Starfield by Bethesda -- PC and Xbox exclusive -- 2023...?

Agreed. That is a worthy improvement.

Todd Howard on the gains from new consoles.

“It’s led to our largest engine overhaul since Oblivion, with all-new technologies powering our first new IP in 25 years, Starfield, as well as The Elder Scrolls 6.”

He doesn’t say it’s not Gamebryo…

Because it’s definitely Gamebyro.

I thought it was confirmed Gamebryo - I just saw a quote about how it was the biggest leap in the engine’s revision since Oblivion yesterday, something like that.

It’s an overhaul, so yes.

Doesn’t he say that for every new game launch? I swear I heard it for Skyrim, Fallout 4, and then Fallout 76.

He seems to trot that out a lot. :)

And, you know, I’m sure there’s truth to it. You have an engine that ancient, I’m sure there’s a lot of work and overhauling necessary to make it do what modern games need to do. Which is kind of the problem, IMO. I mean, what, they’re gonna have to figure out how to get ray tracing to work in Gamebryo now or something? Poor engineers.

Wouldn’t starting over on a new engine take even longer though? That was one of the main problems over at EA, where they insisted everyone switch over to Frostbite. Maybe it’s better that Bethesda continue to toil over an older engine. At least it’s got it’s good points, like being highly moddable, for instance. And I can’t deny that each time it does get a lot better in specific areas.

I don’t really know because I’m not there but I see this all the time in the field. Yes, it’s more of an investment upfront but the payoff is you don’t spend the next two decades twisting ancient code into knots trying to get it to do things it was never intended to do. At a certain point it’s better to have a new solid foundation than to spend another ten years bolting on feature after feature onto a rickety platform and keeps tipping over and falling apart. Look at all the problems FO76 suffered when they tried to cram a client/server architecture into their codebase.

The problem with EA and Frostbite isn’t that they wanted everyone to switch over to a single engine per se, it’s that the engine is good at doing Battlefield but really bad at doing the kind of things that Bioware needs and wants to do.

Bethesda has the money to have a separate engine/tools team building their next engine while other teams work on games. This could have happened in the 9 years since Skyrim shipped, but they just don’t seem to want to invest up front. Also a common thing I see among a lot of the executive types who see the cost but don’t really understand the current difficulties or the benefits of investing.

Again, I don’t have any firsthand knowledge of Bethesda so this is just my own two cents.

Well, Skyrim was finally their huge breakout monster hit. So I’m hoping that they’ve been doing exactly that, having a separate engine/tools team building their next engine while the design teams have been working with the new Gamebryo updates to finish Starfield and ES6. That way the new engine is ready for when they start work on Fallout 5 once Starfield is out the door.

What I want Bethesda to do is get rid of the jankiness. I love their Fallout games but watching my companions walk into walls and other objects when moving towards my character starts to take me out of the game after about the 100th time. Watching an enemy bounce 100 feet into the air because I just hit him with a shotgun also doesn’t work for me. There’s just too much jank in their games.

Yeah, that’s kind of where my disappointment is. It’s going to be over a decade between ES5 and ES6, I had hoped that crazy gap was them building either a new engine or taking an existing engine (Id, Unreal, whatever) and customizing it to their unique needs. That doesn’t seem to be the case, though. ES6 is still a year away at minimum and is going to use the same decrepit Gamebryo engine, same with Starfield. I do suspect that part of the reason it’s taking over a decade to produce a new Elder Scrolls is dealing with their creaky codebase.

That sort of thing is just a personal bugaboo to me, I guess. My company used to be owned by a very big multinational and we fought with them for over 15 years about the need for a new codebase built from the ground up instead of the cheaper (initially, anyway) bolting on code that was never meant to do what we were trying to do… I mean parts of this code base came from our DOS-era product! I’m working in that part of the code today, in fact. It blows. :)

We purchased our company back and are independent now and we’re finally investing like we knew we needed to. Frustratingly, today we also have a small fraction of the resources we used to have when we were initially acquired so the job is harder, we have to juggle both new development with building a new foundation with few resources. But at least the leadership now fully understands the cost of the 20 years of technical debt we’ve piled up and why we need to dig our way out of it, as painful and expensive as it is now.

Obsidian did it with a vastly smaller team and less money for The Outer Worlds. So maybe it isn’t that big of a deal.

Of course Bethesda no doubt has extensive content creation tools and a pipeline customized for their current engine so the technical debt might be much higher than just the engine itself.

Skyrim was a huge hit, no doubt, but both Morrowind and Oblivion were immensely successful for their time.

Well, the Outer Worlds is a much, much smaller game. Just the sheer scale doesn’t even compare. Plus I doubt it’s as moddable either.

Maybe that should have been the way to go though. Do a smaller, less ambitious game to work out the kinks on an engine change, and then do the massive games in that new engine with the new content pipelines worked out.

We don’t have any idea how moddable it is. They didn’t release their tools, which bodes not well. But it’s clear you can make a game like that with a brand new engine. Obsidian is not exactly known for its swift releases at high quality and yet The Outer Worlds is a gem.

Can you pick up, drop, throw, etc. random household objects in Outer Worlds? I think Gamebryo at least tries to do a bit of (janky) physics.

PC Gamer quotes Todd Howard talking a little about Starfield at the Brighton Developer Conference:

During an interview for this year’s Develop: Brighton conference, Todd Howard shared some more info on the eagerly awaited space RPG Starfield. The literally biggest news is the expected size of the map, which will be even larger than that of Fallout 76 (itself four times the size of Skyrim). Whether this refers to a galaxy map or individual planets is not clear.

In order to effectively create that much landmass, part of it will be procedurally generated. According to Howard, procedural generation, which has been a tool in past games, is a good tool to create large landmasses quickly, but of course distinct landmarks are still handcrafted, which is standard practice for most open-world games nowadays. In Starfield and The Elder Scrolls VI, cities should become more lively both by being larger in size and by featuring more NPCs, who themselves will be playing a “larger role”—hopefully meaning they’ll be more fun to interact with.

Howard also mentioned a planned overhaul to Bethesda’s Creation engine, calling the improvements a “larger jump [in quality] than that from Morrowind to Oblivion.” In particular, he named rendering, animation, and AI as areas where the most significant improvements have been made.

Although I don’t think procedural generation was on anyone’s list of wants the other parts sound ok.


I wouldn’t necessarily worry about this part. I believe much of Obliv/Skyrim landmass was started with procedural generation, then hand touched to bring it to final quality and desired look/layout. This is probably similar, although perhaps with more space between heavily hand touched areas.

Yes, procedural landscape generation was a huge talking point in the PR campaign for Oblivion.

I believe he trots this out every single release. It always ends up being the same janky 20 year old engine with a couple improvements bolted on the rickety mess here and there.