrim has and none of them have the kind of massive, open world that Skyrim and Fallout have (I’d love more games like those).
The sad thing is enterable buildings is pretty achievable. They could have an algorithm to generate a generic interior based on a seed at the time that the player enters the building. It’s not exciting, and it would add complexity but it at least would make enterable buildings without massively impacting install size nor save size.
You can play with procedural generated cities and structures here for example to see the potential. Just imagine that instead of a top down map it generates an interior with generic furniture, clutter and possibly loot. Add in different sets of assets it fills in with based on building type and location and verify it generates the same interior every time it gets the same seed and you have a very efficient system for at least being able to enter every random boring building
Alternatively there is a gameplay argument to be made for only being able to enter buildings that actually do something gameplay-wise so you know what buildings you need to enter
It’s not fantasy themed, but the city feels pretty dense in cyberpunk too.